Citation
A cognitive model of journal writing of college students in an introduction to literature course

Material Information

Title:
A cognitive model of journal writing of college students in an introduction to literature course
Creator:
Cole, Margaret A
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
xiii, 357 leaves : ill., forms ; 29 cm.

Thesis/Dissertation Information

Degree:
Doctorate ( Doctor of philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
School of Education and Human Development, CU Denver
Degree Disciplines:
Administration, curriculum, and supervision

Subjects

Subjects / Keywords:
Literature -- Study and teaching (Higher) ( lcsh )
Literature -- Study and teaching (Higher) ( fast )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Colorado at Denver, 1992.
Bibliography:
Includes bibliographical references.
General Note:
Submitted in partial fulfillment of the requirements for the degree, Doctor of Philosophy, School of Education and Human Development, Administration, Curriculum, and Supervision
Statement of Responsibility:
by Margaret A. Cole.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
28483419 ( OCLC )
ocm28483419

Downloads

This item has the following downloads:


Full Text
A COGNITIVE MODEL OF JOURNAL WRITING OF COLLEGE STUDENTS
IN AN INTRODUCTION TO LITERATURE COURSE
by
Margaret A. Cole
B. A., Western Maryland College, 1963 M. A., University of Arizona, 1965 B. A., University of Colorado at Denver, 1975
A thesis submitted to the Faculty of the Graduate School of the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Doctor of Philosophy
Administration, Curriculum, and Supervision
1992


1992 by Margaret A. Cole All rights reserved.


This thesis for the Doctor of Philosophy degree by Margaret A. Cole has been approved for the School of Education
by
Brent G. Wilson
Martin A. Tessmer
Ellen A. Stevens
Date


Cole, Margaret A. (Ph. D., Administration, Curriculum, and Supervision)
A Cognitive Model of Journal Writing of College Students in an Introduction to Literature Course
Thesis directed by Associate Professor Brent G. Wilson
ABSTRACT
The purpose of this research was to study how college students use journals in understanding stories. Such information is prerequisite to developing more effective instructional/learning strategies to scaffold (support) students in comprehending literature. The literature describes the use of journals in all levels of education, but provides little information about (a) the relationships of journal writing to individual learning characteristics and task difficulty, or (b) the types of questions and comments students naturally generate when they write journals to help them comprehend reflective short stories in the context of a credit-bearing course (i.e., an ill-defined domain, an appropriate incentive, an ecologically valid context, and the opportunity for the learner to monitor his/her comprehension and generate whatever questions or comments he/she wishes). The psychology literature on adjunct questions, self-explanations, schema theory, and the generative model of leaming/constructivism was incorporated to shed light on journal writing.
The methodology was primarily qualitative. Data included students journals and
f
class comments; surveys of the helpfulness of journal writing and job aids; individual learning characteristics (e.g., reading ability, field dependence/ independence, academic
IV


record, course grade, and confidence); and case studies of five students. Findings indicated that (a) students generally viewed journal writing as a flexible cognitive tool which helped them construct the meaning of stories; however, students perceptions of journal writing 'are particularly sensitive to implementation; and (b) writing journals scaffolds (supports) students in attending to details, asking questions, and answering
l1
their own questions; however, high-ability readers tended to engage in different cognitive activities than lower-ability readers.
Journal writing is a very dynamic process; no single cognitive model can describe journal writing either within or across students. A two-part model was developed to describe the process: (a) factors which influence journal writing (task difficulty, individual learner characteristics, teacher expectations, student strategies, external resources, overt activities), and (b) components of journal writing (establishing a goal, constructing the textbase, constructing the situation model, predicting outcomes, identifying significant elements, reflecting on meaning, and assembling the schema).
Recommendations for implementation of journal writing are included.
This abstract accurately represents the content of the candidates thesis. I recommend its publication.
Signed
Brent G. Wilson


DEDICATION
I dedicate this thesis to my husband, Jim, and
to all the students whose journal writing inspired this study.
i


ACKNOWLEDGMENTS
I wish to express my appreciation to the members of my committee for their varying perspectives, thoughtful suggestions, and support.
Brent G. Wilson Martin A. Tessmer Dian E. Walster Ellen A. Stevens Edith W. King
vu


CONTENTS
CHAPTER
1. INTRODUCTION.....................................................1
Limitations of Prior Research.................................6
Journal Writing............................................6
Learner-generated Questions................................6
Justification of the Study....................................8
Research Questions for a Study to Develop a Cognitive Model of Joimial Writing..............................................16
Purpose of the Study......................................16
I
2. REVIEW OF RELATED LITERATURE.................................. 18
Journal Writing..............................................19
Adjunct Questions.......................................... 27
Question Position.........................................28
Question Level/Task Difficulty............................29
Additional Findings on Adjunct Questions..................31
Self-Explanations.......................................... 32
Limitations of Research on Adjunct Questions and Self-Explanations............................................33
Schema Theory................................................34
Relationship of Levels of Questions to Levels of Processing.34
Definitions of Levels of Processing..................... 35
The Generative Learning Model................................36
vrn


I
I. '
Learners Dispositions and Goals.................... 37
Focus of Research on Learner-generated Questions.........39
Learner-generated Questions as Study Aids................39
Summary of Learner-generated Questions...................44
Summary.....................................................45
3. METHOD........................................................ 46
Conceptual Framework........................................46
Background and Setting..................................... 50
Background of the Researcher.............................50
, Setting for the Study...................................51
Subjects/Respondents.....................................53
Materials and Resources.....................................54
Data Collection.............................................59
Procedure................................................62
4. CASE STUDIES.............................................. 82
Selection of Interviewees...................................82
lisa........................................................84
Personal Background.................................... 84
[Responses to Interview (Fifth Week).....................88
Summary..................................................93
Colleen.....................................................96
Personal Background......................................97
Responses to Interview (Fifth Week).................... 98
IX


Summary..................................................100
Trish..................................................... .101
Personal Background......................................101
Responses to Interview (Sixth Week)......................103
Summary..................................................106
Frances......................................................107
Personal Background......................................107
Responses to Interview (Eleventh Week)...................110
Summary................................................. 115
Eric.........................................................115
Personal Background......................................116
Responses to Interview (Eleventh Week)...................119
Summary ............................................... 122
Summary......................................................123
5. ANALYSIS OF RESEARCH QUESTIONS..................................125
Research Question 1: What Metacognitive and Cognitive
Activities Do Students Engage in When They Write Journals?...125
Metacognitive and Cognitive Strategies...................126
Cognitive Activities.....................................131
Research Question 2: How Are These Activities Related to Individual Learner Characteristics?........................ 141
Research Question 3: What Types of Questions and Comments Do Students Write in Their Journals?....................... 146
Focus of Cognitive Activities............................146
Comparisons of Journals 3 and 4..........................147
x


i Story-specific Categories in Journals 3 and 4.................163
Omission of Significant Elements............................ 164
Research Question 4: How Are Questions and Comments Related to Individual Learner Characteristics?...........................166
Research Question 5: Can a Taxonomy Describe Students'
Questions and Comments?..........................................174
Research Question 6: How Do Students Perceive Journal
Writing?....................................................... 178
Group 1.......................................................179
Group 2.......................................................186
Relationship Between Grades and Ratings.......................190
Most Helpful Aspect of Journal Writing .......................192
Least Helpful Aspect of Journal Writing.......................194
; Students Suggestions for Making Journal Writing More
Helpful.......................................................196
Research Question 7: How Do Ratings of Journal Helpfulness
Relate to Other Variables?.......................................199
Research Question 8: How Do Students Rate the Job Aids?..........202
Student Accounts of Journal Writing...........................202
Catechism.....................................................204
Research Question 9: How Do Ratings of the Job Aids Relate to Other Variables?............................................. 207
Student Accounts of Journal Writing...........................207
Catechism.....................................................208
Research Question 10: Is There a Cognitive Model That Can
Capture the Processing of Most Students' Journal-Writing
Activity?........................................................210
xi


i1
Summary........................................................221
6. DISCUSSION........................................................229
Interpretation of Findings.....................................229
A Flexible Cognitive Tool................................. 229
A Valuable Instructional Tool...............................232
Individual Learner Characteristics..........................239
Social or Individual Level of Knowledge Building?...........243
Limitations of the Study and Recommendations for Future
Research.......................................................245
r
Recommendations for Implementation of Journal Writing..........253
Alternate Strategies........................................263
Summary........................................................264
APPENDIX
i 'j
A. Faculty Index of Story Difficulty................................ 265
B. Survey of Journal-Writing: Beginning of Term.......................267
C. Survey of Journal-Writing: End of Term........................... 270
D. Survey of Background...............................................273
E. Examples of Short-answer Quizzes...................................276
F. Guidelines for Semistandardized Interviews After First and Last
Journals......................................................... 278
G. Catechism for Stories..............................................280
H. Student Accounts of Journal Writing.............................. 282
I. On "Fleur," by Louise Erdrich......................................285
xu


J. Stories on Which Students Wrote Journals.................287
K. Sample Journals From Group 1...............................310
L. Taxonomy Developed From Students' Journals.................320
M. Examples of Questions and Comments in the Taxonomy...........322
N. Biographies of Students in Groups 1 and 2....................330
GLOSSARY OF SELECTED TERMS.........................................337
REFERENCES.........................................................343
xui


CHAPTER 1
INTRODUCTION
The general acceptance among educators of writing as a learning tool (e.g., Anson & Beach, 1990; Emig, 1977; Fulwiler, 1980,1987,1989; Irmscher, 1979; Moss, 1979) has led to widespread implementation of writing-across-the-curriculum programs. One of the most popular manifestations of writing as a learning tool in liberal arts programs is journal writing (see Glossary). Articles on the subject abound in professional publications, reflecting diverse philosophical and pedagogical positions. Assignments range from totally open-ended journal writing to highly directive writing, such as expressing gut reactions, writing summaries, elaborating (e.g., comparing/ contrasting content to ones own experiences, or applying concepts), evaluating content, or even answering specific questions (e.g., Duke, 1982; Flitterman-King, 1988; McCormick, 1985; Nugent & Nugent, 1985). However, McCormick (1985) observes that a iriajor criticism of the dominant approach to using journals in literature courses is that it calls merely for associative, touchy-feely reactions, and rather than opening up students responses to texts, they restrict them to what students already know (p. 837). Thus, she advocates writing which is given focus by the instructor.
The diverse applications of journal writing extend beyond literature courses. A literature search of the ERIC database (CD-ROM, 1983-1990) using the term journal writing illustrates the scope of interest. The search located 372 items; another 19 were listed (under journals) in the 1991 ERIC printed index. However, barely one third of the entries reported formal research. Table 1.1 suggests the breadth of interest in
1


journal writing. Of the 372 entries in the 1983-90 CD-ROM database, 25 (6.7%) related to literature courses in higher education; 7 of the 25 (1.8% of all the entries) reported formal research. A search of the same database located no entries using the terms journal writing and instructional technology or educational technology.
In spite of jthe widespread use of journals as an instructional strategy, Anson and Beach (1990) reported that only a handful of the more than 600 publications on writing across the curriculum or writing to learn, provided any empirical support for an ovemhelmingly endorsed.. .practice (p. 2). They added that the research studies were not conclusive and raised.. .many interesting problems about the nature of writing and learning (p. 2). The only consistent findings on journal writing seem to be (a) that, when they have been asked, students have said that they found journal writing helpful (e.g., Cole, 1991; Cummins, 1989; Hettich, 1990; Schwartz, 1989), and (b) that reviewing all journals is very time consuming (e.g., Hettich, 1990; Roth, 1985).
The term journal writing is not even used consistently in the literature. As Hedlund, Furst, and Foley (1989) reported, the terms journal, diary, and log are often used synonymously. The 372 entries mentioned above in the ERIC database included many whose texts actually used the terms diary or log. Some used the term journal to describe almost any writing assignment other than an essay or term paper, such as responses to adjunct questions (e.g., Hettich, 1990) and dialectical reasoning (Jolley & Mitchell, 1990). Others emphasized the traditional features of a journal which are also salient in the current study; for example, the personal nature of its content (Hedlund, Furst, & Foley, 1989), active involvement (Hedlund, Furst, & Foley, 1989; Roth,
2


Table 1.1. Sample of ERIC Publications on Journal Writing
DOMAIN / MAJOR FUNCTION
LEVEL DISCIPLINE
1st grade mathematics
lst-3rd, mixed reading
6th grade reading
6th grade mathematics
1st-12th (all)
junior h.s. reading
middle sch. English (literature)
high sch. English
high sch. English (literature)
high sch. geometry
higher ed. biology
higher ed. communication
higher ed. engineering
higher ed. civil engineering
higher ed. English
higher ed. English as a second language
higher ed. geography
higher ed. linguistics
higher ed. literature
higher ed. mass media
higher ed. develop, reading
higher ed. physical education
higher ed. psychology
higher ed. psychology
higher ed. sociology (sex/gender)
higher ed. speech/ communication
higher ed. study skills & literature
higher ed. teacher education (curriculum evaluation)
higher ed. teachtr education
general general
*Research article
reflective practice comprehension
conative, self-concept (learned
helplessness)
problem solving
metacognition
metacognition
comprehension
communication & writing
comprehension
thinking skills
concepts; attitudes
reflective practice writing
problem solving, communication
composition
cultural understanding
expressive writing for students, diagnostic tool for teacher (student determined) comprehension
diagnosis of teaching effectiveness
& student learning
metacognition
affective, conative
course concepts
critical thinidng
reflective and critical thinking
multiple (e.g., brainstorming, self-evaluation)
(effectiveness of annotation vs. journal writing strategies) clarify understandings & relate learning to personal and professional lives reflective practice
independent thinking; writing skills
AUTHOR
Wason-Ellam, 1987* Barone, 1990
Coley & Hoffmann, 1990
Risk, 1988 Kuhrt & Farris, 1990 Pyle, 1990 Myers, 1988 McGuire, 1990 Wilson, 1989 Linn, 1987*
Trombulak & Sheldon, 1989
Buell, 1989 Rumpf et al 1988 Selfe & Arbabi, 1983 Stanley, 1989 Steffensen, 1988
Sublett, 1988
Anson & Beach, 1990* Flitterman-King, 1988; Holland, 1989; Nugent & Nugent, 1989 Tamove, 1988
Feathers & White, 1987* Hedlund, 1990 Hettich, 1990*
Jolley & Mitchell, 1990 Roth, 1985
Julian, 1989
Hynd & Chase, 1990*
Carswell, 1988
Bean, 1989; Manley-Casimir & Wassermann, 1989 Robinson-Armstrong, 1991
3


1985), reflective thinking (Anson & Beach, 1990; Hedlund, Furst, & Foley, 1989), a dialogue with oneself (Roth, 1985), and independent thinking (Robinson-Armstrong, 1991).
Figure 1.1 depicts a model of journal writing in academic environments. In the broadest sense, journal writing is intended to facilitate learning. But in a narrower sense, it has had three major purposes: (a) to facilitate text comprehension, (b) to develop thinking skills (e.g., metacognition, analytical thinking, critical thinking, and problem solving), (c) and to develop language/writing skills (e.g., essay writing and creative writing). (I use the term text in the broadest sense, meaning any information provided by the' instructional environment. It includes everything from a short story to a problem in math and steps in writing a research paper.) Because journal writing may serve the first two purposes without leading to the third, a dotted arrow connects parts B and C of the model.
Figure 1.1. A Model of Journal Writing
4


Unresolved research issues. Following is a list of the most important issues on
I
journal writing which research must address:
1. What are the salient attributes of journal writing? under what conditions? for what learners?
2. Does journal writing indeed facilitate learning? in all disciplines? in ill-defined as well as well-defined domains (e.g., Spiro, Coulson, Feltovich, & Anderson, 1988)? with all types of objectives (see Gagne, Briggs, & Wager, 1988)? with content of varying degrees of difficulty (e.g., Miyake & Norman, 1979)? with learners of varying characteristics (e.g., ability, motivation, cognitive style)?
3. Can effective journal-writing strategies be taught?
4. What type(s) of feedback/evaluation systems are appropriate? under what conditions? for what learners? Should journals be graded? If so, how often? with what criteria?
5. What specifically do learners find helpful about journal writing? Are learners perceptions of helpfulness tied to objectives? difficulty of material? evaluation procedures (including type of tests)? individual learner characteristics?
6. Is collaborative journal writing effective? under what conditions? for what learners?
Little if any research has been conducted on the cognitive foundation of journal writing as an instructional/leaming strategy. There is no qualitative or quantitative research basis for advocating one method of using journal writing over another. Is it the novelty of journal writing that attracts its proponents and seems to facilitate student performance? Or is there a yet-to-be identified cognitive basis for journal writing? If it
5


is the latter, wiiat is that basis, and does it involve a single strategy for all learners in all situations? Or does it consist of a range of strategies which interact with content of the materials presented, type of expected tests and other evaluation strategies, and individual learner differences? A rigorous investigation into the cognitive foundations
of journal writing is warranted.
i
Limitations of Prior Research
Journal Writing
Although journal-writing assignments already have a wide range of applications in higher education, tittle is known about their efficacy. Do they help students learn? But before we can ask that question, we must develop a description of journal-writing strategies used by learners and identify those which seem most successful, for whom, and under what circumstances. We must also identify the types of questions and comments students write in their journals; as Stein and Bransford (1979) observed, An emphasis oh the types of questions students ask themselves may.. .have important implications for understanding individual differences in learning and retention (p. 776). Having done these two tilings, we will have a foundation for developing instruction to guide learners in applying effective strategies in journal writing.
Learner-generated questions
' ^
In addition to its focus on journal writing, this study differs significantly from earlier investigations of learner-generated questions which focus on what a teacher wants a student to learn. For example, M. E. D. A. Andre and Anderson (1978-79) and Frase and Schwartz (1975) directed students to generate questions which they
6


believed would help them prepare for a comprehension test of expository passages; models of teacher questions were provided. This study focuses on the types of questions and comments students wish to ask for any reasonperhaps as much as any graded class assignment allows, each students goals and interests fuel journal writing. Furthermore, while most studies of learner-generated questions have focused on what Ng and Bereiier (1991) call extracting knowledge from text (p. 268; e.g., Van der Meij, 1992; Wong, 1985), this study focuses on comprehending literature.
Moreover, the studies have generally been empirical, conducted in artificial settings (e.g., laboratories and with little or no incentive) rather than in ecologically valid settings. Hicken, Sullivan, and Klein (1992) demonstrated the importance of adequate incentives in research on learning. The artificiality of the settings was also apparent in other ways which may have influenced the outcomes. For example, although researchers have instructed learners to generate questions, they often have not answered those questions (e.g., Miyake & Norman, 1979). Gamer (1990) found that context (e.g., artificial versus ecologically valid setting) plays a significant role in subjects use of learning strategies. Proponents of situated cognition (e.g., Collins, Brown, & Newman, 1989) would raise similar concerns about the validity of the setting.
My exploratory studies found that students questions and comments in required journals in literature classes do not merely reflect the types of questions they might expect on a quiz or examination, but often their questions and comments seek (a) to clarify facts, and allusions, or to verify low-level inferences which will help them construct the meaning of a story, (b) to relate the story to their lives, (c) or to evaluate it in terms of, for example, its artistic merit, its plausibility, or its subject.
7


Finally, this study differs significantly from earlier investigations of learner-generated questions by going beyond an information-processing cognitive model of learning, and acknowledging ways in which journal writing is consonant with the constructivist paradigm (see discussion below).
Justification of the Study
Using journals of adult students to investigate learner-generated questions and comments seems justified on many counts, even though this method is very time-consuming for both student and teacher:
1. The nature of complex short stories. Reading a complex reflective short story (as opposed to escapist literature) is a problem-solving activity which requires the reader to construct its meaning with few structural cues. The reader cannot engage in well known strategies for reading expository prose (such as reading chapter outlines, advance organizers, summaries, and headings and subheadings; skimming to identify topic sentences; and looking at adjunct questions). The differences between expository prose (which has been the subject of most of the research on reading, adjunct questions, and learner-generated questions) and fiction have not been adequately acknowledged in research. Meyer (1975,1977) provided convincing analyses of the structure of nonfictional prose. But even a simple short story does not obey the rules of nonfiction, and good fiction [the type taught in college literature classes] operates by dynamic, subtle processes (Bergstrom, 1983, p. 750). Even though most stories utilize common elements (such as setting, plot and character) and conventions (e.g., Culler, 1975; Rabinowitz, 1987), each
8


story combines them in unique ways, requiring readers to discover its specific structure as they construct its meaning.
2. Intuitive appeal of journals. Journals are an intuitively appealing teachingfleaming strategy. Journals (as well as diaries and notebooks) have been used by some of the greatest thinkers and writers throughout history (e.g., Pascal, Benjamin Franklin, Nathaniel Hawthorne, Henry Thoreau, Fyodor Dostoevsky, Bertrand Russell, Joseph Conrad, and Albert Einstein). The use of journals has not seemed to impair learning and, according to the most widely accepted theory of information processing (E. D. Gagnd, 1985), should facilitate learning by extending the limits of working memory and by encouraging learners to attend to the subject matter and elaborate upon it (this facilitation is also consonant with constructivism). Similarly, asking questions in journals should encourage students (1) to pause frequently while reading; and (2) to locate an appropriate answer to ah understanding question that probably is closely related to the study criteria (T. H. Anderson, 1980, p. 500). The constructivist paradigm would say that during frequent pauses learners are reflecting on what they have read, for example, trying to construct meaning by relating it to their prior knowledge.
Depth-of-processing theories also seem to support journal writing, whether the focus be Rosss (1981) importance of number of decisions, Jacoby, Craik and Beggs (1979) decision difficulty, or Johnson-Laird and colleagues (1978a, 1978b) amount of processing time. Furthermore, as Prawat (1989) notes, there is a growing body of literature suggestive of how writing can promote
9
I.


understanding of subject matter content (Britton, Burgess, Martin, McLeod, & Rosen, 1975; Tchudi & Tchudi, 1983) (p. 14).
3. Generative model/constructivist paradigm of instruction. Requiring students to write journals is in keeping with a generative model of instruction; it places much of the locus of control for learning in the student and encourages the student to interact with the text (see, e.g., Palincsar & Brown, 1984). Wittrock (1974) heralded the paradigmatic shift toward cognitivism, toward reinstating the learner
I
and his cognitive states and information processing strategies, as a primary
i
determiner of learning with understanding and long-term memory (p. 87; emphasis added).
Journal writing is also consonant with constructivist explanations of learning, which are rapidly replacing information-processing models (e.g., Clancey, 1992; Duffy & Jonassen, 1992; Jonassen, 1991a). The generative model of learning has instructional origins, while the constructivist paradigm has strong philosophical roots (personal communication, D. H. Jonassen, August 1991).
4. Individual-learner differences: ability. Journals allow the students to respond to a literary work at their own levels of understanding. When they initially respond to a work, they are neither overwhelmed nor bored by the questions typically found in literary texts which generally seem too sophisticated or too trivial for any given student. After the first reading of a work, high-ability students might be able to tackle some of the most sophisticated questions (e.g., What does the first-person point of view contribute to the meaning and effect of the story?), while low-ability students might ask questions which clarify the facts (What did Fortunate do to
10


deserve to be killed, in The Cask of Amontillado?), paraphrase segments, summarize the plot, or rely on elaboration strategies, trying to relate the content to familiar experiences. The cognitive structures, strategies, and working-memory capacity of the low-ability students might not be adequate to engage in higher-level processing during the earlier stages of study. The low-level processing activities that research suggests we will find with these students might be prerequisite to the development of deeper processing at later stages of study. At the least, asking or commenting at their individually appropriate levels (Miyake & Norman, 1979) might be expected to benefit students more than study alone (Frase & Schwartz, 1975).
5. Individual-learner differences: prior knowledge. Mivake and Normans (19791 study and other research on individual-learner differences suggest that the levels of questions in students journals will vary, depending on the expertise of the student Students with little general knowledge and/or little expertise in literature might tend to summarize, ask low-level questions (e.g., Why is death wearing white in The Appointment in Samarra? Did the servant know Miss Emily had
ir
killed Homer in A Rose for Emily?), or impose their value systems on the characters. Students with extensive general knowledge or an extensive background in literature might tend to have few questions on the easy works (and focus on the theme or respond with critical analyses), but to ask incisive questions about difficult works. Middle-range students might ask middle-level questions for easy to moderate works (e.g., Who is the you that Montresor is speaking to in The Cask of Amontillado?) but respond like novices on difficult stories.
11


6. Activating, building, and scaffolding the development of schemata. Journals provide a mechanism for the teacher (a) to identify problems individual students encounter in interpreting the text and (b) to provide feedback in a timely and appropriate manner (correcting the student if need be, answering questions and encouraging extensions of the students questions and comments). Class discussion is a totally inadequate vehicle for most students to communicate their understandings of the work under discussion. Providing a check on the students encodings of the stories is essential in terms of the generative learning model and schema theory, both of which stress that learners interpret new experiences in light of their prior knowledge (e.g., R. C. Anderson, Reynolds, Schallert, & Goetz, 1977; Jonassen, 1985; Mayer, 1975,1979,1980; Rumelhart & Ortony, 1977; Wittrock, 1974). Providing a check is even more critical in the constructivist paradigm of students learning: the internal neural processes [are] constantly changing and always freshly created (albeit it out of previous activations) (Clancey, 1992, p. 153).
Because students frequently lack adequate knowledge of allusions (as well as other general knowledge), they impose inaccurate schemata on them Journals provide insight into the misconceptions, as well as the exceptional insights, that students generate from their reading. The journals have provided me an opportunity to identify unique misconceptions as well as strong opinions and personal experiences that students otherwise would probably not express. In 20 years of teaching without journals, I rarely heard students express most of the questions and comments that they express in journals (e.g., Oedipus the King is
12


trash [because it deals with the subject of incest]. Why would the whole town [in William Faulkners A Rose for Emily] turn out for the funeral of a black woman? Until the very end, I thought Montresor [in Edgar Allan Poes The Cask of Amontillado] was just planning a big joke on Fortunato; I didnt really believe he was going to kill him. Sometimes students elaborate, connecting events in stories to their own lives, such as their own suicidal tendencies or even suicide attempts, or abortions).
Knowing students questions and comments allows a teacher to correct many misconceptions that a teacher, as an expert reader, is unable to predict less expert and novice students have. According to traditional schema theory, if the students have activated the wrong schemata, they will not be able to correctly assimilate the new information presented during class discussions. Spiro and his colleagues (e.g., Spiro, Coulson, Feltovich, & Anderson, 1988; Spiro & Jehng, 1990) concept of schema assembly offers a more dynamic explanation of schema theory for ill-structured domains (such as literature). Learners dont merely activate schemata, they create new schemata by combining relevant aspects of many schemata (this is consonant with Clanceys [1992] description, above). Knowing students misinterpretations also allows the teacher to emphasize important reading strategies, such as looking up words they do not know rather than guessing from context and looking up terms they are sure they know but which puzzle them in the story.
7. Learners dispositions. Research suggests that there are two types of learners: performance-oriented and mastery learners. Performance-oriented learners focus
13


on learning surface material, that is, facts or disconnected information; they are concerned with get[ting] the job done as painlessly as possible, with learning serving as a means to an end (Prawat, 1989, p. 33). Mastery learners use deep-level processing; they want to increase competence, to become more knowledgeable about or skillful at something (Prawat, 1989, p. 33). However, performance-oriented learners may become mastery learners if their study strategies facilitate learning (Prawat, 1989). In keeping with learners dispositions and schema theory, Shuell (1988) emphasizes that successful learners must play an active role in learning. Thus, by encouraging the students to interact with a literary work at their own levels, and by providing appropriate feedback, journal-writing should facilitate mastery learning.
8. Implications of research on adjunct questions. Research on adjunct questions has demonstrated the facilitative effects on learning of conceptual prequestions and frequent questioning during reading. (Research on learner-generated questions has not investigated these issues.) Furthermore, Rickards (1976) has theorized that conceptual prequestions induced readers to derive a relevant schema for passage information (p. 217). Since one might expect journal-writing to encourage at least mastery learners to generate and focus on conceptual prequestions and to ask frequent questions, journal writing should encourage them to derive a relevant schema for the story. Readers of a complex, reflective short story (perhaps even performance learners) might be expected to begin with at least one conceptual prequestion: What is the meaning of this story? Then one might predict that they will try to construct the meaning of the story as they read, asking questions and
14


generating self-explanations which will shed light on the overall question as they proceed. Unless they are performance-oriented, they will also ask other questions, such as, How do the parts achieve the meaning? Is the story effective? Would the story be more effective if various parts were changed? As the students ask themselves questions relevant to these issues, they might be more likely to notice situations where they need further clarification (Stein & Bransford, 1979, p.
775), for example, flagging points of confusion.
9. Depth of processing. Several explanations of depth of processing seem to be consonant with journal writing. Jacoby, Craik, and Beggs (1979) distinctivejness-of-encoding hypothesis seems to have bearing on journal writing. Requiring [students to write journals would seem to increase the difficulty of the task of reading/studying a literary work and thus result in higher levels of retention. The focus of Johnson-Laird and his colleagues (1978a, 1978b) on the amount of processing time also has bearing, since writing a journal about a literary work will require more processing time than merely reading the work. And Rosss (1981) focus on number of decisions seems relevant, since each journal entry will potentially reflect at least one decision about the work. Fisk and
Schneider (1984) hypothesized that [cjontrolled processing [the type most
i
learners would have to do in reading a reflective story] is characterized as effortful,
I
slow, serial, and capacity limited (p. 196). Writing a journal would seem to scaffold learners in focusing their attention and reducing the burden on working memory. :
15


As I have shown, prior research has failed to address several critical issues which suggestjthat journal writing might serve as an effective instructional/leaming tool in either an information-processing model or a constructivist paradigm of learning.
Next I summarize the specific questions which this study addressed.
Research Questions for a Study to Develop a Cognitive Model of Journal Writing in a College Introduction to Literature Course
Anecdotal reports often attest to the value of having students respond in journals to
works they read (e.g., Duke, 1982; Flitterman-King, 1988; Nugent & Nugent, 1985;
Weiner, 1986). My own instructional experiences with journal writing support such
reports. But no one has studied the cognitive basis of journal writing. In order to
optimize the benefits of journal writing assignments for diverse students, it is first
necessary to develop a cognitive model of journal writing as a learning and study
strategy.
Purpose of the Study
This study examined journal writing as a learning and study strategy of college students in the ecologically valid setting of an introduction to literature class. The study focused on students use of journals during the first half of the course, in the unit on complex reflective short stories. This study addressed several questions:
1. What metaeognitive and cognitive activities do students engage in when they write their journals?
f
2. How are these activities related to individual learner characteristics (e.g., age, reading comprehension level, prior knowledge of literature, students evaluations
i
of journal writing as a learning strategy, grade in the course)?
16


3. What types of questions and comments do students write in their journals?
4. How are questions and comments related to individual learner characteristics?
5. Can a taxonomy describe students questions and comments? Are there concepts or ways of seeing that students have in common?
6. How do students perceive journal writing (e.g., Do they find it helpful? Why or why not? What do they find most helpful and least helpful? What recommendations do they have for making journal writing a more helpful experience?)
7. How do ratings of journal helpfulness relate to other variables?
8. How do students rate the job aids (Students Accounts of Journal Writing and the Catechism for Stories)?
9. How do ratings of the job aids relate to other variables?
10. Is there a cognitive model that can capture the processing of most students journal-writing activity?
17


CHAPTER 2
REVIEW OF RELATED LITERATURE
This study focused on the use of journals in the study of fiction in a college-level introduction to literature course (an ecologically valid setting and an ill-defined domain). Although many teachers direct students to focus on gut reactions (Duke, 1982; McCormick, 1985; Nugent & Nugent, 1985), others emphasize cognitive processing, directing students responses with adjunct questions or soliciting learner-generated questions and comments (Flitterman-King, 1988). This study focused on cognitive activities. But to understand this focus, one must put it into the perspective of the earlier research on adjunct questions and self-explanations, schema theory, and the paradigmatic shift from mathemagenic instruction to generative learning.
Adjunct questions direct the learners processing of material; they precede, are embedded in, or follow text. A self-explanation is a comment generated in the process of learning something new^ going beyond what is stated. Self-explanations are generated in the context of learning something new, unlike elaborations, which generally refer to the use of existing knowledge to embed or embellish a piece of information in a larger context so that it is more memorable (Chi & VanLehn, 1991, pp. 71-72). Schema theory assumes that learners interpret new experiences in light of their prior knowledge. In mathemagenic instruction, the locus of control is external to the learner (e.g., in the teacher or text); the instruction seeks to control the informationprocessing activities of the learner (Jonassen, 1988), and thus to give birth to learning (Rothkopf, 1970). Generative learning places the locus of control in the
18


learner, who generates (constructs) meaning from new experiences in light of prior knowledge (e.g., Jonassen, 1988; Wittrock, 1974). Thus, research on journal writing must address the relevant literature on journal writing, adjunct questions, selfexplanations, schema theory, and the generative model of learning (and the constructivist paradigm). This chapter therefore addresses each in turn.
Journal Writing
In spite of the relative dearth of empirical support, almost without question educators have accepted journal writing as an effective tool for learning (Anson & Beach, 1990). A plethora of publications describe teachers experiences with journal writing, including, for example, their trials and errors (e.g., Heath, 1988; Holland, 1989), goals, strategies, and methods for incorporating them into courses (e.g., Bauso, 1988; Jolley & Mitchell, 1990; Myers, 1988; Roth, 1985), and effects on student engagement with assignments (e.g., Roth, 1985; Wilson, 1989). This approach is in keeping with the orientation of many humanities teachers and professors who follow a criticism model of scholarly inquiry.
However, advocates of journal writing too often describe their experiences as if they were generalizable instructional strategies. For example, Wilson (1989) reported that when 1 lth-grade literature students were instructed to reflect, ruminate, and question, to listen carefully to yourself and attempt to describe the effect the book is having on you and encouraged to respond deeply, write honestly in their journals, they generally (a) admitted confusion, (b) asked questions, (c) made inferences in trying to answer their questions, (c) often noticed words or phrases that the teacher
19


would not have pointed out, (d) identified with characters or the author, (e) made
l1
predictions, and (f) looked for evidence to support their opinions. Every literature teacher would surely welcome such behavior, but would these findings generalize to other 1 lth-grade students and teachers, to traditional and non traditional college students, and to other domains? What other factors (e.g., grading incentive and wording of feedback) contributed to the results?
Although Wilson (1989) provided insightful examples of the types of journal
I
entries students wrote, she did not study the relationships of important factors such as content difficulty and individual learning characteristics (particularly reading ability) and students entries.
In a descriptive article of journal writing in a college literature survey course,
i
Holland (1989) assumed that maintaining the anonymity of journal writers until the end of the term encourages students to take risks and to ask questions in their journals. His decision to assign anonymous journals was a response to disappointing results he obtained in previous terms when he had not graded signed journals (p. 236).
However, as he was aware when he decided on anonymity, his failure to provide any external incentive in his earlier use of journals may have contributed to those disappointing results (see Hicken, Sullivan, & Klein [1992] on the role of incentives). Since anonymity restricts the information that the teacher has to draw upon in responding to the needs of each student during the term, it is worthwhile to determine if it is a necessary criterion of productive journal writing. Can a supportive environment, combined with signed journals and an appropriate incentive, such as grading, produce the positive results he obtained?
20


With such a small body of research on journal writing, researchers are still struggling to identify the issues they must address in the design of their studies. Aware of the general lack of empirical support for the use of writing as a learning tool, Anson and Beach (1990) undertook a series of studies. They began by investigating a claim by Fulwiler (1989) that students who adopt a more informal, exploratory stance are more likely to formulate the course material in their own words, enhancing their understanding (Anson & Beach, 1990, p. 3). Thus they tried to characterize as precisely as possible those features which [they] could recognize across students academic journals (Anson & Beach, 1990, p. 3). They studied the relationship between the linguistic features of journals and performance in a college linguistics class. They rated students journal entries on six 4-point rating scales: formal/informal, objective/affective, elaborated/unelaborated, predetermined/unfolding, externally defined/intemally constructed knowledge, and negative/positive effect on the rater. The only rating tied to final grade was a relatively low correlation with degree of elaboration. Anson and Beach (1990) inferred that, regardless of their entering conceptions of the journal form, students appeared to make use of whatever features [were] personally most meaningful in a given context (p. 8).
Unfortunately, Anson and Beach made journal writing an artificial task (one not related to students or teachers needs), provided an inappropriate criterion for journal writing, and provided no guidance in journal writing. The only announced criterion was quantity of writing. Students received no feedback to what they had written; if students had wanted to ensure they understood the course material, they would have had to ask (apart from their journals). And since the journals were read only at the end
21


of the term, the instructor could not use journal content to monitor students understanding|of the material and adapt the instruction, a benefit many other writers noted (see below). Thus journal writing would probably have seemed an artificial task to many students. Unless their study strategies already included journal writing or they were mastery learners (see Glossary), students may have treated the assignment as busy work, with an implicit goal of wordiness.
Furthermore, Anson and Beach concluded that the kinds of thinking encouraged by journal writing had little to do with the students learning as measured on the objective tests (p. 8). Unlike Feathers and White (1987), they failed to consider that the content of students journals may not fully reflect the cognitive and metacognitive activities that they engage in.
Hynd and Chase (1990) studied the effects of annotating versus journal writing on multiple-choice, and essay test performance in college study skills classes. They found that students trained to annotate a novel significantly outperformed students who wrote journals on the multiple-choice test but not the essay test. However, their study had several limitations. First, as they noted, students were minimally competent in annotation or journal writing after training; motivation was low (the objectives of the experiment were not tied to the final examination which students were required to pass); and the task was probably too easy (students said the excerpt was not very demanding). Although they recommended that a more demanding passage be used in the future, they did not observe the greater benefit of using several passages with varying difficulty levels to study hot only task difficulty but also individual learner differences.
22


Equally important are limitations Hynd and Chase did not note. First, they did not include a control group. Second, although they determined that there were no statistically significant differences among the groups on basic skills, they did not study the relationship of performance to any individual learning characteristics (reading ability is particularly relevant in this instance). Most importantly, students in the groups were taught to focus on different criteria in studying the excerpt. The annotation groups were
taught to mark and annotate vocabulary, character development, important events, theme statements, use of literary devices like figurative language and symbolism, and to make inferences about relationships between ideas... because of the relation to the types of items commonly found on tests about novels that were given in the study strategies courses, (p. 47)
The journal groups were
taught to write journals exploring the environment for their reading, their vocabulary and comprehension difficulties, the strategies they used for overcoming these difficulties, what they thought was memorable about their reading, and their opinions about what they had read.. .because they are often required in classes where journals are used. (p. 47)
Class discussion addressed only the issues the students raised. Thus both groups did
not address the same objectives in their study strategies or in class. Nevertheless,
Hynd and Chase predicted that students who are taught to annotate as well as to write
journals would perform better overall. They did not consider the possibility that the
journal students would outperform the annotation students if they were taught to write
journals which helped them monitor their comprehension of the same elements
annotators were taught to attend to.
Even excellent research suggests avenues for further exploration. Feathers and White (1987) studied students development of metacognitive awareness in a college
23


developmental reading class. They conducted a case study of six students in a college developmental; reading class. Students entering the course typically did not take responsibility for comprehending a text, were unaware of the factors that affect comprehension, and had few, if any, reading or learning strategies (p. 266). The research suggested that students developed metacognitive awareness of reading and learning during the three-month course. The study is, of course, inherently limited by its focus on low-ability readers and its failure to examine any other individual learning characteristics.
Several writers report that journal writing enabled students to relate course content to their experiences (e.g., Hettich, 1990, college psychology; Roth, 1985, college sociology; Wilson, 1989, high school literature). Many have found that students perceived journal writing to be helpful (e.g., Cole, 1991, college literature; Cummins, 1989, college research writing; Hettich, 1990, college psychology; Rumpf, 1988, college engineering; Schwartz, 1989, college oral communication).
In a research study, Wason-Ellam (1987) found that first-grade students, when asked motivating questions about what they learned in mathematics, used their journals for self-questioning, organizing information, assimilating and accommodating information, arid making guesses (note the similarity to Wilsons description of the content of 1 lth-grade literature journals; see above).
Linn (1987) found that journal writing provided benefits for both high school geometry students and their teacher. Journal writing enhanced the students metacognitive ability, forcing them to become actively involved in learning, making them aware of What they did not know, and helping them identify and take advantage of
24


the strengths in their individual learning styles. Students journals provided a communication link with the teacher, who used it as a tool for monitoring students learning and adapting instruction. Similar teacher benefits have been identified by many others (e.g., Cole, 1991, college literature; Holland, 1989, college literature; Sublett, 1988, college geography; Tamove, 1988, college mass media).
Several writers report that students believed that journal writing affected their critical thinking skills (e.g., Bauso, 1988, college literature; Cole, 1991, college literature; Crismore, 1987, college basic writing). But with so little formal research on journal writing; alleged benefits must be verified, and research findings must be explored in other contexts to determine their transferability or generalizability.
Several additional issues are implicit in the literature. For example, what type(s) of grading/feedback systems are effective, for what purposes, and under what conditions? Some writers have advocated that journals not be graded; others (e.g., Holland, 1989, college literature; Roth, 1985, college sociology) have said they found grading a necessary incentive. Still others have identified graded journal writing as an incentive for students to read assignments and reflect on them before coming to class (e.g., Bauso, 1988, college literature; Schwartz, 1989, college oral communication) and a means of increasing students confidence in their understanding of the domain (e.g., Zuercher, 1989, professional writing). Advice on grading and feedback abound (e.g., Bauso, 1988, college literature; Myers, 1988, high school literature; Roth, 1985, college sociology), yet no one has provided empirical support for particular recommendations.
25


I
As noted above, Hollands (1989) article raised the issue of whether the teacher should know the journal-writers identity during the course. Equally important is another issue implicit in the literature: the effect of other students knowing the content of a journal. What effects do group journal writing (e.g., OSullivan, 1987, college literature) and having students share their journals with other students (e.g., Fulwiler, 1989, college literature; Heath, 1988; Wilson, 1989, high school literature) have on journal content, students beliefs about the helpfulness of journal writing, and individual achievement?
Another question implicit in the literature is What are the defining attributes of journal writing? What kinds of constraints can or should an instructor place on assignments which are called journal writing? At one extreme, some advocates allow the writer total freedom in determining content. Some (e.g., Gowen, 1985) advocate that the teacher never see the journals, only students responses to their own journals at the end of the term. At the other extreme, some treat journals essentially as vehicles to respond to adjunct questions (e.g., identifying previous journal entries on Blooms taxonomy; Hettich, 1990) or to write dialectical reasoning entries (Jolley & Mitchell, 1990). Are there differential benefits in these approaches (e.g., Do they interfere with high-ability students successful strategies)?
Finally, the literature provides no model of the cognitive and metacognitive activities involved in journal writing which is grounded in an analysis of specific journals, although several writers have provided cognitive explanations of journal writing (see Hedlund, Furst, & Foley, 1989, for a comprehensive explanation). Nor does the li terature provide taxonomies of journal content in many disciplines. Anson
26


and Beach (1990) developed a simple five-category taxonomy to describe the discursive characteristics of journals in a college linguistics class; they included a sixth category to classify the subjective responses of journal raters to the overall quality of the entry (p. 6). Hettich (1990) required his students to apply Bloom etal.s (1956) taxonomy to their earlier journal entries.
Adjunct Questions
The adjunct question literature suggests several relevant concerns for journal writing. For example, do findings related to question position, question level and task difficulty, and relationship of levels of questions to levels of processing also apply to learner-generated questions and self-explanations? Following is a brief review of the findings on adjunct questions.
Extensive quantitative research on the use of adjunct questions has culminated in an interest in learner-generated questions and self-explanations. A growing body of quantitative research on learner-generated questions and self-explanationsinfluenced by schema theory (R. C. Anderson, 1977; Rumelhart & Qrtony, 1977), the generative model of learning (Jonassen, 1985; Wittrock, 1974), and the constructivist paradigm (e.g., Clancey, 1992; Duffy & Jonassen, 1992; Jonassen, 1991a)supports their use as leaming/study aids (M. E. D. A. Andre & Anderson, 1978-79; Frase & Schwartz, 1975; Singer & Donlan, 1982). However, research has often been conducted in laboratories rather than more ecologically valid settings, and has focused on learning in well-defined domains. In the first part of this chapter I reviewed the literature on journal writing. In the remaining part, I review the literature on adjunct questions, self-
27


explanations, schema theory, and the generative model of learning (and the constructivist paradigm).
Question Position
The adjunct use of questions in text is the best established, most popular, and most heavily researched of the mathemagenic techniques (Jonassen, 1985, p. 128). Researchers have examined the effects of adjunct questions in terms of (a) question position (e.g., before text, after the complete text, after sections of the text, or adjacent to the text in the margin, (b) frequency, (c) response mode (e.g., multiple choice versus constructed response), (d) level of questionverbatim versus higher level, (e) relevance to criterion items, (f) nature and difficulty of the orienting task, (g) individual learner differences, such as aptitude, prior knowledge, and motivation, and (h) the interaction of those factors. The diversity of factors used in studies makes it difficult to draw consistent conclusions. For example, verbatim questions, which were used in many studies, typically yield different results than do higher-level questions, particularly when coupled with changes in position in the text A few studies are discussed to illustrate the complexity of the issue.
Generally, verbatim postquestions have been found to facilitate learning more than prequestions do (R. C. Anderson & Biddle, 1975; Lindner & Rickards, 1985); in fact, prequestions have been found more frequently [to] inhibit than facilitate performance on new criterion test items (R. C. Anderson & Biddle, 1975, p. 93). Rickards (1976) demonstrated the superiority of verbatim postquestions over verbatim prequestions on
i
recall tests. However, he found an interaction between level of question, question position, and amount of retention. On a test of immediate recall he found verbatim
28


postquestions and conceptual postquestions to be equally facilitative, but on a test of long-term retention, he found that conceptual prequestions produced greater verbatim recall than did conceptual postquestions. He theorized that the conceptual prequestions induced readers to derive a relevant schema for passage information (p. 217).
While Rickards (1976) used a recall test, Felker and Dapra (1975) used a multiple-choice test and obtained different results. They found conceptual postquestions superior to conceptual prequestions for learning to identify examples of concepts named and illustrated in the text One explanation of the discrepancy between their findings and those of Rickards is the difference in test mode. However, this seems unlikely, since an earlier study by Watts and Anderson (1971) also demonstrated that conceptual prequestions facilitate performance more than do conceptual postquestions, on a multiple-choice test of repeated questions as well as new application questions.
Question Level/Task Difficulty
Another explanation of the difference between the findings of Felker and Dapra (1975) and those of Watts and Anderson (1971) and Rickards (1976) may lie in the level of difficulty of the task at the time of encoding. Several studies have demonstrated that an increase in task difficulty results in greater levels of recall than do less difficult tasks; that is, tasks requiring more extensive processing in the semantic domain result in greater recall than tasks requiring less semantic processing (Glover, Plake, Roberts, Zimmer, & Palmere, 1981, p. 742; see also, Glover, Plake, &
Zimmer, 1982; Jacoby, Craik, & Begg, 1979). Rickards (1976) subjects were required to infer concepts, using prior knowledge, while Felker and Dapras (1975) subjects merely applied concepts that were named and illustrated in the experimental
29


passages. Since Rickards subjects were forced to draw on their prior knowledge for the appropriate concept, a related explanation is also implied by the Glover, Flake, Roberts, Zimmer, and Palmere (1981) study, which found that adjunct aids which require the readers schema to interact with the text base have faciMtative effects on recall.
This interpretation is supported by research on the consistency of organizers and text. Mannes and Kintsch (1987) found that consistency between organizers and text facilitated understanding and remembering, while inconsistency facilitated inference and problem solving. When the learners had to deal with inconsistency, they apparently had to draw on their prior knowledge to help them construct meaning. Salomon (1979) reported similar results in a study of randomly organized versus logically sequenced film.
Several other researchers have also demonstrated the superiority of higher-level questions. Benton, Glover, and Bruning (1983) noted that Andre has demonstrated that paraphrase questions are superior to verbatim questions in the learning of sentences (T. Andre & Sola, 1976), and in the learning of text materials (T. Andre, 1979; T. Andre & Womack, 1978) (p. 389). Glover, Flake, Roberts, Zimmer, and Palmere (1981) demonstrated a significant positive effect on free recall when questions required students to generate inferences from text. Friedman and Rickards (1981) found positive effects increasing as the questions progressed from low to higher levels. On a test of both direct and indirect learning, paraphrase questions were superior to verbatim questions, and inference questions were superior to paraphrase questions.
30


Additional Findings on Adjunct Questions
The apparent interaction of test type, task difficulty, learners prior knowledge, and other factors indicates the need for great caution in interpreting findings for future research. Following is a summary of the more consistent findings on the use of adjunct questions (in addition to the studies cited above, see R. C. Anderson & Biddle, 1975, and Lindner & Rickards, 1985).
The closer the adjunct question is to information it asks about, the higher the test performance is on similar criterion items.
The effects are greater for short-answer tests than for multiple-choice tests, on both new and repeated criterion items. One explanation for difference in effects is that adjunct questions primarily act on the retrievability of information... .Another possibility is that short-answer and multiple-choice questions make different processing demands when inserted in text, thereby differentially affecting study activities (R. C. Anderson & Biddle, 1975, p. 98).
Compared to verbatim questions, higher-level questions facilitate greater recall of (a) both new and repeated criterion items and (b) both factual and higher-level material.
[H]igher-level questions [particularly prequestions, as noted by Rickards, 1976] affect not only the level at which material is processed but the manner in which such material is organized in memory (Lindner & Rickards, 1985, p. 139). Rickards (1976) suggests that conceptual prequestions induce readers to derive relevant schemata.
Facilitative effects are a function of relevancy of questions to the material (McGraw & Grotelueschen, 1972, and Rothkopf, 1972, as reported in R. C. Anderson, &
31


Biddle, 1975); Stein and Bransford (1979) suggest this effect is related to readers asking about the potential significance or relevance of facts and to readers noticing situations where they need further clarification (p. 775).
Would these findings on a mathemagenic instructional strategy apply in a constructivist learning environmentthat is, when the learner controls the topic and level of difficulty of his/her questions and self-explanations? Before examining the limitations of these studies, I will briefly describe the research on self-explanations, which are essentially answers to implied questions. Then I will examine the limitations of both strands of research for journal writing.
Self-Explanations
Chi and her colleagues (e.g., Chi & Bassock, 1989; Chi, Bassock, Lewis, Reimann, & Glaser, 1989; Chi, de Leeuw, Chiu, & LaVancher, 1991; Chi &
VanLehn, 1991) studied think-aloud protocols of students generating self-explanations of physics examples. They found that the amount learned was related to the number of self-explanations generated (Chi, Bassok, Lewis, Reimann, & Glaser, 1989). They also found that self-explanations help students construct more complete understandings of the domain principles and concepts, that much of what students learned related to technical procedures for solving physics problems, and that good problem-solvers generated more self-explanations than poor solvers (Chi & VanLehn, 1991). Chi, de Leeuw, Chiu, and LaVancher (1991) found that generating self-explanations benefits both average and high-ability eighth-grade students equally (there were no low-ability students in the study).
32


Limitations of Research on Adjunct Questions and Self-Explanations
Although research on adjunct questions has provided much insight into the topic, it has largely ignored the mediating role of individual learner characteristics (Lindner & Rickards, 1985). Furthermore, it has focused on low-level questions, multiple-choice tests, and what Lindner and Rickards (1985) term the artificiality of the adjunct question paradigm (p. 145) because the limitations placed on the learner in the adjunct question studies in no way resemble the normal reading and/or study behavior of students (p. 14b). The artificiality of the studies seems particularly obvious when one observes students in natural conditions employing the suggested strategy of skimming material before they read it or performance-oriented students merely looking ahead to see what questions they must answer, for both types of students postquestions immediately become prequestions. Moreover, the research on adjunct questions has generally ignored the learning strategies students engage in with others (e.g., other students, teacher, or family) when they are studying.
Stein and Bransford (1979) criticized the focus of adjunct-question research on facts, rather than on understanding the potential significance or relevance of facts (p. 775). Their criticism is related to criticism of instruction for its focus on facts rather than meaningful learning (e.g., R. C. Anderson, 1977; Bransford, Sherwood, & Hasselbring, 1988; Perkins, 1986). Thus, many issues related to adjunct questions remain to be explored.
Although several studies claim to have demonstrated the superiority of higher-level questions over verbatim questions in memorability, Jonassen (1985) highlighted our
33


inability to define [adjunct] questions at various levels (p. 128). Lindner and Rickards
(1985) said that the cognitive processing level is
[pjerhaps the most important issue related to adjunct question research and practice... .What level of processing is necessary to produce knowledge or understanding of passage information?
.. .Little or no research has compared various levels of meaningful processing, whereas it is such comparisons that are educationally relevant (p. 138)
Unlike the research on adjunct questions, the research on self-explanations has
directly addressed several critical issues, including the mediating role of individual
learner characteristics and learner-controlled levels of cognitive processing. The studies
on self-explanations had two limitations particularly critical for using journal writing as
a strategy for comprehending reflective short stories. First, ecological validity was
limited by the use of think-aloud protocols. Second, solving physics problems is a
well-defined domain, one that is very procedurally oriented, while comprehending
reflective short stories is an ill-defined domain.
Schema Theory
Relationship of Levels of Questions to Levels of Processing
Many researchers have discussed levels of questions (e.g., factual, paraphrase, conceptual, problem-solving) in terms of depth of thinking or depth of processing (see chapter 1 and below). Are these terms synonymous and to what cognitive activities do they refer? Researchers and theorists have variously defined these activities in terms of amount of processing time, number of decisions, type of mental activity (such as rehearsing, elaborating, paraphrasing, drawing inferences, evaluating, problemsolving), some other factor, or an interaction between several factors.
34


Schema theory (e.g., R. C. Anderson, 1977; Rumelhart & Ortony, 1977) provides a further connection between levels of questions/thinking and levels of processing. A schema is a structured cluster of knowledge which is embedded within a semantic network. Rumelhart and Ortony (1977) explained that the process of activating related schemata is akin to Craik and Lockharts (1972) notion of depth of processing (p.
105). At deeper levels the subject can make more use of learned cognitive structures so that the item will become more complex and semantic (Craik & Lockhart, 1972, p.
679). Recent discussions of schemata emphasize that they are not stored mental
|
structures but dynamic networks (e.g., Clancey, 1992).
Definitions of Levels of Processing
Craik and Lockhart (1972) conceived of depth of processing in terms of elaboration, which often means the addition of further information, so that the trace becomes richer and more detailed (Jacoby, Craik, & Begg, 1979, p. 597). Craik and Tulving (1975) introduced the notion of distinctiveness of encoding, which emphasizes the contrastive value of information in the trace (Jacoby, Craik, & Begg, 1979, p. 597). The distinctiveness-of-encoding hypothesis suggests that memorability is a function of the difficulty of decisions the learner must engage in; that is, an increase in task difficulty.. .results in higher levels of retention (Jacoby, Craik, and Begg, 1979, p. 586). As reported by Ross (1981), Johnson-Laird and his colleagues (1978a, 1978b), have focused on amount of processing time, while Ross (1981) has focused on the number of decisions involved.
Although several explanations of depth-of-processing have been researched, there remains a need to investigate the potential interactions of those and other factors.
35


According to traditional schema theory (e.g., R. C, Anderson, 1977; Rumelhart & Ortony, 1977), when learners encounter new information, they will potentially activate existing schemata, try to fit the new material into the old schemata, refine existing schemata and/or build new schemata. Thus these cognitive activities would potentially seem to (a) involve differentiating the new information from elements in existing schemata (distinctiveness of encoding), (b) result in many decisions, (c) be quite difficult, (d) require much processing time, and (e) perhaps result in elaborations which make the new material more memorable. (R. C. Anderson [1977] hypothesized that when peoples beliefs are threatened, they may segregate and retain logically inconsistent views in order to protect the ones they believe.) Which of these cognitive activities occur(s) might depend on the particular infonnation to be learned, the criticality of the information/task (a factor apparently not yet studied), prior knowledge, and other individual differences. These factors are, of course, difficult to isolate, but may be reflected in the comments and questions in students journals.
The Generative Learning Model
The importance of journal writing as a learning and study strategy is also derived from a generative model of learning. The emerging interest in learner-generated questions and self-explanations has followed closely on the heels of the paradigmatic
I
shift in psychology from behaviorism to constructivism (Bruiting, 1983; Jonassen, 1991a; Resnick, 1983) and the accompanying shift from mathemagenie instruction, which seek[s] to control the information processing activities of the learner
(Jonassen, 1985, p. 127), to generative learning, which places the locus of control in
36


the learner. The generative learning model reflects what many educators and researchers view as a desirable (perhaps the primary) aim of instruction: to encourage learners to become independent by learning how to learn (e.g., Brown, Campione, & Day, 1981; Frase & Schwartz, 1975; R. M. Gagnd,1980; Jonassen, 1985).
The generative model asserts that learners, when faced with stimuli.. ..construct and assign meaning to that information based upon prior learning (Jonassen, 1985, p. 11). The emphasis on the active role of the learner has bearing on many current focuses in instructional theory and research, from problem-solving and schema theory, which explains comprehension in terms of the interaction between the reader and the text (R. C. Anderson, 1977; Rumelhart & Grtony, 1977), to the views of transactional analysis which is applied so well to the study of literature by Rosenblatt (1978).
Learners Dispositions and Goals
The generative model also has bearing on the emerging interest in the dispositions of learners and their learning goals. Prawat (1989) described two types of learners: performance-oriented learners and mastery learners (see Glossary). According to Prawat, when performance-oriented learners become meaningfully engaged in learning and can make the necessary connections between the elements under study, they may become mastery learners; on the other hand, if mastery learners find a strategy ineffective, they may adopt lower-level strategies typical of performance-oriented learners. (For related discussion of surface versus deep-level processing see Marton & Saljo, 1976a, 1976b, and Watkins, 1983.)
Bereiter and his colleagues (e.g., Bereiter & Scardamalia, 1989; Ng & Bereiter, 1991) have focused on the role of learners goals. Bereiter and Scardamalia (1989)
37


described two! different learning goals in public schools. Learners with work orientations, the goals that schools typically cultivate, view schoolwork merely as tasks to be completed (ends in themselves), with learning being incidental. Learners who engaged in intentional learning view schoolwork as a means to an end; they pursue cognitive goals over and above the requirements of tasks (p. 385).
Ng and Bereiter (1991) examined the relationship between motivation and learning-
goal orientation. They identified three levels of goal orientation among adult learners
who were presumably equally motivated: task-completion goals, instructional goals,
and personal-knowledge-building goals. The learners with task-completion goals
focused on completion of assigned tasks, while those with instructional goals focused
on the content of instruction. By contrast, the personal-knowledge builders, in addition
to performing better on a posttest, responded more often [than the others] to learning
goal cues than to task goal cues. They actively related new learning to prior knowledge
and they posed and tried to solve problems and questions (p. 243). While goal
orientation and achievement were not related to the students levels of education and
prior computer experience, they were positively related to previous experience of
independent learning (p. 243). However, rather than focusing on strategies which
will promote independent learning, Ng and Bereiter concluded:
If there are educational approaches that target [personal-knowledge building], the most likely candidates are ones that operate at a social rather than an individual levelones that promote a community of learners (Brown & Campione, 1990) or a knowledge-building community (Scardamalia & Bereiter, in press), (p. 269)
Thus, motivation, goals, strategies, and their interactions remain a concern of both
i
researchers and instructional developers.
38


Focus of Research on Leamer-generatedDnesrims
The generative counterpart to adjunct questions [which are mathemagenic aids] requires learners to generate their own questions (Jonassen, 1985, p. 27). Thus, all of the issues related to adjunct questions logically extend to learner-generated questions; yet by the very nature of the generative model, these issues cannot be as readily investigated. Even so, there has been relatively little research on learner-generated questions, particularly related to the study of fiction, drama, and poetry. The focus of most research on questioning (e.g., Van der Meij, 1992; Wong, 1985) is not on understanding literature, but on what Ng and Bereiter (1991) call extracting knowledge from text (p. 268).
Learner-generated Ouestions_as_S_tudy Aids
Adjunct versus learner-generated questions. In developing a model for studying of nonfictional works, T. H. Anderson (1980) conducted an informal study of graduate students engaged in serious study (p. 495). He used varied methods with eight students, such as observation; interviews; having learners place question marks in the margins of their texts; and having learners read aloud, predict content of paragraphs, summarize out loud, relate other thoughts as they read. He concluded that studentgenerated questions and adjunct questions seem to be effective study aids for the same reasons: They encourage students (1) to pause frequently while reading; and (2) to locate an appropriate answer to an "understanding question that probably is closely related to the study criteria (p. 500). However, Frase and Schwartz (1975) found that unless encouraged to ask difficult questions, students tended to generate only verbatim recall questions. The different findings might be a function of student maturity and/or
39


motivation. Andersons subjects were graduate students in a natural situation, while Frase and Schwartzs were high school students participating in an experiment.
Asking or answering questions versus merely studying. Frase and Schwartz (1975) demonstrated the superiority of asking or answering questions over merely studying; students working in pairs who asked or answered each others questions while studying biographical prose passages performed better on a recall test than students who merely studied (students were instructed to ask questions which would help them on a posttest). However, significant effects were limited to content that was directly related to subjects questions (p. 628). The study controlled the number of questions asked (5 or 10) and to some extent the difficulty level of questions asked
The role of prior knowledge. Miyake and Norman (1979) investigated the notion that a prerequisite for asking questions about a new topic matter is some appropriate level of knowledge (p. 357). They found an interaction between prior knowledge and number of questions asked; novices asked more questions on easy material; trained subjects asked more questions on harder material. Although this finding sounds logical, the study had limitations which relate to ecological validity. First, it was set in a research laboratory, with no meaningful incentive for learning (subjects were undergraduates in introductory psychology courses and received course credit or pay merely for participating). As noted in chapter 1, incentives contribute to subjects performances in research (HiCken, Sullivan, & Klein, 1992). Second, learners questions were not answered, even though Miyake and Norman acknowledged the social context of questions: To ask a question of someone implies more than a need for information (p. 357; emphasis added). (This criticism does not imply that learners
40


might not meaningfully ask themselves questions as they read. Indeed, that is one of the functions this study attributes to journal writing. However, the reason for not answering was clearly an artificial one in the learning context: to keep the amount of information available the same from subject to subject (p. 359).
Relationship to ability level. In a study of high school psychology students M. E. D. A. Andre and Anderson (1978-79) found that self-generation of questions during study can lead to improved performance on a test of comprehension (p. 619); appropriate questions were modeled for the students. However, training mainly benefited low and middle-ability students; Andre and Anderson concluded that high-ability students apparently already knew how to generate good comprehension questions; that is they were able to generate questions (on three passages related to psychological concepts, 450 words each) that facilitated good performance on a short (20-item) test; 10 items required application of concepts and/or principles to new examples; 10 detail questions examined which, who, when, where, what or how (p. 612).
While research has generally found differential benefits of strategy training favoring low- and middle-ability students, Chi, de Leeuw, Chiu, and LaVancher (1991) found that prompting eighth-graders to generate self-explanations as a tool in learning physics concepts benefited average-and high-ability students equally.
Similarly, in my preliminary studies of journal writing (Cole, 1991) students of all ability levels (as determined by reading comprehension and grades) thought they benefited from journal writing. A survey (Appendix C) of students in five of my literature classes revealed that students of all levels said they found journal writing
41


1'
helpful, with 84.6% of the A students and 60% of the B students rating it Very Helpful, the top rating on the survey (with 1 being Not at All Helpful, 3 being Somewhat Helpful, and 5 being Very Helpful); 77% of all respondents gave journal writing a rating of 4 or 5. The only respondent who was earning an F at the time of the survey (just before the final exam) also rated journal writing as Very Helpful. In their survey comments and in class comments, several of the A and B students spontaneously identified self-generated questions as the most helpful aspect of writing journals. One of the A students, who had written journals in two of my courses (Introduction to Literature and Masterpieces of World Literature Since the Enlightenment) wrote that journal writing was the most important thing she had learned in eight semesters as a part-time student. Another A student, who transferred to Amherst College more than a year after she had written journals in the same two literature courses, wrote from Amherst that, when she found herself having difficulty with a philosophy of religion course, she decided to write journals and they solved her problems in learning the material. While these students may have been responding in ways they thought would please the teacher, this seems unlikely, since their typical behavior did not suggest such a motivation. More importantly, students perceptions are one of many sources of questions worthy of research.
Learner-generated versus proposed teacher questions. Singer and Donlan (1982) demonstrated the effectiveness of learner-generated questions compared to prequestions posed by a teacher. High school students were taught to combine a problem-solving schema for comprehending short stories with the construction of schema-general questions for each element in six complex short stories. However, there were several
42


limitations to this study. All students read the story at the same rate (listening to a recording as they read from a copy); thus they could not read at their own pace, look up unknown words, or re-read. Essentially all questions were generated at a given point in the story (the recording was stopped and students were asked to write three questions they would like answered as the story progressed); students were allowed to ask additional questions at the conclusion of the story, but few did. Thus the study paralleled the adjunct-question format of single-position (near the beginning), grouped questions. The authors did provide a list of story-general questions, which vary in difficulty, and examples of three story-specific questions used in training the students to ask questioris; however, there is no way to evaluate the level of students questions, since the authors did not provide any examples. The measure of learning was limited to a 10-item multiple-choice test on each story; nor did the authors itemize to see whether previously taught story elements resulted in improved mastery on subsequent tests. Finally, the authors taught the students to generate questions based on a very simple story schema whose elements were leading character, goal, obstacles, outcome and theme. They added theme to an existing schema for vignettes and fables because they felt that it was an important element for complex stories. However, after they found facilitative effects for learner-generated questions for the other part of their schema but not for theme, they concluded, in retrospect we should not have included theme in our instruction because it is not a compelling aspect of story comprehension (p. 182). Yet theme is the all-embracing element of complex stories (apart from escapist literature). Nor did their schema acknowledge elements that typically contribute to the theme directly or indirectly; for example, setting, symbols, point of view, and irony. (For
43
nrmni


further discussion of this issue in terms of the proposed study, see the Procedure Section.)
Summary of Learner-generated Questions
The above studies, then, suggest that learner-generated questions can be an effective study strategy. In spite of their limitations, the Miyake and Norman (1979) study and the Singer and Donlan (1982) study suggest interesting avenues to explore, particularly in terms of (a) a more ecologically valid study, and (b) the interaction of college-age learners with complex short stories as they attempt to construct meaning.
Implications of Schema Theory. The relevance of the above questions to both cognitive theory and instruction is supported by the generative model of learning (and the constructivist paradigm) and schema theory. For example, Wittrock (1974) observed that people tend to generate perceptions and meanings that are consistent with their prior learning (p. 88). R. C. Anderson, Reynolds, Schallert, and Goetz (1977) similarly observed:
From the perspective of schema theory, the principal determinant of the knowledge a person can acquire from reading is the knowledge s/he already possesses. The schemata by which people attempt to assimilate text will surely vary according to age, subculture, experience, education, interests, and belief systems, (p. 378)
But they cautioned that dominant high-level schemata are often imposed on text even
i
when, according to a third party point of view, some violence is done to the data5 contained in the text (R. C. Anderson, Reynolds, Schallert, & Goetz, 1977, p. 371), and they concluded that It may turn out that many problems in reading comprehension are traceable to deficits in knowledge rather than deficits in linguistic skill narrowly conceived... (p. 378).
44


Thus, it seems imperative that educators learn more about how students actually comprehend textcreative text as well as the more extensively researched expository textso that they can facilitate students construction of its meaning.
Summary
In this chapter I have discussed the literature on journal writing, adjunct questions, self-explanations, schema theory, and the generative model of learning. In many courses, journal writing supports other writing activities. But in literature courses, it is typically used to facilitate comprehension and to communicate problems and understandings to the teacher. I have shown that most of the issues related to journal writing as an instructional/leaming tool in college literature courses (and other disciplines) have not been addressed adequately, if at all, by researchers. Writing journals to facilitate comprehension of reflective short stories (a very ill-defined domain) may involve unique cognitive and metacognitive strategies. The literature provides no research-based model of these activities. Nor does it provide a suitable taxonomy for classifying the content of journals on short stories.
I presented literature on adjunct questions, self-explanations, schema theory, and the generative model of learning because as students read reflective short stories, they engage in a dynamic cognitive process which relates to all of these. They monitor their comprehension; generate questions and self-explanations (answers to their explicit or implicit questions); and activate, refine, or create schemata.
In the next chapter, I describe the methods I used to study college students use of journal writing in comprehending reflective short stories.
45


CHAPTERS
METHOD
Conceptual Framework
I1 used predominantly qualitative research methods in the study, employing the conceptual framework reflected in Figure 3.1. Although the figure reflects a generally linear process, beginning with theory and moving clockwise, the driving force is the question focus. Thus, a change in the focus of the question(s) may force the researcher into an iterative process, particularly in developing tools, indices and instruments; in observation and data collection; and in data analysis. (Ultimately I hope to explore the topic of journal writing, as defined in this study, in a quantitative study; for example, assessing the effectiveness of journal writing as a learning strategy in literature courses, examining whether learners can be taught to improve the effectiveness of their journal writing while studying literature, and examining the transferability of the strategy to other courses.)
Qualitative research has become accepted as a legitimate alternative to quantitative research, not merely an adjunct (Eisner & Peshkin, 1990). Donald Campbell and Lee Cronbach, who were considered major spokesmen for quantitative research in the past, *
*1 am using a predominantly subjective rather than an objective stance in chapters 3,4 and S to emphasize the role of the researcher as a smart instrument (Guba & Lincoln, 1982, p. 240) in qualitative research, or as Guba and Lincoln prefer, naturalistic research. They emphasize that, rather than trying to overcome the inquirer-respondent interactivity, the naturalist exploits it (p. 240).
46


have both advocated the usefulness of qualitative research (Eisner & Peshkin, 1990; Patton, 1980).
Figure 3.1. Conceptual Framework of the Study*
Educational technology is following the trend toward greater flexibility in research methods. Salomon and Gardner (1986) advocated that computer researchers should .. .utilize holistic as well as standard experimental research paradigms, particularly during the early phases of research (p. 13). Even a cursory examination of studies published in the Journal of Experimental Psychology (e.g., Stadler, 1989; Willingham,
47


Nissen, & Bullemer, 1989) reveals the incorporation of qualitative methods such as self-reports iniexperimental research, in an almost matter-of-fact way. The Association of Educational Communications and Technology (the primary professional organization of education teachers) provides a special award for a qualitative research investigation in the field of educational communications and technology to encourage doctoral candidates to engage in qualitative research. Several journals are also encouraging submission of qualitative research articles (e.g., Educational Technology publishes juried articles on qualitative research).
Patton (1980) provided guidance in determining if a qualitative study is justified; he lists 16 appropriate situations, 5 of which relate in varying degrees to the research issue at hand:
the programs emphasize individualized outcomes,
there is a need for detailed, in-depth information.. .about certain client cases,
there is an interest in collection of detailed, descriptive information about the program for the purpose of improving the program,
no valid, reliable, and believable standardized instrument is available or readily capable of being developed to measure the particular program outcomes for which data are needed,
there is a need to personalize the evaluation process by using research methods that require personal, face-to-face contact with the programmethods that may be perceived as humanistic and personal because participants are not preordinately labeled and numbered, and methods that feel natural, informal, and understandable to participants (pp. 88-89; emphasis in original).
48


However,1 researchers must tailor the designs of their studies to their needs (see, for example, Clark, 1989; Patton, 1980). Thus, I used a mixed methodology, quantifying data whenever appropriate while generally adopting the stance of a phenomenologistconcerned with understanding human behavior from the actors own frame of reference (Bogdan & Taylor, 1975, p. 2, as cited in Patton, 1980, p.
45; emphasis in Patton). I utilized student self-reports, surveys, interviews, content of journals, and students oral comments during class discussions, during office hours, and in spontaneous conversations before and after class.
The need for ecologically valid methods in studying journal writing has a parallel in
the adjunct question literature (E. D. Gagnd, Broughton, Eggleston, Holmes, Hawkins
& Sheldon, 1979). Lindner and Rickards (1985) state:
It is readily apparent [from a review of the literature] that the limitations placed on the learner in adjunct question studies in no way resemble the normal reading and/or study behaviors of students. Subjects can neither re-read nor underline nor take notes in the typical experiment, nor are they allowed to review text passages once they have a question. Generalizations to real-life settings from the results of adjunct question experiments are, consequently, rather limited, (p. 146)
Although Lindner and Rickards were referring to adjunct questions, the studies of leamer-generatgd questions (see chapter 2) suggest that ecological validity is of equal concern to that area of investigation. While there has been much naturalistic research on the process of writing (e.g., Emig, 1971; Graves & Murray, 1980; Hayes &
Flower, 1980), there has been little if any naturalistic research on the strategies students use in the comprehension of fiction, especially relating to learner-generated questions and self-explanations. As discussed in chapter 2, T. H, Anderson (1980) conducted an informal study of graduate students engaged in serious study of nonfictional works
49


(p. 495); he usied varied methods with eight students, such as observation; interviews; having learners place question marks in the margins of their texts; and having learners read aloud, predict content of paragraphs, summarize out loud, and relate other thoughts as they read. I have used some of his methods in the present study. For example, instead of placing questions marks in their texts, journal directions asked students to:
record questions you have about an assigned story (any questions whose answers you believe will help you understand the story), and
ask questions during class discussion.
Students could also
summarize the story in writing, and
elaborate on the story in writing and during class discussion.
For most of the students, surveys replaced Andersons interview method; I also interviewed five students to obtain more information about their journal-writing strategies.2 Because of the nature of the reading-writing assignments, I did not observe students during actual reading and writing.
Background and Setting
Background of the Researcher
When I initiated the study, I had taught college English and literature full time for 25 years, the last 23 of which were at Arapahoe Community College, a public community college in a large metropolitan area. I have a BA and an MA in English,
2See below,and chapter 4 for an explanation of the criteria used to select students for interviews.
50


and a BA in Psychology; I have done additional graduate work in English and Education and am a doctoral candidate in Instructional Technology at the University of Colorado, Denver Center. My special focuses are cognition and instruction and instructional design.
Setting for the Study
I conducted preliminary studies of journal writing in literature classes for three semesters at Arapahoe Community College (ACC). I conducted the current study during the 15-week spring semester, 1992. Located in a white-collar suburb, the college enrolled 7,614 students spring 1992; full-time-equivalent enrollment was 3,744.3 The older students are generally more motivated and higher achievers than the recent high school graduates.4 Table 3.1 compares the college population to Group 1, the primary focus of this study.
At the college, an English/literature instructor teaches five classes with a maximum of 23 students per semester. Technological support is traditional (chalk board, overhead projector, film projector, slide projector, tape recorder, and limited access to VCR). Although the college had extensive computer resources for its occupational-vocational programs, there was essentially no computer support for the academic program until the middle of the semester during the current study.
3FuIl-time equivalency is based on a 12-credit load.
^This conclusion is based on my own observations and on conversations with faculty in many other disciplines. We base our conclusions on issues such as students preparedness (e.g., Do they come to class prepared? Do they submit assignments on time? Do they learn what is expected or continually challenge requirements or make excuses?), their participation in class discussions, and in-depth rather than surface learning.
51


Table 3.1. Comparison of ACC Population and Group 1
ACC Group 1
Number Enrobed 7.614 15
Average Age 31.2 28
Average # Credits
i Spring 1992 7.4 8.4
Employment
Full time 44% 67%
Part-time 18% 26%*
Unemployed 19% 0%
Unknown 19% 7%
Marital Status
Married 42% 40%
Single 56% 60%
Unknown , 2% 0%
Sex
Female 59% 80%
Male 40% 20%
Unknown .004%
* The minimum number of hours employed was 18.
Within this setting, I have used journal assignments as an inexpensive strategy to respond to the needs of individual students; although journal writing is time-consuming for both students and teacher,5 it does not require expensive technological support. For the previous seven years I required students in my literature classes (and most of my composition classes) to write journals on most of their reading assignments. The primary intent has been to engage the student with the work as much as possible within the constraints of a traditional college learning environment; that is, an instructional environment with high student-faculty ratios and little technological support. 10
5Although no surveys addressed the amount of time spent, some students reported spending up to
10 hours on journal assignments. I spent as little as five minutes grading short, superficial journals, and up to an hour grading long, probing journals. I probably averaged 10 to 15 minutes per short-story journal.
52


Subiects/Resoondents6
In choosing the subjects, I used purposeful sampling of typical cases, as described by Patton (1980); Berg (1989) uses the term purposive. The subjects of the study were students in two of my classes at the urban public community college where I teach. All students were enrolled in freshman-level Introduction to Literature, a 3-credit semester course in the study of fiction, drama and poetry, meeting 150 minutes a week. There were 157 students in the primary section (Group 1), which met from 5:30 to 6:45 p.m., Mondays and Wednesdays. I selected this class because it was the largest one I had at the time of the study8 and was fully constituted on the first day of class.9 10
To assess typicality, I included a second section of Introduction to Literature with 10 students,10 which met from noon to 12:50 p.m., Mondays, Wednesdays, and
6Guba and Lincoln (1982) prefer the term respondents because they believe it more accurately describes the interactive relationship which characterizes qualitative (naturalistic research). They believe the term subjects is more suited to experimental (rationalistic) inquiry. To reflect the ecological validity of my study, I generally use the term students.
^Sixteen students enrolled, but one had to drop after the first few weeks for financial reasons.
Only 14 students completed the two journals which were the primary focus of the journal analysis.
8The small sizes of my classes was an unfortunate circumstance, since it affected the statistical significance of many of the correlations between factors of interest. Normally, my classes enroll at least 20 students, but this term started with 16,10, and 8. Although I normally teach five classes, I taught only three at the time of the study; I was released from two classes to work on other projects.
9My second largest class, which I included as Group 2, was constituted from the wait list of another class. Unfortunately, the instructor did not immediately send all the students to the new section. Thus, 50% of the students missed up to three days of the first week of class, including many of the class orientation activities.
10I had planned to include a third section, a sophomore-level course in world literature. Although I did assign students to write journals, tested their reading levels and field dependence/independence, and gave them the same surveys I gave the other two classes, the class was too small to warrant inclusion in the study.
53


Fridays. Students in college-level classes at the college typically represent a broad range of ages (18-45 or older, sometimes retired senior citizens), abilities and backgrounds (e.g., veterans, housewives entering college for the first time or returning to college after 10-20 years, students who have taken developmental composition or developmental reading, and outstanding students who ultimately earn full scholarships to distinguished colleges such as Amherst, St. Johns College, Georgetown University, Ripon College, and Colorado College).
At least 75% of the students in the Introduction to Literature course take it to satisfy part of the Humanities requirement; other students take the course for general-elective credit or for personal enrichment (occasionally senior citizens, most of whom already have bachelors or masters degrees, take the courses for enrichment). Although students are advised to have college-level reading skills before enrolling in college-level courses, some students disregard the advice. During fall semester 1990,15% of the students in my two sections of Introduction to Literature entered with reading skills below the 9th-grade level; 25-30% read at the 10th- or llth-grade level. However, that distribution was an extreme one. During the current study, only 14% (2) of the 14 students in Group 1 who took the reading test read below the 12th-grade level (vocabulary level, comprehension level, or total reading level), and 35% (5) read at or above the 16.9th-grade level. None of the students in Group 2 read below the 12th-grade level on any measure.
Materials and Resources Materials and resources included the following:
54


Faculty Index of Story Difficulty (Appendix A);
Nelson-Denny Reading Test, both the Vocabulary and Comprehension/Rate subtests (Form E, 1981 edition). Although the college uses the ASSET tests for academic advising, I used the Nelson-Denny test for two reasons. First, it is used by the Reading Department at the college. Second, according to Ysseldyke (1985), it is one of the few measures of reading achievement designed for use with
.. .college-age students and is carefully constructed (p. 1037). Grade equivalents for the test range as follows: vocabulary, 3.6 -16.9; comprehension, 3.6 -16.9; total reading, 3.7 -16.9. Ysseldyke noted that the Vocabulary test has a test-retest reliability range of .89-.95, and the Comprehension test-retest reliability range is .75-.82. Since the Reading Rate is not of concern in this study, its lower reliability range (.62-.82) does not affect the study.
Ysseldyke (1985) observed two problems with the Nelson-Denny Reading Test which potentially impact this study. First, the college and university sample on which the test was standardized were not truly representative; colleges and two-year schools with fewer than 5,000 students were over-represented, and Hispanics and Blacks were underrepresented. Since this is an exploratory, qualitative study, the representativeness of the norming sample is not of great concern (it is important only that the test yield a wide range of abilities). The second concern relates to validity and has greater significance for the study; Ysseldyke (1985) reported that there is little data in support of [claims for] intended uses of the tests, including identification of both superior students and those who may need special help in reading (p. 1037). However, the test should still facilitate identifying the best and
55


poorest readers on a relative scale; I did not use it for academic placement purposes. Given the lack of a demonstrably more valid test, I chose this one.
Hidden Figures Test (Ekstrom, French, & Harman, 1976), Part 1. I chose this test to assess the cognitive style of field dependence/independence for two reasons. First, research has demonstrated an association between the Hidden Figures Test and field dependence/independence (Nowrozi, 1980; Jonassen & Grabowski, in press; also see Glossary). Second, I wanted to minimize the amount of class time devoted to testing in order to maximize the amount of time for instruction. Since this test takes less than 15 minutes of class time, it was possible to administer it to Group 1 during the same class session as the Nelson-Denny Reading Test Furthermore, I didnt want testing to potentially create an environment that might suggest that the students learning was not my primary focus. At the beginning of the course, I told students that I would be collecting information that would help me help them as well as future students.
Although the Hidden Figures Test has been used as a measure of field dependence/independence in many published research studies, I could not locate any data on reliability or validity, even in the Buros mental measurements yearbooks (7th through 10th editions). However, Jonassen and Grabowski (in press) report that it has high reliability and good construct validity.
an introduction to literature textbook containing all or most of the stories to be read but generally without study questions (Barnet, Berman, & Burtos [1989] An Introduction to Literature) and, The Appointment in Samarra, reproduced from another text. Stories were selected (a) which, with two exceptions, none of the
56


students had read or studied before the semester began,11 and (b) for which, with one exception,12 the text did not include adjunct questions and/or commentary, since one would expect that reading stories with which one is familiar or which have adjunct questions and/or commentary would reduce the number of questions and comments as well as affect the type of questions and comments.
handouts to define journal writing, to describe the process used by three former students Student Accounts of Journal Writing (Appendix H), to provide examples (Appendix I), and to provide prompts for thinking about stories (Catechism for Stories, Appendix G);
The Student Accounts of Journal Writing is an optional job aid to scaffold students in developing their schemata for journals, including a flexible procedure to meet their needs. The Accounts were three former students descriptions of how they wrote their journals.
^Dinah had studied Hills Like White Elephants and The Cask of Amontillado. Frances had studied The Story of an Hour in my composition course the prior semester. She read The Cask of Amontillado voluntarily during the current study before I decided to assign it Lisa had read The Cask of Amontillado. Sheila had studied The Cask of Amontillado, but she did not submit a journal on it Trish had also studied The Cask of Amontillado.
Because three students had already studied it I initially planned not to assign The Cask of Amontillado. However, for several reasons, I decided to assign it: (a) I felt the remaining students deserved the experience of studying it; (b) it is such a difficult story students often seem to benefit from studying it a second time; and (c) many teachers often do not provide adequate insight into the story. I base my conclusions on comments prior students made when we studied the story and on my observations (as former chair of the English department) of English classes at the college.
12The Story, of an Hour had adjunct questions. I assigned this story because it provides excellent examples of dramatic and situational irony. But to eliminate the confounding variable of adjunct questions in other journals, I purposely avoided assigning other stories with accompanying questions. Although the journals on this story were not included in the primary analysis, they did suggest some relationships between students use of adjunct questions and field dependence/ independence. See discussion in chapter 5.
57


The Catechism for Stories is an optional job aid to scaffold students in (a) writing their journals and (b) developing their schemata for analyzing stories.
two HyperCard tutorial programs for students to use as resources if they wished: Writing Journal Questions and Writing Journal Comments',;13
a HyperCard resource on fiction (definitions and examples of 54 terms; practice applying many of the terms; brief discussion of journal writing; expanded discussion of the Catechism for Stories);
objective short-answer quizzes (Appendix E), to assess factual knowledge of the stories after students have written journals but before class discussion (these quizzes, of course, also served as one means to motivate students to read the stories carefully).
a multiple-choice mid-term and final examination, to assess mastery of the course objectives (the examinations also served as means to motivate students to focus on the objectives of the course as they read the stories and attended class); recall items were included to determine if students had mastered the terminology; application items relating to stories discussed assessed near transfer; application items relating to an unfamiliar story assessed far transfer, which is widely accepted as a primary aim of problem-solving (e.g., Hayes, 1985; Newell, 1980) and of education in general (e.g., Prawat, 1989).
13Until the fourth week of the term (when the college opened a laboratory with the colleges first Macintosh computers), these were accessible only by making an appointment with me to use them on a faculty computer. Only one student took advantage of the opportunity to use them.
58


questionnaires and semistandardized interview forms to gather demographic data and data about students uses of journals and their perceptions of the value of writing journals specifically in the study of fiction (Appendices A, B, C, D, and F).
Data Collection
In the tradition of qualitative methodology, I triangulated (e.g., Berg, 1989), using a variety of data sources, all listed below. I piloted all of the paper-and-pencil questionnaires in the fall of 1989, then revised and piloted them on students in three literature courses in the spring of 1990. I further revised surveys of journal writing (Appendices B and C) to include questions about handouts on journal writing, which I developed primarily in response to students suggestions in the spring of 1990.
1. Student journals. Students wrote their questions about and comments on each story in a journal before class discussion. I chose an unobtrusive journal format rather than think-aloud protocols to study the types of questions and comments students have, because writing questions and comments in journals parallels the normal study behavior of students more than do think-aloud protocols. Students had the opportunity to read the stories as many times as they felt necessary, to look back at passages, to look at their notes, etc. Intuitively, journals are ecologically valid teaehing/leaming strategies; the very nature of their instructional use also reflects one function of direct quotations in qualitative methods as described by Patton (1980):
[Direct quotations are a basic source of raw data]... revealing respondents level of emotion, the way they have organized their experiences, and their basic perceptions [e.g., schemata relevant to the story]. The task for the qualitative methodologist is to provide a framework within which people can
59


respond in a way that represents accurately and thoroughly their points of view about the world, or that part of the world about which they are talking.... (p. 28)
Although I cannot be sure that the students journals accurately and thoroughly represent[ed] their points of view, use of journals would seem to allow and encourage this more than think-aloud protocols. It is important that the journals be confidential (between the student and me). I do not share them with other students, as is often the case with instructional methods which incorporate journal-writing (e.g., Duke, 1982; Nugent & Nugent, 1985). From personal experience, I have found that students are more willing to share their comments and questions with the instructor than with other students;14 and although no student may object in class when the instructor asks, Does anyone not want to share his journal with other students [in groups], at the end of one semester in which I had students share their journals in small groups, one student did indicate his objection on a survey I administered. Time constraints made it equally unrealistic to expect students to write journals in the presence of the researcher, who would be able to observe the actual activities students engage in. I chose to trade off (a) the ability to directly observe students writing journals on much shorter and much easier stories under artificial conditions for (b) ecologically more valid conditions for students studying reflective short stories of the type and length normally assigned in the course.
2. Surveys of Students
a. First day of class, a checklist of prior reading, to determine stories students had read (I used responses to select stories most students had not yet read) and
14See Lauras comment on the most helpful aspect of journal writing, in chapter 5.
60


a questionnaire to collect demographic data, including major, background in literature and number of college-level credits (see Appendix D). Orally and in print on various surveys, I told the students that I was collecting data so that I could learn more about how students learn to understand stories and so that I would have information to help me help them and future students, b. After first journal and last journal, survey of journal writing to collect data on students procedures in writing journals and on their reactions to writing journals (see Appendices B and C).
3. Interviews to probe students journal writing behaviors. During the term, I interviewed five students who differed across a range of dimensions (see discussion of selection criteria in chapter 4). I used a semistandardized interview format (Berg, 1989; see Appendix F).
4. Paper and pencil quizzes to assess students factual knowledge of each story; I administered a five-item objective quiz at the beginning of class after students submitted their journals but before class discussion of each story (see Appendix E). The purpose was to motivate students to read the story carefully as a basis for class discussion.
5. An objective multiple-choice mid-term and final examination to assess students mastery of the course objectives (e.g., verbal information on concepts and principles; application of concepts such as protagonist, irony, and theme; and problem solving). The midterm (the end of the eighth week) addressed only the short-story objectives; the final (on the last day of class) was comprehensive.
61


6. Nelson-Dennv Reading Test (Vocabulary and Comprehension/Rate; Form E, 1981 edition) during the third week of the study, to assess the typicality of the respondents (students) and to help determine cognitive ability.
7. Hidden Figures Test (Ekstrom, French, & Hannan, 1976), Part 1, during the third week of the study, to assess cognitive style.
8. Student records, to identify exact age, and semester and grade-point averages.
Procedure
I served as researcher and teacher, teaching the course in the same basic manner that I have used for about 20 years in all of my coursespredominantly teacher-led discussion (I have also had students write journals in most of my courses during the past seven years). Students studied fiction during the first seven weeks of the course, reading a story and writing a journal about once a week. They wrote two additional journals on plays.
Table 3.2 provides an overview of the research procedure. This overview elaborates three parts of the Conceptual Framework in Figure 3.1: developing tools and indices and instruments for observation and measurement; observation and data collection; and data analysis. Following is a detailed discussion of the procedure used with Group 1, which met twice a week for 75 minutes:
1. Established difficulty level of stories. Weeks 1 6:
I planned to sequence stories along a continuum of generally increasing difficulty. In determining the difficulty level of stories, I rejected readability indices although they have been used in studies of fiction (e.g., Shannon, Kameenui, & Baumann, 1988). Readability indices were derived from studies of expository prose, whose
62
I


Table 3.2. Overview of the Research Procedure_________________________________
Established the difficulty level of stories potentially to be assigned.
Oriented students to the course and to the study, and collect demographic data. Selected and sequenced stories to be assigned.
Administered tests of reading ability and field dependence/independence.
Presented schema for classically-structured stories.
Established baseline of students journal writing activities.
Administered Survey of Journal Writing, After First Journal.
Conducted class and collected journals for analysis:
Assigned story and journal; gave quiz and discussed answers; discussed story. Interviewed and tutored students.
Administered mid-term examination.
Administered End-of-Journal-Writing Survey.
Analyzed and interpreted data.
Developed taxonomy of questions and comments in students journals.
Developed cognitive model of journal writing.______________________________
purpose is to transmit information as clearly as possible, typically in a top-down structure. The indices are based on the average number of syllables and average sentence length. However, research showing the importance of general world knowledge in comprehending expository prose (e.g., Brown, Campione, & Day, 1981) calls into question the usefulness of any of the currently existing readability scales even for expository prose.15
15Brown, Campione, and Day (1981) note that the teacher may select texts that deal with familiar material or actively attempt to provide the requisite background knowledge for a particular text, but she/he cannot always do this. For several reasons, this problem would seem to be exacerbated in the i study of literature. First, college-level literature makes extensive use of allusions. Second, many writers (e.g., Donald Hall and Alexander Solzhenitsyn) have argued that literature is a vehicle for expanding knowledge of people, history, and other cultures. Third, teaching declarative knowledge out of context (i.e., as background knowledge for a particular piece of literature) is, at best, artificial. Moreover, such an approach does not teach students to learn to identify and debug their
63


In contrast to the purpose and techniques of expository prose, fiction requires the reader to construct meaning in a way that expository prose does not The basic structure is never top-down (although the structure of individual paragraphs may be), and the author often delays or even withholds information for artistic or affective purposes. Thus the difficulty level of stories for a given readership (e.g., college-leyel readers) is more a function of the demands placed on the reader by factors such as the narrative structure; the point of view; and the use of allusion, symbolism, and irony.
Faculty ratings. I had planned instead to use faculty ratings of story difficulty (Appendix A) to guide my selection and sequencing of stories. During the first week of the term, I administered the survey to the other six full-time members of the college-level English program. The survey revealed no consensus,16 17 and suggests the need for a future study to try to develop a tool for identifying difficulty. The range of responses provides further justification of the usefulness of journals in communicating students difficulties. If teachers with so much experience^7 do not agree on the difficulty level, then at least some of them must have inaccurate schemata of their students needs.
comprehension problems. Learning how to leam and to debug ones own knowledge has been called the most important goal of instruction (e.g., Norman, Gentner, & Stevens, 1976).
16For 29% of the 14 stories rated by more than one person, the ratings spanned the range. Including my ratings for those stories increased the percentage to 36. The faculty did not agree unanimously on the rating of any story, although 83% (5 out of 6) did rate Faulkners The Bear as Difficult.
Note: I asked respondents to focus on the difficulty of understanding a storys meaning, not on its vocabulary level.
17Four of the teachers had taught full-time more than 20 years; the fifth had taught full time more than 15 years; the sixth had taught full time less than 2 years.
64


Since there was no consensus, I decided to rely on my own perception of the difficulty levels based on the journals and class discussions from prior semesters.18 I wanted to minimize assigning stories students were already familiar with, because I wanted students journals to be as free as possible of the influence of prior exposure and study. (See Step 3 for the actual selection and sequencing.)
Results based on assumptions of story difficulty are necessarily tentative. However, with many years experience reading students journals on four of the stories, I am confident of the rankings I assigned. Furthermore, my estimations of story difficulty correlated 1.0 with the ratings assigned by an adjunct faculty member who requires her students to write more structured journals than I assign and who has many years experience teaching college English and literature. She based her ratings on criteria similar to mine. I asked her to rate the stories, without knowledge of my ratings, during the data analysis phase.
2. Oriented students to the course and to journal writing. Week 1. class 1:
I oriented students to the course and the study. Orally during the first day of class, I told students that I was collecting data so that I could learn more about how students learn to understand stories and so that I would have information to help me help both them and future students. The syllabus also included directions for writing journals (see Step 6 below).
I administered the survey of students backgrounds in literature and the checklist of prior reading (see Appendix D).
18I had taught most of these stories to at least 30 classes in the previous six to eight years.
65
nnrnu


I gave students in-class practice in writing a journal on a haiku poem19 (asking them to write any questions or make any comments that they would like to make). To introduce journal writing, I chose a haiku poem rather than a short story primarily because of time constraints. The brevity of a haiku poem allows time for thoughtful reading, writing, and discussion within one class session. That particular haiku also stimulates discussion of the problems that arise when a reader has an inadequate schema for concepts in a work of literature.
I collected their journals, and, without reading their names, read what each student had written. I noted that many students had written similar things, but that some had written quite different questions and comments. I emphasized that each journal was acceptable. Then using students questions and comments, I helped students construct more complete understandings of the poem.
3. Selected and sequenced stories. Weeks 2-6:
I drew on my prior experience in teaching stories (see Step 1) and the student checklist of prior reading (Appendix A) as a guide in selecting and sequencing stories for study. My primary goal was instructional; each story satisfied one or more major instructional goals.20 Within that constraint I selected and sequenced stories that would also satisfy my research goals. Following is an explanation for
19Diagon Flies, an 18th-century Japanese haiku.
20Except for Cathedral, each of the stories has stood the test of time and is generally regarded as a classic work in the canon of short stories. Although I do not consider Cathedral as outstanding as the other works, it is highly regarded and is included in many anthologies for college-level fiction and introduction to literature courses.
66


the selection and sequencing of specific stories; with one exception (story b), none
of the stories had adjunct questions (a copy of each story is included in Appendix
J):
a. The Appointment in Samarra is a relatively difficult one-paragraph tale. I wanted the first story to be challenging enough that all students would be likely to have some questions or points of confusion, but short enough that students would not become overwhelmed. Thus, I would have the opportunity to observe how they responded to a story they could not readily comprehend. For example, would they communicate their confusion through their journals? What types of questions, if any, would they ask?21 Would they generate self-explanations? This story involves schemata that students typically have difficulty comprehending (e.g., the concept of unavoidable fate, and the association of death with white rather than black); furthermore, the outcome relies on irony, and the author does not follow the convention of enclosing dialogue in quotation marks.22
b. The Story of an Hour. I selected this story predominantly for pedagogical, rather than research, reasons. It is a short, relatively easy story, yet its use of irony and its tum-of-the-century schema for women always pose problems for many students. Although this story had adjunct questions, I decided to assign it because it is the best story in the students text to illustrate dramatic
21 They might ask questions directly or indirectly, and they might ask open or closed questions (see definitions and examples in the Glossary).
22Instructionally, the story fostered discussion of several story elements; for example, setting, protagonist, antagonist, conflict
67


and situational irony. Many students find these concepts difficult to understand as well as apply, yet they are important for students to master early in the course since so many stories exploit irony. The adjunct questions ultimately provided insight into students journal writing, since some students apparently used them as prompts (see discussion in chapter 5).
c. Hills Like White Elephants. I chose this story to help students develop their concept of point of view. But it is, in my view, the most difficult of the five stories, not only because it uses the objective point of view but also because it withholds the specific topic of the conversation (an abortion). Generally students are unable to infer the correct topic. A short journal from a former student illustrates the reaction of many students:
I dont care how renown E. Hemingway is, he has no right to write a story which no reader can grasp. It is one thing to be intrigued by some nebulous thing in a story, but it is another to have the whole story be about that thing and to never find out what it is. Marriage, an illness, duo suicide? There is no way of telling. And the story is pointless without knowing! Ick!
d. Cathedral. I chose this story to build students confidence in their ability to understand short stories while I also introduced symbolism and gave students the opportunity to read a story with a happy outcome.23 Placing it after the most difficult story allowed me to examine interactions between journal items and difficulty level.
e. The Cask of Amontillado. I selected this story for two reasons. First, it allowed me to address major instructional goals which students had not yet
^Students often complain that all of the stories deal with death or other sad outcomes. Indeed, most anthologies have few happy or humorous stories.
68


mastered Second, it is a masterpiece which every student should have the opportunity to study, not merely read. I placed it fifth24 because it is difficult for many reasons: its vocabulary; its unfamiliar schemata (e.g., the brotherhood of the masons; coats of arms; family mottoes; and Mardi Gras, which is identified only as carnival season); its symbolism; its subject (murder for revenge); and the other problems it poses (e.g., Does the murderer achieve revenge as he defines it, and why does he want revenge?). Initially I did not intend to assign this story because three students had studied it in another course, but I decided that all the students would profit from studying it (see footnote 11).
4. Tested reading ability and field dependence/indeoendence. Week 3, class 1 A counselor/academic advisor at the college administered the Nelson-Denny Reading Test and the Hidden Figures Test of field dependence/independence while I was attending a professional meeting. I had told students that I planned to use the test results to help both them and future students.
5. Introduced classic short story schema and oriented students to predicting outcomes. Week 1, class 2; and week 2, class 1:
I read Petrbnius tale The Widow of Ephesus in class, stopping periodically to have students predict whether the story was finished and why they thought so.
24I have often taught this story first, thinking that class discussion would help students realize the types of problems one can encounter in trying to understand a story. But I now believe that placing such a difficult story at the beginning of the course is counter-productive. It de-motivates some students (they believe that all the stories will be totally beyond their grasp), and it fosters overgeneralization in some others (they believe that all stories must contain esoteric elements and seem to try to find ways to justify totally unwarranted interpretations of various elements). Sequencing should try to optimize the level of cognitive dissonance (see discussion in chapter S) and thus be responsive to students zones of proximal development (see discussion in chapter 5).
69


Then I explained the classic short story schema, applying it to the story. I also introduced the concept of plausibility, applying it to the story. The classic short story schema includes the following elements:
Exposition
Setting
Denouement
Symbols
Irony
Tine
Place
Verbal
Dramatic
Situational
Protagonist
Antagonist
Conflict
Crisis
Climax
Consequences
Point of View
Atmosphere
Theme
For several reasons, I refined and expanded the various story schemata used by many investigators (e.g., Adams & Collins, 1977; Rumelhart, 1977; Singer & Donlan, 1982; and Thomdyke, 1977). First, Singer and Donlan (1982) found the schemata of the other researchers and theorists inadequate for comprehending complex short stories and thus developed a new schema. Singer and Donlan emphasized that the other schemata focus on plot and do not acknowledge the importance of theme in complex short stories (or, for that matter, any adult, reflective fiction). However, as noted in chapter 2, after Singer and Donlan found facilitative effects for learner-generated questions for the other parts of their schema but not for theme, they concluded, in retrospect we should not have included theme in our instruction because it is not a compelling aspect of story comprehension (p. 182); they cited Adams and Collins (1977) for support Yet every college literature text with instructional apparatus (and/or teachers manual)
70


presents theme as the all-embracing element of college-level stories (apart from escapist literature).
Second, none of the schemata cited above acknowledge elements that typically contribute to the theme directly or indirectly (e.g., setting, symbols, point of view, and irony). The inability or failure to recognize irony can result in a total misunderstanding of a story. I wanted to provide a schema which would encourage students to think about all relevant aspects of a story and the way they interrelate.
6. Assigned story 1 and journal 1: conducted Survey of Journal Writing. After First Journal. Week 2, class 2:
Storv 1 and journal 1: For homework students read a one-paragraph story and wrote their first journal. I scaffolded students in this assignment by giving them a copy of the Catechism for Stories (Appendix G), three students accounts of journal writing from spring 1989 (Appendix H), and a copy of a journal from spring 1989 (Appendix I).25
The syllabus included the following instructions for journal writing:
Write a journal for each story, at least a half-page longbut there is no maximum length. Write your journal in ink, using complete sentences otherwise I will not be able to understand what you are asking or saying. Write any questions you haveexcept the definition of a word, unless you are unsure of the relevant definitionor any comments you would like to make about the story. The journal is not busy work, but is intended to help
^During the fall, 1990,1 gave students copies of these documents after they had written their first journalI had wanted them to try to discover their own approaches to journal writing before giving them accounts and models from other students. Although several students said they found the documents helpful, they, nevertheless, said they wished they had had them before they wrote their first journal. Since then, I give these to students before they write their first journal.
Note: The HyperCard tutorials Writing Journal Questions and Writing Journal Comments were not readily available when the semester began. A major question is whether examples facilitate or impede students own journal writing. The tutorials now provide an alternative learning strategy whose benefits must be investigated.
71


me know what you need help with in understanding the story. Write any questions you believe will help you understand the story now as well as later, for the examinationor any other questions you are just curious about Journals are worth a maximum of 10 points each. If a journal meets the minimum requirements, you will receive 7.5 pointsa grade of C; you will receive additional credit depending on how much your questions and comments indicate that you are paying attention to all the elements as you try to understand the story. There are no wrong questions or comments (but there, are questions and comments which ignore the details of a story).
Journals are a tool for you to leam about what you dont know, not to show me that you already know everything. Even if you dont understand a story at all but ask questions about what you dont understand and indicate that you are paying attention to details, you can receive a 10.
I will collect your journals at the beginning of the class on which they are due. If you miss that class, you may submit your journal at the beginning of the next class you attend.26
(Note: Journals counted approximately 20% of the course grade.27)
Survey of Journal Writing. After First Journal: To preclude any effect of feedback and grade on students evaluations of journal helpfulness, I administered the first journal survey right after I had collected the first set of journals..
Although student anonymity on surveys is generally considered desirable in experimental research, qualitative research does not impose such constraints. Furthermore, in this ecologically valid setting (the students were enrolled for credit and were entitled to as much help as I could give them), I needed to know students identities for two reasons.
First, I needed to know how individuals responded so that I could do followup interviews and adapt in-class instruction or offer tutorial assistance as
26Later submissions defeat the purpose of the journalsto have students think for themselves rather than merely record class comments.
-' There were journals on five stories and two plays.
72


necessary. Second, I needed to be able to identify respondents so I could analyze the data from the various sources.
7. Assigned stories 2 5 and their journals. Weeks 4 7:28
Students read a short story for homework and wrote a journal, about one a week (see instructions in Step 6). I photocopied each journal after I had graded it
Students took a five-question objective quiz on the story (to ensure careful reading of the story). Quizzes counted five points each, totaling approximately 10% of the course grade; I dropped the lowest quiz grade.29
We discussed each story (usually a week each) and applied the story schema. Discussion addressed questions and comments from the journals of current and former students and students in-class questions and comments. Class participation counted 10% of the grade.
I responded in writing to students journal questions and comments and returned journals at the next session. If the same question or misconception appeared in several journals, I included it in the class discussion, without identifying the student(s) who wrote it.
8. Administered midterm examination. Week 8, class 1:
I administered an objective, multiple-choice midterm examination on short stories
(40 questions, 100-points). The midterm and final examination each counted about
27% of the grade.
^There was no class session 2 of week 3; I was attending a professional meeting. I dismissed all students in lieu of out-of-class interviews later in the term.
A blizzard forced the closure of the college for session 1 of week 7.
29There were quizzes on six stories and one play.
73


9. End-of-Joumal-Writing Survey. Week 12, class 1:
f
After students had written their last journal (on Sophocles Oedipus the King), I administered the End-of-Joumal-Writing Survey. I administered it at this point rather than after the unit on short stories for two reasons. First, I wanted to give students as much opportunity as possible to incorporate journal writing into then-study repertoires. Duffy and Roehler (1989) found that students may require several months to incorporate new learning strategies. Second, because short stories and plays exploit many of the same techniques, I believed that writing journals on plays would be a relatively near transfer task.
f
Since I wanted students to be as honest as they could and not fear that their comments might influence their grades in the course, before they took the survey I explained the process that we would use. They would put their surveys in a packet which a student would seal and take to the division office where it would be held until after I had submitted students grades for the course.
t
10. Researcher activity. On-going:
Developed additional instructional materials and quizzes. After I had selected stories, I developed an additional factual quiz (see Appendix E)30 31 and an objective midterm3;1identification of definitions, principles of fiction, and application of terms to stories discussed in class and to one new story; applications included problem-solving (e.g., identifying the theme of the new story). With Trishs
had previously developed quizzes for all but Cathedral.
31I adapted an existing midterm to the works I had assigned.
nmirni
74


permission, I also typed a copy of her fourth journal (see Appendix K) and
II.
distributed it to students as an additional example of journal writing.32
Conducted interviews. I interviewed five students to probe their journal-writing strategies. Students schedules and mine did not permit me to interview more students during the weeks they were writing journals. (See case studies in chapter 4.)
Tutored any students who asked for help. I tutored the one student who asked for help (Lisa).33 The purpose of tutoring was to help her develop more effective strategies for reading stories and writing journals. I modeled the process of generating questions and self-explanations, and scaffolded and coached her, in keeping with the model of cognitive apprenticeship described by Collins, Brown, and Newman (1989).34
11. Analyzed and interpreted findings. After all the data were collected:
I developed a taxonomy of journal questions and comments. My primary goal was to try to increase understanding about the cognitive processing of students in
32Some students (particularly in Group 2) still seemed to be having difficulty writing productive journals. Others (e.g., Sheila) had been productive on more difficult stories but on the easiest story (4) wrote mainly affective or judgmental statements. By contrast, Trish understood the story well but took the opportunity to explore higher-level issues, such as the role of the setting.
33Tutoring was in my office before class, six times during a two-month period. Sessions generally lasted at least one hour.
^In the past, I have only provided advice on journal writing to the class as a whole (no one ever asked for additional help). As I conducted this study, I wanted to take a more proactive position. In my comments to the class as a whole and in my written feedback on journals of students who seemed to be having difficulty, I encouraged students to come in for help if they wanted to, so that I could work with them individually. I also encouraged students to view the tutorials on writing journal questions and comments. As noted elsewhere, Lisa was the only student to ask for help or to use the tutorials, although on the End-of-Joumal-Writing Survey (Appendix C) several students said they did not understand the purpose or grading criteria and suggested that I provide more help in the future.
75


their journal writing. I particularly hoped to be able to identify levels of questions and comments and to determine any relationships these had to individual learner characteristics and other variables (e.g., students ratings of journal helpfulness). But I also wanted to develop a taxonomy that would have practical value for a literature teacher.
Although I applied all but the final taxonomy to most of the journals, for several reasons I focused my analysis on journals 3 and 4. First, I wanted to examine journal writing after students had had an opportunity to develop their skills and become comfortable with the process. In journal 1, and to some extent journal 2, students were learning how to write journals.35 Second, journals 3 and 4 were on the most difficult and the easiest stories, respectively. They allowed me to examine potential effects of difficulty level. Third, I wanted to maximize the sample size; thus I chose journal 3 rather than journal 5 (also a difficult story) because more students submitted journal 3.36
Like Chan, Burns, Scardamalia, and Bereiter (1992), I began by trying to understand the data [in my case, the content of journals] at an intuitive level and then.. .moved toward a more objective scale that preserved as much of the intuitive understanding as possible (p. 101). Moving toward a more objective scale progressed through three major incarnations and several minor ones. For
35 Also, in journal 2, some students were apparently influenced by adjunct text questions. See discussion in chapter 5.
'Sample size for comparing journals 3 and 4 was 14; sample size for comparing journals 4 and 3 would have dropped to 12. Sample size would have dropped to 10 or 11 for some of the correlations with individual learner characteristics (e.g., reading ability and field/ dependence/independence). I wanted to maximize possible statistical significance of correlations.
76


example, after several attempts to operationalize shallow versus deep processing on two or three levels, I determined that the questions and comments in students journals were too remote from the actual cognitive processes to reliably determine the level.37 (See Appendix K for samples of students journals.)
The unit of analysis was generally a sentence; however, if a sentence seemed to have more than one purpose, I classified each one separately.38 Recognizing the importance of the inductive approach of qualitative methods, I wanted to guard against imposing an existing taxonomy on the data.39 (See Appendix L for a list of the categories in the taxonomy and Appendix M for examples.)
I correlated findings with individual learner differences. I calculated Pearson product-moment correlation coefficients to determine if there were any relationships between categories in the taxonomy and individual learner characteristics such as age, ability (vocabulary, reading comprehension, and total reading levels as measured by the Nelson-Denny Reading Test), college grade-point average, class performance (quiz scores, journal scores, and examination scores). The sub-groups were too small to analyze relationships between other variables and sex or background in literature. I also calculated Pearson product-moment correlation coefficients to determine if there were any relationships
3 ^Reliability problems were pronounced when I reapplied the taxonomy to particular journals; I often made contradictory classifications or found myself puzzled over which level applied. See the related discussion of cognitive artifacts in chapter 5.
38Ihis occurred so infrequently I did not statistically analyze the occurrences.
^'Initially, I had also hoped to apply several existing taxonomies (e.g., Bloom et al., 1956; Gagne, Briggs, & Wager, 1988) as points of comparison, but I found them impracticable for the complexity and diversity of students questions and comments.
77


between journal performance (e.g., grades and content) and other factors (e.g., student ratings of journal helpfulness and their ratings of job aids).
To identify potential interaction of ability level, story-difficulty level and question frequency (see discussion of Miyake and Norman, 1979, in chapter 2), as well as frequency of self-explanations, I split Group 1 into good40 versus poorer reading levels and performed several analyses, including Pearson product-moment correlations and one-way ANOVAs.
I developed a cognitive model of journal writing. I had hoped to develop a model such as a flowchart, which depicted a clear sequence with decision points. However, when I examined data from the student surveys, interviews, and journals, I discovered that no single model could describe journal writing either within or across students.41 For example, the difficulty level of a story affected the process in a variety of ways (type of activities, order of activities, and number and types of recursions) for almost every student42 At least one student (Foster) varied heir procedure so she wouldnt get bored.
It is possible that students would develop a more stable procedure over time, as they incorporated journal writing into their repertoire of learning strategies.43
^Only 12 of the 15 students wrote all five journals. The top six students all scored at the 16.9 grade level on reading comprehension, vocabulary and total reading ability. The bottom six students were distributed over as many as six grade levels.
41I tried developing several traditional models; for example, a very complex flow chart, a chart of the type used in an extended task-analysis procedure (Jonassen, Hannum, & Tessmer, 1989), and a complex circular model of the type popularized by Romiszowski (1981).
42Eric was the only one who seemed to follow the same basic procedure for each work.
43 As noted above, Duffy and Roehler (1989) found that students may require several months to incorporate new learning strategies into their repertoires.
78


However, it seems more likely that, at least for college-level students, journal writing in literature courses is an ill-defined domain (see Spiro et al., 1988), responsive to the demands a particular piece of literature makes on the student. (See chapter 5 for a description and discussion of the generic model I developed.)
12. Ensured the trustworthiness of the study. Throughout the study:
In this section, I am applying the terminology Guba and Lincoln (1982) advocated for naturalistic (qualitative) research. They emphasized the need for terminology separate from that used to describe rationalistic (scientific) research because the two types of research make different assumptions about the nature of reality, the relationship between the inquirer and the object of study/respondent, the nature of truth statements, the nature of causal relationships, and the role of values within disciplined inquiry. Trustworthiness addresses four issues (Guba, 1981): truth value (internal validity in rationalistic research versus credibility in naturalistic research), applicability (external validity versus transferability), consistency (reliability versus dependability), and neutrality (objectivity versus confirmability).
I implemented the following strategies, which Guba and Lincoln (1982) proposed for naturalistic research, to try to ensure the trustworthiness of my study.
To establish and test credibility (internal validity), I
had a prolonged engagement with the students; the short-story journal-writing segment spanned 7 weeks; total data collection spanned 15 weeks, beginning with the background survey and concluding with the last-day survey of improvement in confidence in ability to write journals;
79


engaged in persistent observation; I met with the students as a group 150 minutes a week for 15 weeks, 7 of which directly involved the study of short stories and 9 of which involved journal writing; and interviewed five students outside of class;
engaged in peer debriefing; I met with my dissertation advisor and another member of my committee to help gain perspective on my study as it was evolving;
used triangulation of data sources (e.g., surveys; interviews; content analysis of journals; comments during and outside class; and objective tests of students reading abilities and field dependence/independence);
used referential adequacy materials; I recorded field notes (more than 40,500 words) in a word-processing file after each class session and tutoring session. I kept a separate word processing file (more than 16,000 words) of notes on relationships of data to theory and prior research, possible significances of the data, futiire research, and recommendations for future implementation of journal writing. I did not audiotape or videotape class sessions because I did not want to interfere with the primary purpose of the classto facilitate student learning in an academic course. On the first night, when I was prepared to audiotape the class session, I asked the students, Would you tend to participate less if I audiotaped the class? Since a few said or nodded their heads yes,441 said I would not audiotape any of the class sessions. I used semi standardized interviews; I printed the questions and took notes during each interview. And I photocopied each journal in Group 1 after I had graded it
^One of those who said she would talk less was Frances, one of my composition students from the previous semester. See her case study in chapter 4.
80


To support transferability (external validity), I used
purposive sampling in interviews; based on my extensive teaching experience and knowledge of the population, I selected students who represented a cross-section of the population (random selection of only five students from such a small sample could have failed to provide the necessary diversity);
thick description; I grounded interpretations in the data, taking care to provide the reader with enough low-inference descriptions of behavior and excerpts from...interviews (Donmoyer, 1990, p. 196) or, in Phillips (1990) term, low-level observations (p. 26) that the reader will be able to evaluate the conclusions which I draw or generate his/her own. I provide thick description either in the body of the text or in footnotes, and included several journals in Appendix K.
To establish dependability (reliability), I used
overlap of methods, a form of triangulation; I used surveys, interviews, analysis of journal content and of comments during and outside class, and
dependability audit; I left an audit trail of my methodology and decision points.
To establish confirmability (objectivity), I
used triangulation (see above),
practiced reflexivity; I continuously examined the purpose, methods, potential biases, etc., and
used a confirmability audit; I traced each finding to its original data source; I have described many of these in the body of the text or in footnotes, and have included several journals in Appendix K.
81


CHAPTER 4
CASE STUDIES
The purpose of this study was to develop a cognitive model of journal writing in a college introduction to literature class. In this chapter I present the case studies of five students in Group 1. In chapter 5,1 systematically address each of the main research questions of the study.
Selection of Interviewees
I would like to have interviewed all of the students. However, students work and academic schedules and my professional obligation to assist students during office hours before and after class limited my access to them through interviews. In selecting students to interview, I tried to span as broad a range of variables as possible. However, my primary focus was on the following traits (a summary of those selected is in parentheses):
field dependence/independence (2 of the most field dependent, 2 of the most field independent, 1 in the middle),
reading-comprehension level (1 sub-college level, 1 college-sophomore level, 3 at least college-senior level),
age (2 who were below both the median and the mean, and 3 who were above),
academic experience (2 who had essentially failed in early college experiences; 1 with a mixed record who had returned to college; 2 who had first enrolled in
82


college as' nontraditional students but both of whom had high grade-point averages).
Secondarily, I was interested in the following traits:
sex (3 of the 12 females, 1 of the 3 males),
journal skill (2 very astute but not seeming to perform up to their potential in their journals, 1 very imperceptive but probably performing up to her potential, 2 seeming to perform up to their potential), and
class participation (the 2 most actively involved in class discussions, the 1 least actively involved, and 2 somewhat-to-moderately involved).
I was able to interview students who satisfied all the primary traits except age, and all the secondary traits. I would like to have interviewed at least one traditional-age student (particularly Todd or Bonnie, who were also the most field dependent in either group and who earned the heaviest total loads); however, our schedules did not permit it. Because Todd was extremely field dependent, in chapter 51 draw many examples from his journals to contrast the journals of the most field independent students.
I interviewed Lisa and Colleen after they had written their third journal on a story, and Trish after she had written her fourth. Because of the time constraints, I was not able to interview Frances and Eric while they were writing journals on stories; I interviewed them after they had written their sixth journal, on the play night, Mother. For brief biographies of all the students in Groups 1 and 2, see Appendix N.
83


i
\
These five cases illustrate that although instructional implementation could
I'
benefit by ensuring that all students understand the purpose and criteria of journal writing, journal writing can be a flexible cognitive tool. These students all found journal writing helpful. With the exception of Eric, who never read the Catechism for Stories, they relied more on job aids, such as the Catechism, at the beginning of the term than at the end, gradually becoming more responsive to their internal learning needs and turning to the Catechism only occasionally to get their wheels turning.
Lisa
i
... At the end of page 245 Its just to let the air in. Something soon after that comment leads me to believe there is one particular thing wrong in [the couples] relationship but again Hemingway doesnt tell us directly what it might be.
If; I am correct in saying this is the objective point of view. It does make this piece of work hard to put all the pieces of the story together. Because the author does not give us any thoughts or feelings regarding what they are intending to do....
Excerpt from Lisas third journal
i
PersonalBackgroynd
At the age of 36, Lisa had returned to college after 12 years. Because she often came to my office for help before class, I came to know more about her personally than I did about most of the other students. I share some of the more personal details of her life to illuminate characteristics about her reading and journal writing strategies later in this chapter and in chapter 5 and to demonstrate that being frank about ones
i
personal life does not necessarily mean that a student is totally frank about his/her understanding of course material.
1
84


Lisa is a gentle person who always seemed sensitive to other peoples feelings, but generally lacked self confidence. She is in her second marriage, and about midterm contemplated leaving her husband, particularly because she felt their six-
r-
year-old daughter would suffer in a marriage filled with strife. Her husband is a professional niusician who doesnt talk enough and who has been unwilling to
i
listen to any of her anxieties and problems with her school work. She worked 18 hours a week as a secretary. Being a mother, working, studying and requiring extra rest to combat a serious case of lupus did not leave much time for her with her husband. Moreover, she said their marriage had been one loss or other problem after another: they lost one child, they had had financial problems which resulted in the loss of the house they were buying; she has lupus, etc. Apparently our discussions helped Lisa resolve her marital problems at least for the remainder of the semester.
Lisa was registered for six credits (Introduction to Literature and Beginning Algebra); although she struggled with both, she earned Cs. Her educational goals fluctuated during the term. At the beginning of the term, she said that her goal was a doctorate in psychology. Later in the term, she focused on becoming an elementary school teacher., She had 68 cumulative credits, with a 2.8 GPA. She was field dependent, with a score of 4 (out of 16) on the Hidden Figures Test. On the Nelson-
I.
Denny Reading Test, she had a total reading level of 11.1 (vocabulary level, 12.7;
comprehension level, 9.4). She had perfect attendance in class, was almost always
|
cheerful and participated moderately, although she was probably the least perceptive student in the class.
85


Although she is not academically gifted, she was probably one of the most highly motivated and diligent students I have ever taught; her journals were the longest in the class (515; 1,332; 1,159; 1,722; and 1,173 wordsall but the last were at least 2.3 standard deviations longer than the respective averages). About midterm she said that the most important thing in her life was to complete her education and to become a teacher. She tried to relate whatever she was reading to her own life so that it would be more meaningful and easier to remember; however, this sometimes caused her to impose her life schema on the work, not attending to the differences between her life and the story.
She generally tried to apply whatever advice was given and to use the various aids available to her. She was the only one in either group to view the HyperCard tutorials Writing Journal Questions and Writing Journal Comments and the Introduction to Fiction program, and one day in class she told the students how helpful she had found the programs. However, she refused to accept the fact that she really needed to take a developmental reading course; I recommended that she take such a course before taking any other college courses because her reading problems were obviously interfering with her ability to learn. She was often unable to connect pronouns with their correct antecedents and to make other very low-level inferences. She generally had a much larger percentage of incorrect journal items than did other students; for example, in journal 3,42.9% of her items were incorrect, 3.3 times the standard deviation for journal 3. (But at least she was trying to make sense of the story; her 63 items were 3.6 times the standard deviation.)
86


In addition to staying after class occasionally to talk or to work on the HyperCard
(
programs in the English lab, she came for help outside class eight times, 30 to 60
i
minutes each. My extra help included asking her to elaborate on elements from the works we read in class as well as reading and discussing other works. For example, as she read The Chrysanthemums (a short story by John Steinbeck) out loud, I asked her questions about setting, motivation, meaning of specific passages, theme, and symbolism. We also read and analyzed several poems.
But her high motivation to succeed, coupled with her lack of self confidence
§
might have interfered with her optimal performance (see Yerkes & Dodsons 1908 law on the relationship between arousal and performance). For example, she said she suffered from test anxiety and discussed her academic problems with an academic advisor at least twice during the semester. She spent up to ten hours working on her journals for the class; she even took off work one day to write her journal because she had been ill the day before, her day off (she lost $40 pay). She felt the suggestions I gave her during the interview about journal writing were so helpful she asked if she could share them with the class.
Although she seemed comfortable talking with me, she did not give me a complete description of her efforts to understand assignments, either in personal conversations or in her journals. She was so anxious about understanding the stories when she wrote her journals that she discussed at least some of them with others. She did tell me her husband refused to discuss her school work any more, but she did not
lr
say that she discussed any of the assignments with other students while she was
l
writing her journals. In journal 5, she did not give credit to anyone for insight into the
87


Full Text
explanations. Reinforce students honesty and their efforts at inquiry and selfexplanations. Use questions and comments in the journals to help the class construct the meaning of the work.
i.
Depending on students needs, you might also want to have students collaboratively write a journal in class as a way of transferring the complete task from teacher/class to the individual students. I have not tried this, so I caution that it might actually interfere with students learning to monitor and respond to their own needs, particularly field dependent students, who seem to rely on discussion with others.
i.
5. Provide additional instructional support. Following are examples of the types I have used:,
a. Modelthe type of cognitive activities that journal writing typically involves by leading students through the reading of part or all of a short story in class.
b. Provide tutorials on asking questions, on generating self-explanations, and on reading a story.
c. Provide sample journals on stories the students will not be assigned to read. (Give students two or three samples that indicate the wide range of journal
l
content so that they dont develop narrow concepts of journal writing.)
d. Provide a job aid (such as the Catechism for Stories) to stimulate thinking.
e. Tutor students who need remediation in how to read a story.
f. Caution students if they are likely to find a work particularly difficult Remind them to look up unknown words and words whose meanings they think they
255



PAGE 1

A COGNITIVE MODEL OF JOURNAL WRITING OF COLLEGE STUDENTS iN AN INTRODUCTION TO LITERATURE COURSE by Margaret A. Cole B. A., Western Maryland College, 1963 M. A.,University of Arizona, 1965 B. A., University of Colorado at Denver, 1975 A thesis submitted to the Faculty of the Graduate School of the UniversitY of Colorado at Denver in partial fulfillment of the requirements for the degree Doctor of Philosophy Adniinistration, Curriculum, and Supervision 1992

PAGE 2

1992 by Margaret A. Cole All rights reserved.

PAGE 3

I'. This thesis for the Doctor of Philosophy degree by Margaret A. Cole has been approved for the School of Education by Brent G. Wilson Martin A. Tessmer / Dian E. Walster Ellen A. Stevens Edith w. King If )Jt?tl Date

PAGE 4

Cole, A. (Ph. D., Administration, Curriculum, and SupeiVision) A Cognitive Model of Journal Writing of College Students in an Introduction to Literature Course Thesis directed,by Associate Professor Brent G. Wilson ABSTRACT The of this research was to study how college students use journals in i understanding stories. Such information is prerequisite to developing more effective instructionai/leaming strategies to scaffold (support) students in comprehending literature. The literature describes the use of journals ln. all levels of education, but provides little information about (a) the relationships of journal writing to individual learning characteristics and task difficulty, or (b) the types of questions and comments students generate when they write journals to help them comprehend reflective short stories in context of a credit-bearing (i.e., an ill-defmed domain, an appropriate incentive, an ecologically valid context, and the opportunity for the learner to monitor his/her comprehension and generate whatever questions or comments he/she wishes). The psychology literature on adjunct questions, self-explanations, schema theory, and the model of learning/constructivism was incorporated to shed light on journal writing. The methodology was primarily qualitative. Data included students' journals and I class comments; of the helpfulness of journal writing and job aids; individual learning characteristics (e.g., reading ability, field dependence/ independence, academic IV

PAGE 5

record, course grade, and confidence); and case studies of five students. Findings indicated that (a) students generally viewed journal writing as a flexible cognitive tool which helped them construct the meaning of stories; however, students' perceptions of journal writing 'are particularly sensitive to implementation; and (b) writing journals scaffolds (supports) students in attending to details, asking questions, and answering ,, their own questions; however, high-ability readers tended to engage in different cognitive activities than lower-ability readers. Journal writing is a very dynamic process; no single cognitive model can describe journal writing within or across students. A two-part model was developed to describe the process: (a) factors which influence journal writing (task difficulty, individual learner characteristics, teacher expectations, student strategies, external resources, overtiactivities), and (b) components of journal writing (establishing a goal, constructing the ,textbase, constructing the situation model, predicting outcomes, identifying significant elements, reflecting on meaning, and assembling the schema). Recommendations for implementation of journal writing are included. This abstract accurately represents the content of the candidate's thesis. I recommend its publication. Signed Brent G. Wilson v

PAGE 6

DEDICATION I dedicate this thesis to my husband, Jim, and to all the students whose journal writing inspired this study.

PAGE 7

ACKNOWLEDGMENTS I wish to express my appreciation to the members of my committee for their v,arying perspectives, thoughtful suggestions, and support. Brent G. Wilson Martin A. Tessmer Dian E. Walster Ellen A. Stevens Edith w. King vii

PAGE 8

CONfENTS CHAPTER 1. ................................................................. 1 Limitations of Prior Research ................................................ 6 Journal Writing ........................................................... 6 ... Learner-generated Questions ............................................ 6 Justification of the Study ..................................................... 8 Research Questions for a Study to Develop a Cognitive Model of Jo4rnal Writing ................................................................ 16 ':Purpose of the Study ..................................................... 16 2. REVIEW OF RELATED LITERATURE. ...................................... 18 Writing ................................................................ 19 Adjunct Questions ............................................................. 27 Question Position ......................................................... 28 Question Level/Task Difficulty .......................................... 29 Additional Fin':fings on Adjunct Questions ............................ 31 Self-Explanations ............................................................. 32 Limitations of Research on Adjunct Questions and Self-Explanations ............................................................. 33 Theory ................................................................ 34 Relationship of Levels of Questions to Levels of Processing ....... 34 Pefmitions of Levels of Processing .................................... 35 The. Generative Learning Model. ............................................ 36 Vll1

PAGE 9

1 I Learners' Dispositions and Goals ...................................... 37 Focus of Research on Learner-generated Questions .................. 39 Learner-generated Questions as Study Aids ........................... 39 Summary of Learner-generated Questions ...... : ...................... 44 stinunary ....................................................................... 45 3. METHOD ........................................................................... 46 Conceptual Framework ....................................................... 46 and Setting ...................................................... 50 Background of the Researcher .......................................... 50 Setting for the Study ..................................................... 51 : Subjects/R.espondents .................................................... 53 Materials and Resources ...................................................... 54 Data Collection ................................................................ 59 :Procedure .................................................................. 62 4. CASE STUDIES .................................................................. 82 Selection of Interviewees ..................................................... 82 ............................................................................. 84 Personal Background .................................................... 84 iResponses to Interview (Fifth Week) .................................. 88 'sUIIlll13I)' .. 93 Colleen ......................................................................... 96 I fersonal Background .................................................... 97 Responses to Interview (Fifth Week) .................................. 98 I IX

PAGE 10

Summary .................................................................. 100 Trish ............................................................................ 101 Personal Background .................................................... 101 Responses to Interview (Sixth Week) .................................. 103 ,: Summary .................................................................. 106 ......................................................................... 107 Personal Background .................................................... 107 Responses to Interview (Eleventh Week) .............................. 110 Summary .................. : .............................................. 115 Eric .............................................................................. 115 Personal Background .................................................... 116 Responses to Interview (Eleventh Week) .............................. 119 Summary .................................................................. 122 Suininary ....................................................................... 123 5. ANAIJYSIS OF RESEARCH QUESTIONS .................................. 125 Research Question 1: What Metacognitive and Cognitive Activities Do Students Engage in When They Write Journals? .......... 125 :Metacognitive and Cognitive Strategies ................................ 126 1Cognitive Activities ....................................................... 131 Research Question 2: How Are These Activities Related to Individual Leamer Characteristics? ......................................... 141 Res.earch Question 3: What Types of Questions and Comments Do 'Students Write in Their Journals? ....................................... 146 Focus of Cognitive Activities ........................................... 146 I Comparisons of Journals 3 and 4 ....................................... 147 X

PAGE 11

, Story-specific Categories in Journals 3 and 4 ......................... 163 Omission of Significant Elements ...................................... 164 Research Question 4: How Are Questions and Comments Related to. Individual Leamer Characteristics? ...................................... 166 Question 5: Can a Taxonomy Describe Students' QUestions and Comments? ................................................... 17 4 Research Question 6: How Do Students Perceive Journal Writing? ........................................................................ 178 Group 1 .................................................................... 179 Group 2 .................................................................... 186 Relationship Between Grades and Ratings ............................ 190 Most Helpful Aspect of Journal Writing, .............................. 192 Least Helpful Aspect of Journal Writing ............................... 194 :Students' Suggestions for Making Journal Writing More Helpful ..................................................................... 196 Research Question 7: How Do Ratings of Journal Helpfulness to Other Variables? .................................................... 199 Research Question 8: How Do Students Rate the Job Aids? ............ 202 Student Accounts of Journal Writing ................................... 202 :Catechism .................................................................. 204 Research Question 9: How Do Ratings of the Job Aids Relate to Other Variables? ............................................................... 207 Student Accounts of Journal Writing ................................... 207 I .................................................................. 208 Question 10: Is There a Cognitive Model That Can Capture the Processing of Most Students' Journal-Writing Activity? ........................................................................ 210 xi

PAGE 12

Summary ....................................................................... 221 I II 6. DISCUSSION ..................................................................... 229 Interpretation of Findings .................................................... 229 A Flexible Cognitive Tool ............................................... 229 A Valuable Instructional Tool ........................................... 232 Individual Leamer Characteristics ...................................... 239 : Social or Individual Level of Knowledge Building? ................. 243 Liplitations of the Study and Recommendations for Future Research ........................................................................ 245 I Recommendations for hnplementation of Journal Writing ............... 253 : Alternate Strategies ....................................................... 263 Summary ....................................................................... 264 APPENDIX I 0 A. Faculty Index of Story Difficulty ................................................ 265 B. Survey of Journal-Writing: Beginning of Term ................................ 267 C. Survey :Of JournalWriting: End of Term ....................................... 270 D. Survey.ofBackground ............................................................ 273 I E. of Short-answer Quizzes ............................................. 276 F. Guidelines for Semistandardized Interviews After First and Last Journals ............................................................................. 278 G. for Stories ............................................................. 280 H. Accounts of Journal Writing ............................................ 282 I. On "FleUr," by Louise Erdrich ................................................... 285 Xll

PAGE 13

J. Stories on Which Students Wrote Journals ..................................... 287 K. Sample Journals From Group 1 ................................................. 310 L. Taxonomy Developed From Students' Journals ............................... 320 M. Examples of Questions and Comments in the Taxonomy .................... 322 N. Biogr:aphies of Students in Groups 1 and 2 .................................... 330 GLOSSARY OF SELECTED TERMS .................................................... 337 ............................................................................... 343 XIll

PAGE 14

CHAPTER 1 INTRODUCTION The general acceptance among educators of writing as a learning tool (e.g., Anson & Beach, 1990; Emig, 1977; Fulwiler, 1980, 1987, 1989; Irmscher, 1979; Moss, 1979) has led to widespread implementation of writing-across-the-curriculwn programs. One.of the most popular manifestations of writing as a learning tool in liberal arts prowams is journal writing (see Glossary). Articles on the subject abound in professional publications, reflecting diverse philosophical and pedagogical positions. Assigrunents range from totally open-ended journal writing to highly directive writing, such as expressing "gut" reactions, writing summaries, elaborating (e.g., comparing/ contrasting to one's own experiences, or applying concepts), evaluating content, or even, answering specific questions (e.g., Duke, 1982; Flitterman-King, 1988; McCormick, 1985; Nugent & Nugent, 1985). However, McCormick (1985) observes that a major criticism of the dominant approach to using journals in literature courses is that it calls merely for "associative, 'touchy-feely' reactions, and rather than opening up students' responses to texts, they restrict them to what students already know" (p. 837). Thus, she advocates writing which is given focus by the instructor. The diverse applications of journal writing extend beyond literature courses. A literature search of the ERIC database (CD-ROM, 1983-1990) using the term journal writing illustrates the scope of interest. The search located 372 items; another 19 were listed (under journals) in the 1991 ERIC printed index. However, barely one third of the entries reported formal research. Table 1.1 suggests the breadth of interest in 1

PAGE 15

journal writing. Of the 372 entries in the 1983-90 CD-ROM database, 25 (6.7%) related to literature courses in higher education; 7 of the 25 (1.8% of all the entries) reported formal research. A search of the same database located no entries using the terms journal writing and instructional technology or educational technology. In spite of 1the widespread use of journals as an instructional strategy, Anson and Beach (1990) reported that only a "handful" of the more than 600 publications on writing across tne curriculum or writing to learn, provided any empirical support for an "oveiWhelmingly endorsed ... practice" (p. 2). They added that the research studies were not conclU;sive and "raised ... many interesting problems about the nature of writing and learning" (p. 2). The only consistent fmdings on journal writing seem to be (a) that, when they have been asked, students have said that they found journal writing helpful (e.g., Cole, 1991; Cummins, 1989; Hettich, 1990; Schwartz, 1989), and (b) that reviewing all journals is very time consuming (e.g., Hettich, 1990; Roth, 1985). The term journal writing is not even used consistently in the literature. As Hedlund, Furst, and Foley (1989) reported, the terms journal, diary, and log are often used synonymously. The 372 entries mentioned above in the ERIC database included many whose texts actually used the terms diary or log. Some used the term journal to describe almost writing assignment other than an essay or term paper, such as responses to adjunct questions (e.g., Hettich, 1990) and dialectical reasoning (Jolley & Mitchell, 1990) .. Others emphasized the traditional features of a journal which are also salient in the current study; for example, the personal nature of its content (Hedlund, Furst, & Foley, active involvement (Hedlund, Furst, & Foley, 1989; Roth, 2

PAGE 16

Table 1.1. Samele of ERIC Publications on Journal Writing LEVEL DISCIPLINE DON.UUN/NUUORFUNCTION AU1HOR 1st grade mathematics reflective practice Wason-Ellam, 1987* lst-3rd, readin g comprehension Barone, 1990 mixed 6th grade readin g conative, self-concept (learned Coley & Hoffmann, 1990 helplessness) 6th grade mathematics problem solving Risk, 1988 1st-12th (all) metacognition Kuhrt & Farris, 1990 junior h.s. reading metacognition Pyle, 1990 middle sch. English (literature) comprehension Myers, 1988 high sch. English communication & writing McGuire, 1990 high sch. English (literature) comprehension Wilson, 1989 high sch. thinking skills Linn, 1987* highered. biology concepts; attitudes Trombulak & Sheldon, 1989 highered. communication reflective practice Buell, 1989 highered. engineering writing Rumpf et al., 1988 higher ed. civil 'engineering problem solving, communication Selfe & Arbabi, 1983 higher ed. English composition Stanley, 1989 higher ed. English as a second cultural understanding Steffensen, 1988 language higher ed. geography expressive writing for students, Sublett, 1988 diagnostic tool for teacher higher ed. linguistics (student determined) Anson & Beach, 1990* highered. literature comprehension Flitterman-King, 1988; Holland, 1989; Nugent & Nugent, 1989 highered. mass media diagnosis of teaching effectiveness Tamove,l988 & student learning higher ed. develop. reading metacognition Feathers & White, 1987* higher ed. physiciU education affective, conative Hedlund, 1990 highered. logy course concepts Hettich, 1990* highered. psychology critical thinking Jolley & Mitchell, 1990 higher ed. sociology reflective and critical thinking Roth, 1985 (sex/gnder) highered. speech/ multiple (e.g., brainstorming, selfJulian, 1989 commUnication evaluation) highered. study skills & (effectiveness of annotation vs. Hynd & Chase, 1990* literature journal writing strategies) highered. teacher education clarify understandings & relate Carswell, 1988 (curricUlum learning to personal and professional lives higher ed. teacher education reflective practice Bean, 1989; Mantey-Casimir & Wassermann, 1989 general general independent thinking; writing skills Robinson-Armstrong, 1991 *Research article 3

PAGE 17

1985), reflective thinking (Anson & Beach, 1990; Hedlund, Furst, & Foley, 1989), a dialogue with oneself (Roth, 1985), and independent thinking (Robinson-Armstrong, 1991). Figure 1.1 :depicts a model of journal writing in academic environments. In the broadest writing is intended to facilitate learning. But in a narrower sense, it has had three major purposes: (a) to facilitate text comprehension, (b) to I develop thinking skills (e.g., metacognition, analytical thinking, critical thinking, and problem solving), (c) and to develop language/writing skills (e.g., essay writing and creative writing). (I use the term text in the broadest sense, meaning any information provided by the instructional environment. It includes everything from a short story to I a problem in math and steps in writing a research paper.) Because journal writing may serve the first tWo purposes without leading to the third, a dotted arrow connects parts i B and C of the model. Figure 1.1. A Model of Journal Writing B A 4

PAGE 18

research issues. Following is a list of the most important issues on I journal writing which research must address: 1. What are the salient attributes of journal writing? under what conditions? for what learners? 2. Does journal writing indeed facilitate learning? in all disciplines? in ill-defmed as well as well-defined domains (e.g., Spiro, Coulson, Feltovich, & Anderson, 1988)? with all types of objectives (see Gagne, Briggs, & Wager, 1988)? with content of varying degrees of difficulty (e.g., Miyake & Norman, 1979)? with learners ofvarying characteristics (e.g., ability, motivation, cognitive style)? 3. Can effectiye journal-writing strategies be taught? 4. What type(S) of feedback/evaluation systems are appropriate? under what conditions?, for what learners? Should journals be graded? If so, how often? with what 5. What specifically do learners fmd helpful about journal writing? Are learners' perceptions of helpfulness tied to objectives? difficulty of material? evaluation procedures (including type of tests)? individual learner characteristics? 6. Is collaborative journal writing effective? under what conditions? for what learners? Little if any research has been conducted on the cogQ.itive foundation of journal 1writing as an instructiona]Jleaming strategy. There is no qualitative or quantitative research basis for advocating one method of using journal writing over another. Is it the novelty of writing that attracts its proponents and seems to facilitate student peiformance? or is there a yet-to-be identified cognitive basis for journal writing? If it 5

PAGE 19

is the latter, what is that basis, and does it involve a single strategy for all learners in all situations? ar::does it consist of a range of strategies which interact with content of the materials presented, type of expected tests and other evaluation strategies, and individual Ieamer differences? A rigorous investigation into the cognitive foundations of journal writing is warranted. i Limitations of Prior Research Journal Writin& Although jpumal-writing assignments already have a wide range of applications in higher little is known about their efficacy. Do they help students learn? But before we can a.Sk that question, we must develop a description of journal-writing strategies used by learners and identify those which seem most successful, for whom, and under what 'circumstances. We must also identify the types of questions and comments studdnts write in their journals; as Stein and Bransford (1979) observed, "An emphasis oh the types of questions students ask themselves may ... have important implications for 1understanding individual differences in learning and retention" (p. 776). Having dqne these two things, we will have a foundation for developing instruction to guide learners in applying effective strategies in journal writing. I QJlestions lIn addition to its focus on journal writing, this study differs significantly from earlier investigations of learner-generated questions which focus on what a teacher wants a student to learn. For example, M. E. D. A. Andre and Anderson (1978-79) and Frase and Schwartz (1975) directed students to generate questions which they 6

PAGE 20

believed would help them prepare for a comprehension test of expository passages; models of teacher questions were provided. This study focuses on the types of questions and comments students wish to ask for any reason-perhaps as much as any graded class assignment allows, each student's goals and interests fuel journal writing. Furthermore, while most studies of learner-generated questions have focused on what Ng and Bereiter (1991) call "extracting knowledge from text" (p. 268; e.g., Vander Meij, 1992; Wong, 1985), this study focuses on comprehending literature. Moreover, ,the studies have generally been empirical, conducted in artificial settings (e.g., laboratories and with little or no incentive) rather than in ecologically valid ,' settings. Hicken, Sullivan, and Klein (1992) demonstrated the importance of adequate incentives in research on learning. The artificiality of the settings was also apparent in other ways which may have influenced the outcomes. For example, although researchers have instructed learners to generate questions, they often have not answered those questions (e.g., Miyake & Norman, 1979). Gamer (1990) found that context (e.g., artificial versus ecologically valid setting) plays a significant role in subjects' use of learning strategies. Proponents of situated cognition (e.g., Collins, Brown, & Newman, 1989) ,would raise similar concerns about the validity of the setting. My exploratory studies found that students' questions and comments in required journals in literature classes do not merely reflect the types of questions they might expect on a quiz:or examination, but often their questions and comments seek (a) to clarify facts, and.'allusions, or to verify low-level inferences which will help them construct the meaning of a story, (b) to relate the story to their lives, (c) or to evaluate it in terms of, for example, its artistic merit, its plausibility, or its subject. 7

PAGE 21

Finally, this study differs significantly from earlier investigations of learner generated questions by going beyond an information-processing cognitive model of learning, and acknowledging ways in which journal writing is consonant with the constructivist (see discussion below). Justification of the Study Using journals of adult students to investigate learner-generated questions and comments seems justified on many counts, even though this method is very time consuming for both student and teacher: 1. The nature pf complex short stories. Reading a complex reflective short story (as opposed to escapist literature) is a problem-solving activity which requires the reader to construct its meaning with few structural cues. The reader cannot engage in well known strategies for reading expository prose (such as reading chapter outlines, advance organizers, summaries, and headings and subheadings; skimming to identify topic sentences; and looking at adjunct questions). The differences between expository prose (which has been the subject of most of the research on reading, adj\lDct questions, and learner-generated questions) and fiction have not been adequately acknowledged in research. Meyer (1975, 1977) provided convincing analyses of the structure of nonfictional prose. But even a simple short story does not obey the rules of nonfiction, and "good fiction [the type taught in college classes] operates" by "dynamic, subtle processes" (Bergstrom, 1983, p. Even though most stories utilize common elements (such as setting, plot and character) and conventions (e.g., Culler, 1975; Rabinowitz, 1987), each 8

PAGE 22

story comBines them in unique ways, requiring readers to discover its specific structure as they construct its meaning. 2. Intuitive appeal of journals. Journals are an intuitively appealing teaching/learning : strategy. Journals (as well as diaries and notebooks) have been used by some of the thinkers and writers throughout history (e.g., Pascal, Benjamin Franklin, Nathaniel Hawthorne, Henry Thoreau, Fyodor Dostoevsky, Bertrand Russell, JO,seph Conrad, and Albert Einstein). The use of journals has not seemed to impair and, according to the most widely accepted theory of processing (E. D. Gagne, 1985), should facilitate learning by extending (he limits of working memory and by encouraging learners to attend to the subject'matter and elaborate upon it (this facilitation is also consonant with constructivism). Similarly, asking questions in journals should "encourage students (1) to pause frequently while reading; and (2) to locate an appropriate answer to a:n 'understanding question' that probably is closely related to the study criteria" (T; H. Anderson, 1980, p. 500). The constructivist paradigm would say that during ;requent pauses learners are reflecting on what they have read, for example, trying to construct meaning by relating it to their prior knowledge. Depth-of-processing theories also seem to support journal writing, whether the focus be Ross's (1981) importance of number of decisions, Jacoby, Craik and Begg's (1979) decision difficulty, or Johnson-Laird and colleagues' (1978a, 1978b) amo.unt of processing time. Furthermore, as Prawat (1989) notes, "there is a growing body of literature suggestive of how writing can promote 9

PAGE 23

understanding of subject matter content (Britton, Burgess, Martin, McLeod, & I Rosen, 1975; Tchudi & Tchudi, 1983)" (p. 14). 3. Generative mocieVconstructivist of instruction. Requiring students to I write jourilals is in keeping with a generative model of instruction; it places much of the locu.s of control for learning in the student and encourages the student to interact with the text (see, e.g., Palincsar & Brown, 1984). Wittrock (1974) heralded the paradigmatic "shift toward cognitivism, toward reinstating the learner I and his cognitive states and information processing strategies, as a primary I determiner of learning with understanding and long-term memory" (p. 87; emphasis added). Journal writing is also consonant with constructivist explanations of learning, which are rapidly replacing information-processing models (e.g., Clancey, 1992; Duffy & 1992; Jonassen, 1991a). The generative model oflearning has instructional origins, while the constructivist paradigm has strong philosophical roots communication, D. H. Jonassen, August 1991). 4. lndiyidual-iearner differences: ability. Journals allow the students to respond to a literary wm:k at their own levels of understanding. When they initially respond to a work, they are neither overwhelmed nor bored by the questions typically found in literary texts which generally seem too sophisticated or too trivial for any given student. the first reading of a work, high-ability students might be able to tackle some of the most sophisticated questions (e.g., What does the f'rrst-person point of view contribute to the meaning and effect of the story?), while low-ability students might ask questions which clarify the facts (What did Fortunato do to 10

PAGE 24

deserve to be killed, in "The Cask of Amontillado"?), paraphrase segments, the plot, or rely on "elaboration" strategies, trying to relate the content ,, to familiar experiences. The cognitive structures, strategies, and working-memory capacity of the low-ability students might not be adequate to engage in higher-level processing: during the earlier stages of study. The low-level processing activities that reseaich suggests we will find with these students might be prerequisite to the development of deeper processing at later stages of study. At the least, asking or comment:iAg at their individually appropriate levels (Miyake & Norman, 1979) 1' 1 might be expected to benefit students more than study alone (Frase & Schwartz, 1975). 5. Indiyidua}Jiearner differences: prior Miyake and Norman's (1979) I' study and other research on individual-learner differences suggest that the levels of questions in students' journals will vary, depending on the expertise of the student StUdents with little general knowledge and/or little expertise in literature might tend 'to summarize, ask low-level questions (e.g., Why is death wearing white in "The Appointment in Samarra"? Did the servant know Miss Emily had killed Homer in "A Rose for Emily"?), or impose their value systems on the characters. 'Students with extensive general knowledge or an extensive backgroun&in literature might tend to have few questions on the easy works (and focus on theme or respond with critical analyses), but to ask incisive questions about difficUlt works. Middle-range students might ask middle-level questions for 1' easy to moqerate works (e.g., Who is the you that Montresor is speaking to in "The Cask of Amontillado"?) but respond like novices on difficult stories. 11

PAGE 25

6. and scaffolding the development of schemata. Journals provide a Plechanism for the teacher (a) to identify problems individual students encounter in interpreting the text arid (b) to provide feedback in a timely and appropriate manner (correcting the student if need be, answering questions and encouraging extensions of the student's questions and comments). Class discussion is a totally inadequate vehicle for most students to communicate their understanclings of the work under discussion. a check on the students' encodings of the stories is essential in terms of the generative learning model and schema both of which stress that learners interpret new experiences in light of their prior knowledge (e.g., R. C. Anderson, Reynolds, Schallert, & Goetz, 1977; Jonassen, 1985; Mayer, 1975, 1979, 1980; Rumelhart & Ortony, 1977; Wittrock, 1974). Providing a check is even more critical in the constructivist paradigm ofstudents' learning: "the internal neural processes [are] constantly changing and always freshly created (albeit it out of previous activations)" (Clancey, 1992, p. 153). Because students frequently lack adequate knowledge of allusions (as well as other general knowledge), they impose inaccurate schemata on them. Journals provide insight into the misconceptions, as well as the exceptional insights, that students generate from their reading. The journals have provided me an opportunity to identify unique misconceptions as well as strong opinions and personal experiences that students otherwise would probably not express. In 20 I years of teaching without journals, I rarely heard students express most of the questions and comments that they express in journals (e.g., "Oedipus the King is 12

PAGE 26

trash [because it deals with the subject of incest]." "Why would the whole town [in Faulkner's "A Rose for Emily"] tum out for the funeral of a black woman?" '"Until the very end, I thought Montresor [in Edgar Allan Poe's "The Cask of Amontillado"] was just planning a big joke on Fortunato; I didn't really believe he was going to kill him." Sometimes students elaborate, connecting events in stories to their own lives, such as their own suicidal tendencies or even suicide attempts, or abortions). Know'ing students' questions and comments allows a teacher to correct many that a teacher, as an "expert" reader, is unable to predict less expert and students have. According to traditional schema theory, if the students have activated the wrong schemata, they will not be able to correctly assimilate the new information presented during class discussions. Spiro and his colleagues' (e.g., Spiro, Coulson, Feltovich, & Anderson, 1988; Spiro & Jehng, 1990) concept of schema assembly offers a more dynamic explanation of schema theory for ill-structured domains (such as literature). Learners don't merely activate schemata, they create new schemata by combining relevant aspects of many ,, schemata (this is consonant with Clancey's [1992] description, above). Knowing students' rrlisinterpretations also allows the teacher to emphasize important reading strategies, such as looking up words they do not know rather than guessing from context and looking up terms they are sure they know but which puzzle them in the story. 7. Learners' dispositions. Research suggests that there are two types of learners: performance-oriented and mastery learners. Performance-oriented learners focus 13

PAGE 27

on surface material, that is, facts or disconnected information; they are concerned with "get[ting] the job done as painlessly as possible, with learning serving as a means to an end" (Prawat, 1989, p. 33). Mastery learners use deeplevel processing; they want "to increase competence, to become more about or skillful at something" (Prawat, 1989, p. 33). However, performance-oriented learners may become mastery learners if their study strategies facilitate learning (Prawat, 1989). In keeping with learners' dispositiol'ls and schema theory, Shuell (1988) emphasizes that successful learners must play active role in learning. Thus, by encouraging the students to interact with a literary work at their own levels, and by providing appropriate feedback, journal-writing should facilitate mastery learning. 8. Implications of research on adjunct questions. Research on adjunct questions has demonstrated the facilitative effects on learning of conceptual prequestions and frequent questioning during reading. (Research on learner-generated questions has not these issues.) Furthermore, Rickards (1976) has theorized that conceptual prequestions induced readers to "derive a relevant schema for passage information" (p. 217). Since one might expect journal-writing to encourage at least mastery learners to generate and focus on conceptual prequestions and to ask frequent q*stions, journal writing should encourage them to derive a relevant schema for:ilie story. Readers of a complex, reflective short story (perhaps even ,. performance learners) might be expected to begin with at least one conceptual prequestion: What is the meaning of this story? Then one might predict that they will try to construct the meaning of the story as they read, asking questions and 14

PAGE 28

. 1 generating self-explanations which will shed light on the overall question as they I proceed .. Unless they are performance-oriented, they will also ask other questions, such as, How do the parts achieve the meaning? Is the story effective? Would the story be more effective if various parts were changed? As the students ask themselves questions relevant to these issues, they might "be more likely to notice situations where they need further clarification" (Stein & Bransford, 1979, p. 775), for flagging points of confusion. 9. De.pth of processing. Several explanations of depth of processing seem to be i: II consonant with journal writing. Jacoby, Craik, and Begg' s (1979) hypothesis seems to have bearing on journal writing. Requiring !Students to write journals would seem to increase the difficulty of the I task of reading/studying a literary work and thus result in higher levels of retention. The focus of Johnson-Laird and his colleagues (1978a, 1978b) on the amount of processing time also has bearing, since writing a journal about a literary I' work will require more processing time than merely reading the work. And Ross's (19,81) focus on number of decisions seems relevant, since each journal I I entry will potentially reflect at least one decision about the work. Fisk and Schneider (1984) hypothesized that "[c]ontrolled processing [the type most I learners wQuld have to do in reading a reflective story] is characterized as effortful, I slow, serial, and capacity limited" (p. 196). Writing a journal would seem to scaffold in focusing their attention and reducing the burden on working memory. 15

PAGE 29

As I shown, prior research has failed to address several critical issues which suggest;.that journal writing might serve as an effective instructional/learning tool in either an in(ormation-processing model or a constructivist paradigm of learning. Next I summaiize the specific questions which this study addressed. Research Questions for a Study to Deyelo.p a Cowitive Model of Journal in a Collece Introduction to Literature Course Anecdotal .reports often attest to the value of having students respond in journals to works they reatl (e.g., Duke, 1982; Flitterman-King, 1988; Nugent & Nugent, 1985; Weiner, 1986).: My own instructional experiences with journal writing support such reports. But no one has studied the cognitive basis of journal writing. In order to optimize the benefits of journal writing assignments for diverse students, it is first ,; necessary to develop a cognitive model of journal writing as a learning and study strategy. Pw;pose of the Study This study examined journal writing as a learning and study strategy of college students in the ecologically valid setting of an introduction to literature class. The study focused on use of journals during the first half of the course, in the unit on complex reflective short stories. This study addressed several questions: 1. What metac.ognitive and cognitive activities do students engage in when they write their journa.s? li 2. How are these activities related to individual learner characteristics (e.g., age, reading comprehension level, prior knowledge of literature, students' evaluations of journal writing as a learning strategy, grade in the course)? 16

PAGE 30

3. What types of questions and comments do students write in their journals? I 4. How are questions and comments related to individual learner characteristics? 5. Can a taxonomy describe students' questions and comments? Are there concepts or ways of seeing that students have in common? 6. How do perceive journal writing (e.g., Do they find it helpful? Why or why not? What do they find most helpful and least helpful? What do they have for making journal writing a more helpful experience?) 7. How do ratings of journal helpfulness relate to other variables? 8. How do students rate the job aids ("Students' Accounts of Journal Writing" and the "Catechism for Stories")? 9. How do ratings of the job aids relate to other variables? 10. Is there a cognitive model that can capture the processing of most students' journal-writing activity? 17

PAGE 31

CHAPTER2 REVIEW OF RELATED LITERATURE This study focused on the use of journals in the study of fiction in a college-level introduction to literature course (an ecologically valid setting and an ill-defined domain). Although many teachers direct students to focus on "gut" reactions (Duke, 1982; McCormick, 1985; Nugent & Nugent, 1985), others emphasize cognitive processing, directing students' responses with adjunct questions or soliciting learner generated questions and.comments (Flittelman-King, 1988). This study focused on cognitive activities. But to understand this focus, one must put it into the perspective of the earlier research on adjunct questions and self-explanations, schema theory, and the paradigmatic shift from mathemagenic instruction to generative learning. Adjunct questions direct the learner's processing of material; they precede, are embedded in, or follow text. A self-explanation is a comment generated in the process of learning something new; going beyond what is stated. Self-explanations are generated in the context of learning something new, unlike elaborations, which "generally refer to the use of existing knowledge to embed or embellish a piece of information in a larger context so that it is more memorable" (Chi & VanLehn, 1991, pp. 71-72). theory assumes that learners interpret new experiences in light of their prior knowledge. In mathemagenic instruction, the locus of control is external to the learner (e.g.,' in the teacher or text); the instruction seeks to control the information processing activities of the learner (Jonassen, 1988), and thus to "give birth to learning" (Rothkopf, 1970). Generative learning places the locus of control in the 18

PAGE 32

learner, who generates (constructs) meaning from new experiences in light of prior knowledge (e.'g., Jonassen, 1988; Wittrock, 1974). Thus, research on journal writing must address the relevant literature on journal writing, adjunct questions, self explanations, schema theory, and the generative model of learning (and the constructivist paradigm). This chapter therefore addresses each in turn. Journal Writing In spite of 1the relative dearth of empirical support, almost without question educators writing as an effective tool for learning (Anson & Beach, 1990). A plethora of publications describe teachers' experiences with journal writing, including, for example, their "trials and errors" (e.g., Heath, 1988; Holland, 1989), goals, strategies, and methods for incorporating them into courses (e.g., Bauso, 1988; Jolley & 'Mitchell, 1990; Myers, 1988; Roth, 1985), and effects on student engagement with assignments (e.g., Roth, 1985; Wilson, 1989). This approach is in keeping with the orientation of many humanities teachers and professors who follow a criticism model'of scholarly inquiry. However, advocates of journal writing too often describe their experiences as if they were generalizable instructional strategies. For example, Wilson (1989) reported that when literature students were instructed to "reflect, ruminate, and question," to "listen carefully to yourself and attempt to describe the effect the book is having on you" encouraged to "respond deeply, write honestly" in their journals, they generally (a) admitted confusion, (b) asked questions, (c) made inferences in trying to answer'their questions, (c) often noticed words or phrases that the teacher 19

PAGE 33

would not have pointed out, (d) identified with characters or the author, (e) made I' predictions, and (f) looked for evidence to support their opinions. Every literature teacher would,surely welcome such behavior, but would these findings generalize to other 11th-grade students and teachers, to traditional and non traditional college students, and tp .other domains? What other factors (e.g., grading incentive and wording of feepback) contributed to the results? Although Wilson (1989) provided insightful examples of the types of journal I entries wrote, she did not study the relationships of important factors such as content and individual learning characteristics (particularly reading ability) and students' entries. In a descriptive article of journal writing in a college literature survey course, Holland (1989} assumed that maintaining the anonymity of journal writers until the end of the term students to take risks and to ask questions in their journals. His decision to assign anonymous journals was a response to "disappointing results" he obtained in terms when he had not graded signed journals (p. 236). However, as he was aware when he decided on "anonymity," his failure to provide any external incentive in his earlier use of journals may have contributed to those disappointing results (see Hicken, Sullivan, & Klein [1992] on the role of incentives). Since anonymity restricts the information that the teacher has to draw upon in responding to needs of each student during the term, it is worthwhile to determine if it is a necessary criterion of productive journal writing. Can a supportive environment, combined with signed journals and an appropriate incentive, such as grading, produce the positive results he obtained? 20

PAGE 34

With sucq a small body of research on journal writing, researchers are still struggling to identify the issues they must address in the design of their studies. A ware of the general of empirical support for the use of writing as a learning tool, Anson and Beach undertook a series of studies. They began by investigating a claim by Fulwiler (1989) that students who "adopt a more informal, exploratory stance are more likely to formulate the course material 'in their own words,' enhancing their understanding" (Anson & Beach, 1990, p. 3). Thus they tried "to characterize as precisely as those features which [they] could recognize across students' academic jourrials" (Anson & Beach, 1990, p. 3). They studied the relationship between the linguistic features of journals and performance in a college linguistics class. They rated students' journal entries on six 4-point rating scales: formal/informal, objective/affective, elaborated/unelaborated, predetermined/unfolding, externally defined/internaUy constructed knowledge, and negative/positive effect on the rater. The only rating tied .to fmal grade was a relatively low correlation with degree of elaboration. Anson and Beach (1990) inferred that, regardless of their entering conceptions of the journal form, students appeared to "make use of whatever features [were] personally most meaningful" in a given context (p. 8). Unfortunately, Anson and Beach made journal writing an artificial task (one not related to studeqts' or teachers' needs), provided an inappropriate criterion for journal writing, and provided no guidance in journal writing. The only announced criterion was quantity of Writing. Students received no feedback to what they had written; if students had to ensure they understood the course material, they would have had to ask (apart 'from their journals). And since the journals were read only at the end 21

PAGE 35

of the tenn, the instructor could not use journal content to monitor students' understanding: of the material and adapt the instruction, a benefit many other writers noted (see beltiw). Thus journal writing would probably have seemed an artificial task to many students. Unless their study strategies already included journal writing or they were mastery learners (see Glossary), students may have treated the assignment as busy work, an implicit goal of wordiness. Furthermore, Anson and Beach concluded that "the kinds of thinking encouraged by journal writing had little to do with the students' learning as measured on the objective tests" (p. 8). Unlike Feathers and White (1987), they failed to consider that the content of students, journals may not fully reflect the cognitive and metacognitive activities that t:q.ey engage in. Hynd and (1990) studied the effects of annotating versus journal writing on multiple-choice. and essay test performance in college study skills classes. They found that students to annotate a novel significantly outperformed students who wrote journals on the multiple-choice test but not the essay test. However, their study had several limitations. First, as they noted, students were minimally competent in annotation or jo:umat writing after training; motivation was low (the objectives of the experiment were not tied to the fmal examination which students were required to pass); and the task was probably too easy (students said the excerpt was not very demanding). Although they recommended that a more demanding passage be used in the future, they did not observe the greater benefit of using several passages with varying difficulty levels to study riot only task difficulty but also individual learner differences. 22

PAGE 36

Equally iffiportant are limitations Hynd and Chase did not note. First, they did not include a control group. Second, although they detennined that there were no statistically significant differences among the groups on basic skills, they did not study the relationship of perforinance to any individual learning characteristics (reading ability is particularly relevant in this instance). Most importantly, students in the groups were taught to focus on different criteria in studying the excerpt. The annotation groups were taught to mark and annotate vocabulary, character development, important events, theme statements, use of literary devices like figurative language and symbOlism, and to make inferences about relationships between ideas ... because of the relation to the types of items commonly found on tests about novels that were given in the study strategies courses. (p. 47) The journal groups were taught to write journals exploring the environment for their reading, their vocaoulary and comprehension difficulties, the strategies they used for overcoming these difficulties, what they thought was memorable about their reading, and their opinions about what they had read ... because they are often required in classes where journals are used (p. 47) Class discussion addressed only the issues the students raised. Thus both groups did not address the same objectives in their study strategies or in class. Nevertheless, Hynd and Chase predicted that students who are taught to annotate as well as to write journals would perform better overall. They did not consider the possibility that the journal students would outperform the annotation students if they were taught to write journals which helped them monitor their comprehension of the same elements annotators were taught to attend to. Even excellent research suggests avenues for further exploration. Feathers and White (1987) studied students' development of metacognitive awareness in a college 23

PAGE 37

developmental, reading class. They conducted a case study of six students in a college reading class. Students entering the course "typically did not take responsibility for comprehending a text, were unaware of the factors that affect and had few, if any, reading or learning strategies" (p. 266). The research suggested that students developed metacognitive awareness of reading and learning during the three-month course. The study is, of course, inherently limited by its focus on low-ability readers and its failure to examine any other individual learning characteristics.. Several wnters report that journal writing enabled students to relate course content to their experiences (e.g., Hettich, 1990, college psychology; Roth, 1985, college sociology; Wilson, 1989, high school literature). Many have found that students writing to be helpful (e.g., Cole, 1991, CQllege literature; Cummins, 1989, college research writing; Hettich, 1990, college psychology; Rumpf, 1988, college engineering; Schwartz, 1989, college oral communication). In a research study, Wason-Ellam (1987) found that frrst-grade students, when asked motivating questions about what they learned in mathematics, used their journals for organizing information, assimilating and accommodating information, arid making (note the similarity to Wilson's description of the content of 11th-grade literature journals; see above). Linn (1987) found that journal writing provided benefits for both high school geometry students and their teacher. Journal writing enhanced the students' metacognitive aoility, forcing them to become actively involved in learning, making them aware of what they did not know, and helping them identify and take advantage of 24

PAGE 38

the strengths in their individual learning styles. Students' journals provided a conununicatio,-. link with the teacher, who used it as a tool for monitoring students' learning and adapting instruction. Similar teacher benefits have been identified by many others (e.g., Cole, 1991, college literature; Holland, 1989, college literature; Sublett, 1988, college geography; Tamove, 1988, college mass media). Several writers report that students believed that journal writing affected their critical thinking skills (e.g., Bauso, 1988, college literature; Cole, 1991, college literature; Crismore, 1987, college basic writing). But with so little formal research on journal writing; alleged benefits must be verified, and research findings must be explored in other contexts to determine their transferability or generalizability. Several additional issues are implicit in the literature. For example, what type(s) of grading/feedback systems are effective, for what purposes, and under what conditions? Some writers have advocated that journals not be graded; others (e.g., Holland, 1989, college literature; Roth, 1985, college sociology) have said they found grading a necessary incentive. Still others have identified graded journal writing as an incentive for students to read assignments and reflect on them before coming to class (e.g., Bauso, 1988, college literature; Schwartz, 1989, college oral conununication) and a means of increa,sing students' confidence in their understanding of the domain (e.g., Zuercher, 1989, professional writing). Advice on grading and feedback abound (e.g., Bauso, 1988, college literature; Myers, 1988, high school literature; Roth, 1985, college sociology), yet no one has provided empirical support for particular reconunendations. 25

PAGE 39

i As noted above, Holland's (1989) article raised the issue of whether the teacher should know the journal-writer's identity during the course. Equally important is another issue implicit in the literature: the effect of other students' knowing the content of a journal. What effects do group journal writing (e.g., O'Sullivan, 1987, college literature) and1having students share their journals with other students (e.g., Fulwiler, 1989, college llterature; Heath, 1988; Wilson, 1989, high school literature) have on journal content, students' beliefs about the helpfulness of journal writing, and individual achievement? Another question implicit in the literature is What are the defining attributes of journal writing? What kinds of constraints can or should an instructor place on assignments which are called journal writing? At one extreme, some advocates allow the writer total :freedom in detennining content. Some (e.g., Gowen, 1985) advocate I that the teacher never see the journals, only students' responses to their own journals at the end of the At the other extreme, some treat journals essentially as vehicles to respond to adjunct questions (e.g., identifying previous journal entries on Bloom's taxonomy; Hett:j.ch, 1990) or to write dialectical reasoning entries (Jolley & Mitchell, 1990). Are there differential benefits in these approaches (e.g., Do they interfere with high-ability students' successful strategies)? Finally, the literature provides no model of the cognitive and metacognitive activities involved in journal writing which is grounded in an analysis of specific journals, although several writers have provided cognitive explanations of journal writing (see Hedlund, Furst, & Foley, 1989, for a comprehensive explanation). Nor does the literature provide taxonomies of journal content in many disciplines. Anson 26

PAGE 40

and Beach (1990) developed a simple five-category taxonomy to describe the discursive characteristics of journals in a college linguistics class; they included a sixth category to classify "the subjective responses of journal raters to the overall 'quality' of the entry" (p. 6). Hettich ( 1990) required his students to apply Bloom et al. 's ( 1956) taxonomy to their earlier journal entries. AQiunct Questions The adjunct question literature suggests several relevant concerns for journal writing. For example, do findings related to question position, question level and task difficulty, and relationship of levels of questions to levels of processing also apply to learner-generated questions and self-explanations? Following is a brief review of the findings on adjunct questions. Extensive quantitative research on the use of adjunct questions has culminated in an interest in learner-generated questions and self-explanations. A growing body of quantitative on learner-generated questions and by schema theory (R. C. Anderson, 1977; Rumelhart & Ortony, 1977), the generative model ofleaming (Jonassen, 1985; Wittrock, 1974), and the constructivist paradigm (e.g., Clancey, 1992; Duffy & Jonassen, 1992; Jonassen, 1991a}--supports their use as learning/study aids (M. E. D. A. Andre & Anderson, 1978-79; Frase & Schwartz, 1975; Singer &Donlan, 1982). However, research has often been conducted in laboratories rather than more ecologically valid settings, and has focused on learning in well-defmed domains. In the first part of this chapter I reviewed the literature on journal writing ."In the remaining part, I review the literature on adjunct questions, self27

PAGE 41

explanations, schema theory, and the generative model of learning (and the constructivist f>aradigm). Question Position ''The adjubct use of questions in text is the best established, most popular, and most heavily researched of the mathemagenic techniques" (Jonassen, 1985, p. 128). Researchers examined the effects of adjunct questions in terms of (a) question position (e.g., text, after the complete text, after sections of the text, or adjacent to the text in margin, (b) frequency, (c) response mode (e.g., multiple choice versus constructed response), (d) level of question-verbatim versus higher level, (e) relevance to criterion items, (f) nature and difficulty of the orienting task, (g) individual learner differences, such as aptitude, prior knowledge, and motivation, and (h) the interaction of those factors. The diversity of factors used in studies makes it difficult to draw consistent conclusions. For example, verbatim questions, which were used in many studies, tYPically yield different results than do higher-level questions, particularly coupled with changes in position in the text A few studies are discussed to illustrate the complexity of the issue. ,. Generally, postquestions have been found to facilitate learning more than prequestions do (R. C. Anderson & Biddle, 1975; Lindner & Rickards, 1985); in fact, prequestions have been found "more frequently [to] inhibit than facilitate performance on new criteriort test items" (R. C. Anderson & Biddle, 1975, p. 93). Rickards (1976) ,. demonstrated superiority of verbatim postquestions over verbatim prequestions on I recall tests. However, he found an interaction between level of question, question position, and of retention. On a test of immediate recall he found verbatim 28

PAGE 42

postquestions and conceptual postquestions to be equally facilitative, but on a test of long-term retelltion, he found that conceptual prequestions produced greater verbatim recall than did: conceptual postquestions. He theorized that the conceptual prequestions induced readers to "derive a relevant schema for passage information" (p. 217). While Rickards (1976) used a recall test, Felker and Dapra (1975) used a multiple choice test and' obtained different results. They found conceptual postquestions superior to conpeptual prequestions for learning to identify examples of concepts named and illustrated in the text One explanation of the discrepancy between their findings and those of Rickards is the difference in test mode. However, this seems unlikely, since an earlier study by Watts and Anderson (1971) also demonstrated that conceptual prequestions facilitate performance more than do conceptual postquestions, on a multiple-choice test of repeated questions as well as new application questions. Question Levellfask Difficulty Another explanation of the difference between the findings of Felker and Dapra (1975) and those. of Watts and Anderson (1971) and Rickards (1976) may lie in the level of difficulty of the task at the time of encoding. Several studies have demonstrated that an increase in task difficulty results in greater levels of recall than do less difficult tasks; that is, tasks requiring "more extensive processing in the semantic domain" result in greater recall than tasks requiring less semantic processing (Glover, Plake, & Palmere, 1981, p. 742; see also, Glover, Plake, & Zimmer, 1982; Craik:, & Begg, 1979). Rickards' (1976) subjects were required to infer. concepts, using prior knowledge, while Felker and Dapra's (1975) subjects merely concepts that were named and illustrated in the experimental 29

PAGE 43

passages. Rickards' subjects were forced to draw on their prior knowledge for the concept, a related explanation is also implied by the Glover, Plake, Roberts, Zimmer, and Palrnere (1981) study, which found that adjunct aids which require the reader's schema to interact with the text base have facilitative effects on recall. This interpretation is supported by research on the consistency of organizers and text. Mannes Kintsch (1987) found that consistency between organizers and text facilitated and remembering, while inconsistency facilitated inference and problem When the learners had to deal with inconsistency, they apparently had to draw ontheir prior knowledge to help them construct meaning. Salomon (1979) reported similar results in a study of randomly organized versus logically sequenced film. Several researchers have also demonstrated the superiority of higher-level questions. Beqton, Glover, and Bruning (1983) noted that "Andre has demonstrated that 'paraphrase questions are superior to verbatim questions in the learning of sentences' (f. Andre & Sola, 1976), and in the learning of text materials (f. Andre, 1979; T. Andre & Womack, 1978)" (p. 389). Glover, Plake, Roberts, Zimmer, and Palrnere (1981).demonstrated a significant positive effect on free recall when questions required to generate inferences from text. Friedman and Rickards (1981) found positive effects increasing as the questions progressed from low to higher levels. On a test of bo$ direct and indirect learning, paraphrase questions were superior to verbatim questions, and inference questions were superior to paraphrase questions. ,. 30

PAGE 44

Additional Findings on Adjunct Questions The apparent interaction of test type, task difficulty, learners' prior know ledge, and other factors iildicates the need for great caution in interpreting findings for future research. Following is a summary of the more consistent findings on the use of adjunct questions (in apdition to the studies cited above, seeR. C. Anderson & Biddle, 1975, and Lindner & Rickards, 1985). The closer the adjunct question is to infonnation it asks about, the higher the test perfonnance is on similar criterion items. The effects greater for short-answer tests than for multiple-choice tests, on both new and repeated criterion items. One explanation for difference in effects is that "adjunct questions primarily act on the retrievability of information .... Another possibility is !that short-answer and multiple-choice questions make different processing demands when inserted in text, thereby differentially affecting study activities" (R. C. Anderson & Biddle, 1975, p. 98). Compared to.verbatim questions, higher-level questions facilitate greater recall of (a) both new and.repeated criterion items and (b) both factual and higher-level material. "[H]igher-leyel questions [particularly prequestions, as noted by Rickards, 1976] affect not only the level at which material is processed but the manner in which such material is organized in memory" (Lindner & Rickards, 1985, p. 139). Rickards (1976) suggests that conceptual prequestions induce readers to derive relevant schemata. Facilitative effects are a function of relevancy of questions to the material (McGraw & Grotelueschen, 1972, and Rothkopf, 1972, as reported in R. C. Anderson, & 31

PAGE 45

Biddle, 1975); Stein and Bransford (1979) suggest this effect is related to readers' "asking abotit the potential significance or relevance of facts" and to readers' noticing "situations where they need further clarification" (p. 775). Would these findings on a mathemagenic instructional strategy apply in a constructivist learning environment-that is, when the learner controls the topic and level of difficulty of his/her questions and self-explanations? Before examining the limitations of studies, I will briefly describe the research on self-explanations, which are essentially answers to implied questions. Then I will examine the limitations of both strands of research for journal writing. Self-Explanations Chi and her colleagues (e.g., Chi & Bassock, 1989; Chi, Bassock, Lewis, Reimann, & Glaser, 1989; Chi, de Leeuw, Chiu, & LaVancher, 1991; Chi & VanLehn, 1991) studied think-aloud protocols of students' generating self-explanations of physics examples. They found that the amount learned was related to the number of self-explanations generated (Chi, Bassok, Lewis, Reimann, & Glaser, 1989). They also found that help students construct more complete understandings of the domain principles and concepts, that much of what students learned related to technical procedures for solving physics problems, and that good problem-solvers generated more. self-explanations than poor solvers (Chi & VanLehn, 1991). Chi, de Leeuw, Chiu, and LaVancher (1991) found that generating self-explanations benefits both average and high-ability eighth-grade students equally (there were no low-ability students in the study). 32

PAGE 46

Limitations of Research on Adjunct Questions and Self-Explanations Although:research on adjunct questions has provided much insight into the topic, it has largely ignored the mediating role of individual learner characteristics (Lindner & Rickards, 198$). Furthermore, it has focused on low-level questions, multiple-choice tests, and what Lindner and Rickards (1985) term "the artificiality of the adjunct question paradigm" (p. 145) because "the limitations placed on the learner in the adjunct question studies in no way resemble the normal reading and/or study behavior of students" (p. 146). The artificiality of the studies seems particularly obvious when one observes students in natural conditions employing the suggested strategy of skimming material before they read it or "performance-oriented" students merely looking ahead to see what quest:i,ons they must answer, for both types of students postquestions immediately t>ecome prequestions. Moreover, the research on adjunct questions has generally ignoFed the learning strategies students engage in with others (e.g., other students, teacher, or family) when they are studying. Stein and Bransford (1979) criticized the focus of adjunct-question research on facts, rather than on understanding the "potential significance or relevance of facts" (p. 775). Their criticism is related to criticism of instruction for its focus on facts rather than meaningful learning (e.g., R. C. Anderson, 1977; Bransford, Sherwood, & Hasselbring, 1988; Perkins, 1986). Thus, many issues related to adjunct questions remain to be explored. Although several studies claim to have demonstrated the superiority of higher-level questions over verbatim questions in memorability, Jonassen (1985) highlighted "our 33

PAGE 47

inability to define [adjunct] questions at various levels" (p. 128). Lindner and Rickards (1985) said that the cognitive processing level is [p ]erhaps the most important issue related to adjunct question research and practice .... What level of processing is necessary to produce knowledge or understanding of passage information? ... Little or no research has compared various levels of meaningful processing, whereas it is such comparisons that are educationally relevant (p. 138) Unlike the research on adjunct questions, the research on self-explanations has directly addressed several critical issues, including the mediating role of individual learner characteristics and learner-controlled levels of cognitive processing. The studies on self-explanations had two limitations particularly critical for using journal writing as a strategy for comprehending reflective short stories. First, ecological validity was limited by the of think-aloud protocols. Second, solving physics problems is a well-defined domain, one that is very procedurally oriented, while comprehending reflective short stories is an ill-defmed domain. Schema Theory Relationship of Leyels of Questions to Levels of Processing Many researchers have discussed levels of questions (e.g., factual, paraphrase, conceptual, problem-solving) in terms of depth of thinking or depth of processing (see chapter 1 and below). Are these terms synonymous and to what cognitive activities do they refer? Researchers and theorists have variously defined these activities in terms of amount of processing time, number of decisions, type of mental activity (such as rehearsing, elaborating, paraphrasing, drawing inferences, evaluating, problemsolving), some other factor, or an interaction between several factors. 34

PAGE 48

Schema theory (e.g., R. C. Anderson, 1977; Rumelhart & Ortony, 1977) provides I a further connection between levels of questions/thinking and levels of processing. A schema is a stiJictured cluster of knowledge which is embedded within a semantic network. Rumelhart and Ortony (1977) explained that the "process of activating related schemata is akin to Craik: and Lockhart's (1972) notion of depth of processing" (p. 105). "At deeper levels the subject can make more use of learned cognitive structures so that the item will become more complex and semantic" ( Craik & Lockhart, 1972, p. 679). Recent discussions of schemata emphasize that they are not stored mental I, structures but dynamic networks (e.g., Clancey, 1992). Definitions of Ipeyels of Processin& Craik and Lockhart (1972) conceived of depth of processing in terms of "elaboration," which often means "the addition of further information, so that the trace becomes richer and more detailed" (Jacoby, Craik, & Begg, 1979, p. 597). Craik and Tulving (1975):;introduced the notion of "distinctiveness of encoding," which emphasizes "the contrastive value of information in the trace" (Jacoby, Craik, & Begg, 1979, p. 597). The distinctiveness-of-encoding hypothesis suggests that memorability is a function of the difficulty of decisions the learner must engage in; that is, "an increase in task, difficulty ... results in higher levels of retention" (Jacoby, Craik, and I Begg, 1979, p. 586). As reported by Ross (1981), Johnson-Laird and his colleagues (1978a, 1978b),' have focused on amount of processing time, while Ross (1981) has focused on the number of decisions involved. Although several explanations of depth-of-processing have been researched, there remains a need to investigate the potential interactions of those and other factors. 35

PAGE 49

According to traditional schema theory (e.g., R. C. Anderson, 1977; Rumelhart & Ortony, when learners encounter new infonnation, they will potentially activate existing schemata, try to fit the new material into the old schemata, refine existing schemata andl?r build new schemata. Thus these cognitive activities would potentially seem to (a) involve differentiating the new infonnation from elements in existing schemata (distinctiveness of encoding), (b) result in many decisions, (c) be quite difficult, (d) require much processing time, and (e) perhaps result in elaborations which make the new material more memorable. (R. C. Anderson [1977] hypothesized that when people's beliefs are tlrreatened, they may segregate and retain logically inconsistent views in order to protect the ones they believe.) Which of these cognitive activities occW:(s) might depend on the particular information to be learned, the criticality of the information/task (a factor apparently not yet studied), prior knowledge, and other indiv,idual differences. These factors are, of course, difficult to isolate, but may be reflected in the comments and questions in students' journals. The Generative Learnins Model The importance of journal writing as a learning and study strategy is also derived from a generative model of learning. The emerging interest in learner-generated questions and self-explanations has followed closely on the heels of the paradigmatic I shift in psychology from behaviorism to constructivism (Bruning, 1983; Jonassen, 1991a; Resnick, 1983) and the accompanying shift from mathemagenic instruction, which "seek[s] to control the infonnation processing activities of the learner" (Jonassen, 1985, p. 127), to generative learning, which places the locus of control in 36 L .. -------------------------------------------------------

PAGE 50

the learner. The generative learning model reflects what many educators and researchers view as a desirable (perhaps the primary) aim of instruction: to encourage learners to become independent by learning how to learn (e.g., Brown, Campione, & Day, 1981; Frase & Schwartz, 1975; R. M. Gagne, 1980; Jonassen, 1985). '"The generative model asserts that learners, when faced with stimuli ... ,construct and assign meaning to that information based upon prior learning" (Jonassen, 1985, p. 11). The emphasis on the active role of the learner has bearing on many current focuses in instructional theory and research, from problem-solving and schema theory, which explains comprehension in terms of the interaction between the reader and the text (R. C. Anderson, 1977; Rumelhart & Ortony, 1977), to the views of transactional analysis which is applied so well to the study of literature by Rosenblatt (1978). Learners' Diwositions and Goals The generative model also has bearing on the emerging interest in the dispositions of learners and 1their learning goals. Prawat (1989) described two types of learners: performance-oriented learners and mastery learners (see Glossary). According to Prawat, when learners become meaningfully engaged in learning and can make the necessary connections between the elements under study, they may become mastery learners; on the other hand, if mastery learners find a strategy ineffective, they may adopt lower-level strategies typical of performance-oriented learners. (For related discussion of surface versus deep-level processing see Marton & Saljo, 1976a, 1976b, and Watkins, 1983.) Bereiter and his colleagues (e.g., Bereiter & Scardamalia, 1989; Ng & Bereiter, 1991) have focused on the role of learners' goals. Bereiter and Scardamalia (1989) 37

PAGE 51

described two: ,different learning goals in public schools. Learners with work orientations, the goals that schools typically cultivate, view schoolwork merely as tasks I to be completed (ends in themselves), with learning being incidental. Learners who engaged in intentional/earning view schoolwork as a means to an end; they pursue cognitive "over and above the requirements of tasks" (p. 385). I' Ng and Breiter (1991) examined the relationship between motivation and learning goal orientation. They identified three levels of goal orientation among adult learners who were pres';lmably equally motivated: task-completion goals, instructional goals, and personal-lq10wledge-building goals. The learners with task-completion goals focused on completion of assigned tasks, while those with instructional goals focused on the content pf instruction. By contrast, the personal-knowledge builders, in addition I to performing better on a posttest, "responded more often [than the others] to learning goal cues than task goal cues. They actively related new learning to prior knowledge and they posed and tried to solve problems and questions" (p. 243). While goal orientation and achievement were not related to the students' levels of education and prior computer ,experience, they were "positively related to previous experience of independent leJring" (p. 243). However, rather than focusing on strategies which will promote learning, Ng and Bereiter concluded: If are educational approaches that target [personal-knowledge building], the m6st likely candidates are ones that operate at a social rather than an indiviQ.uallevel----ones that promote a "community of learners" (Brown & Campjone, 1990) or a "knowledge-building community" (Scardamalia & in press). (p. 269) Thus, motivation, goals, strategies, and their interactions remain a concern of both researchers and. instructional developers. 38

PAGE 52

Focus of Research on Learner-&enerated Questions 'The generative counterpart to adjunct questions [which are mathemagenic aids] requires learners to generate their own questions" (Jonassen, 1985, p. 27). Thus, all of the issues related to adjunct questions logically extend to learner-generated questions; yet by the very nature of the generative model, these issues cannot be as readily investigated. Even so, there has been relatively little research on learner-generated questions, particularly related to the study of fiction, drama, and poetry. The focus of most research on questioning (e.g., Vander Meij, 1992; Wong, 1985) is not on understanding literature, but on what Ng and Bereiter (1991) call "extracting knowledge from text" (p. 268). Questions as Study Aids Adjunct yersus learner-&enerated questions. In developing a model for studying of nonfictional works, T. H. Anderson (1980) conducted an informal study of graduate students "engag in serious study" (p. 495). He used varied methods with eight students, such as observation; interviews; having learners place question marks in the margins of theU: texts; and having learners read aloud, predict content of paragraphs, summarize out loud, relate other thoughts as they read. He concluded that student generated questions and adjunct questions seem to be effective study aids for the same reasons: They "encourage students (1) to pause frequently while reading; and (2) to locate an appropriate answer to an 'understanding question' that probably is closely related to the study criteria" (p. 500). However, Frase and Schwartz (1975) found that unless encouraged to ask difficult questions, students tended to generate only verbatim recall questions. : : The different findings might be a function of student maturity and/or 39

PAGE 53

motivation .t\nderson's subjects were graduate students in a natural situation, while Frase and Schwartz's were high school students participating in an experiment. Questions yersus "merely Frase and Schwartz (1975) demonstrated the superiority of asking or answering questions over "merely studying"; students working in pairs who asked or answered each other's questions while studying biographical prose passages performed better on a recall test than students who merely studied (students were instructed to ask questions which would help them on a posttest). However, significant effects were limited to "content that was directly related to subjects' questions" (p. 628). The study controlled the number of questions asked (5 or 10) and to some extent the difficulty level of questions asked The role of prior Miyake and Norman (1979) investigated "the notion that a prerequisite for asking questions about a new topic matter is some appropriate level of knowledge" (p. 357). They found an interaction between prior knowledge and number of questions asked; novices asked more questions on easy material; "trained" subjects asked more questions on harder material. Although this fmding sounds logical, the study had limitations which relate to ecological validity. First, it was set in I a research laboratory, with no meaningful incentive for learning (subjects were undergraduates 'in introductory psychology courses and received course credit or pay merely for participating). As noted in chapter 1, incentives contribute to subjects' performances in research (Hicken, Sullivan, & Klein, 1992). Second, learners' questions were not answered, even though Miyake and Norman acknowledged the social context of questions: "To ask a question of someone implies more than a need for information" (p. 357; emphasis added). (This criticism does not imply that learners 40

PAGE 54

might not meaningfully ask themselves questions as they read Indeed, that is one of the functions this study attributes to journal writing. However, the reason for not answering wa8 clearly an artificial one in the leaining context: "to keep the amount of information available the same from subject to subject" (p. 359). Relationship to ability leyel. In a study of high school psychology students M. E. D. A. Andre and Anderson (1978-79) found that "self-generation of questions during study can lead :to improved performance on a test of comprehension" (p. 619); "appropriate" questions were modeled for the students. However, training mainly benefited low middle-ability students; Andre and Anderson concluded that high ability students; apparently already knew how to generate good comprehension questions; that 1is they were able to generate questions (on three passages related to psychological concepts, 450 words each) that facilitated good performance on a short (20-item) test; items required application of concepts and/or principles to new examples; 10 questions examined "which, who, when, where, what or how" (p. 612). While has generally found differential benefits of strategy training favoring lowapd middle-ability students, Chi, de Leeuw, Chiu, and LaVancher (1991) found prompting eighth-graders to generate self-explanations as a tool in learning physics concepts benefited average-and high-ability students equally. Similarly, in my preliminary studies of journal writing (Cole, 1991) students of all ability determined by reading comprehension and grades) thought they benefited from journal writing. A survey (Appendix C) of students in five of my literature revealed that students of all levels said they found journal writing 41

PAGE 55

helpful, with 84.6% of the A students and 60% of the B students rating it "Very Helpful," the on the survey (with 1 being "Not at All Helpful," 3 being "Somewhat Helpful," and 5 being "Very Helpful'); 77% of all respondents gave journal a rating of 4 or 5. The only respondent who was earning an F at the time of the survey Gust before the final exam) also rated journal writing as "Very Helpful." In their survey comments and in class comments, several of the A and B students identified self-generated questions as the most helpful aspect of writing One of the A students, who had written journals in two of my courses (Introduction to Literature and Masterpieces of World Literature Since the Enlightenment) wrote that journal writing was the most important thing she had learned in eight semesters as a part-time student. Another A student, who transferred to Amherst College more than a year after she had written journals in the same two literature wrote from Amherst that, when she found herself having difficulty with a philosopry of religion course, she decided to write journals and they solved her problems in learning the material. While these students may have been responding in ways they thought would "please the teacher," this seems unlikely, since their typical ,. behavior did not suggest such a motivation. More importantly, students' perceptions are one of many sources of questions worthy of research. yersus pre.posed teacher gyestions. Singer and Donlan (1982) demonstrated the effectiveness of learner-generated questions compared to prequestions posed by a teacher. High school students were taught to combine a problem-solving schema for comprehending short stories with the construction of schema-general questions for each element in six complex short stories. However, there were several 42

PAGE 56

limitations to this study. All students read the story at the same rate (listening to a recording as they read from a copy); thus they could not read at their own pace, look up unknown words, or re-read. Essentially all questions were generated at a given point in the story (the recording was stopped and students were asked to write three questions they would like answered as the story progressed); students were allowed to ask additional questions at the conclusion of the story, but few did Thus the study paralleled the adjunct-question format of single-position (near the beginning), grouped questions. authors did provide a list of story-general questions, which vary in difficulty, and examples of three story-specific questions used in training the students to ask questions; however, there is no way to evaluate the level of students' questions, since the authors did not provide any examples. The measure of learning was limited to a 10-item multiple-choice test on each story; nor did the authors itemize to see whether previously taught story elements resulted in improved mastery on subsequent tests. Finally, the autP.ors taught the students to generate questions based on a very simple story schema whose elements were leading character, goal, obstacles, outcome and theme. They added "theme" to an existing schema for vignettes and fables because they felt that it was important element for complex stories. However, after they found facilitative effects for learner-generated questions for the other part of their schema but not for theme, they concluded, "in retrospect we should not have included theme in our instruction because it is not a compelling aspect of story comprehension" (p. 182). Yet theme is the all-embracing element of complex stories (apart from escapist literature). Nor did their schema acknowledge elements that typically contribute to the theme directly or indirectly; for example, setting, symbols, point of view, and irony. (For 43

PAGE 57

further discussion of this issue in terms of the proposed study, see the Procedure Section.) Summazy of IJeamer-&eneratecl Questions The above studies, then, suggest that learner-generated questions can be an effective strategy. In spite of their limitations, the Miyake and Norman (1979) study and the and Donlan (1982) study suggest interesting avenues to explore, particularly in terms of (a) a more ecologically valid study, and (b) the interaction of college-age learners with complex short stories as they attempt to construct meaning. Implications of Schema Themy. The relevance of the above questions to both cognitive theory and instruction is supported by the generative model of learning (and the constructivist paradigm) and schema theory. For example, Wittrock (1974) observed that ":people tend to generate perceptions and meanings that are consistent with their prior learning" (p. 88). R. C. Anderson, Reynolds, Schallert, and Goetz (1977) similarly observed: From .. the perspective of schema theory, the principal determinant of the knowledge a person can acquire from reading is the knowledge s/he already possesses. The schemata by which people attempt to assimilate text will surely vary according to age, subculture, experience, education, interests, and belief systems. (p. 378) But they cautiop.ed that "dominant high-level schemata are often imposed on text even I when, according to a third party point of view, some violence is done to the 'data' contained in the text" (R. C. Anderson, Reynolds, Schallert, & Goetz, 1977, p. 371), and they concluded that "It may tum out that many problems in reading comprehension are traceable to deficits in knowledge rather than deficits in linguistic skill narrowly conceived ... (p. 378). 44

PAGE 58

Thus, it seems imperative that educators learn more about how students actually comprehend text--creative text as well as the more extensively researched expository text-so that they can facilitate students' construction of its meaning. SUilllll3Q' In this chapter I have discussed the literature on journal writing, adjunct questions, self-explanations, schema theory, and the generative model of learning. In many courses, journ3.1 writing supports other writing activities. But in literature courses, it is typically used to facilitate comprehension and to communicate problems and understandings to the teacher. I have shown that most of the issues related to journal writing as an instructiona]Jlearning tool in college literature courses (and other disciplines) have not been addressed adequately, if at all, by researchers. Writing journals to facilitate comprehension of reflective short stories (a very ill-defined domain) may involve unique cognitive and metacognitive strategies. The literature provides no research-based model of these activities. Nor does it provide a suitable taxonomy for classifying the content of journals on short stories. I presented literature on adjunct questions, self-explanations, schema theory, and the generative model of learning because as students read reflective short stories, they engage in a dynamic cognitive process which relates to all of these. They monitor their comprehension; generate questions and self-explanations (answers to their explicit or implicit questions); and activate, refine, or create schemata. In the next chapter, I describe the methods I used to study college students' use of journal writing in comprehending reflective short stories. 45

PAGE 59

CHAPI'ER3 MEIHOD Conce.ptual Framework 11 used predominantly qualitative research methods in the study, employing the conceptual framework reflected in Figure 3.1. Although the figure reflects a generally linear process, beginning with theory and moving clockwise, the driving force is the question focus. Thus, a change in the focus of the question(s) may force the researcher into an process, particularly in developing tools, indices and instruments; in observation and data collection; and in data analysis. (Ultimately I hope to explore the topic of writing, as defmed in this study, in a quantitative study; for example, assessing the effectiveness of journal writing as a learning strategy in literature courses, examining whether learners can be taught to improve the effectiveness of their journal writing while studying literature, and examining the transferability of the strategy to other courses.) Qualitative'tesearch has become accepted as a legitimate alternative to quantitative research, not an adjunct (Eisner & Peshkin, 1990). Donald Campbell and Lee Cronbach, who: were considered major spokesmen for quantitative research in the past, lJ am using subjective rather than an objective stance in chapters 3, 4 and 5 to emphasize the role of the researcher as a "smart instrument" (Guba & Lincoln, 1982, p. 240) in qualitative researcl), or as Guba and Lincoln prefer, "naturalistic research." They emphasize that, rather than ttying to overtome the '"inquirer-respondent interactivity ,' the natwalist exploits it" (p. 240). 46

PAGE 60

have both advocated the usefulness of qualitative research (Eisner & Peshkin, 1990; Patton, 1980). Figure 3.1. Conceptual Framework of the Study* Theory (tacit & formal) Limitations of the study; :.future research Conclusions Description Data analysis Question Focus Observation & data collection Models Concepts Hypotheses Operatonalism in a particular setting & sample for observation & measurement *Adapted from Marshall & Grossman, 1989, p. 23. Educatiomll technology is following the trend toward greater flexibility in research methods. Salo1J1on and Gardner (1986) advocated that "'computer researchers' should ... utilize holistic as well as standard experimental research paradigms, particularly during the early, phases of research" (p. 13). Even a cursory examination of studies published in the Journal of Experimental Psychology (e.g., Stadler, 1989; Willingham, 47

PAGE 61

Nissen, & Bullemer, 1989) reveals the incorporation of qualitative methods such as self-reports in:experimental research, in an almost matter-of-fact way. The Association of Educational Communications and Technology (the primary professional organization of education teachers) provides a special award for a qualitative research investigation in the field of educational communications and technology to encourage doctoral candidates to engage in qualitative research. Several journals are also encouraging submission of qualitative research articles (e.g., Educational Technology publishes juried articles on qualitative research). Patton (1980) provided guidance in determining if a qualitative study is justified; he lists 16 apprc;>priate situations, 5 of which relate in varying degrees to the research issue at hand: the programs "emphasize individualized outcomes," there is a need for "detailed, in-depth information ... about certain client cases," there is an interest in "collection of detailed, descriptive information about the program for the purpose of improving the program," "no valid, reliable, and believable standardized instrument is available or readily capable of being developed to measure the particular program outcomes for which data are needed," there is "a need to personalize the evaluation process by using research methods that require personal, face-to-face contact with the program-methods that may be perceived a.S 'humanistic' and personal because participants are not preordinately labeled and numbered, and methods that feel natural, informal, and understandable to (pp. 88-89; emphasis in original). 48

PAGE 62

I researchers must tailor the designs of their studies to their needs (see, for example, lark, 1989; Patton, 1980). Thus, I used a mixed methodology, quantifying data whenever appropriate while generally adopting the stance of a ,, phenomenologist-" concerned with understanding human behavior from the actor's own frame of reference" (Bogdan & Taylor, 1975, p. 2, as cited in Patton, 1980, p. 45; emphasis in Patton). I utilized student self-reports, surveys, interviews, content of journals, and students' oral comments during class discussions, during office hours, and in spontaneous conversations before and after class. I The need for ecologically valid methods in studying journal writing has a parallel in the adjunct question literature (E. D. Gagne, Broughton, Eggleston, Holmes, Hawkins & Sheldon, 1979). Lindner and Rickards (1985) state: It is readily apparent [from a review of the literature] that the limitations placed on the learner in adjunct question studies in no way resemble the normal reading and/or study behaviors of students. Subjects can neither re-read nor undedine nor take notes in the typical experiment, nor are they allowed to review text passages once they have a question. Generalizations to real-life settings from the results of adjunct question experiments are, consequently, rather limited. (p. 146) Although Lindner and Rickards were referring to adjunct questions, the studies of questions (see chapter 2) suggest that ecological validity is of equal concern to that area of investigation. While there has been much naturalistic research on the process of writing (e.g., Emig, 1971; Graves & Murray, 1980; Hayes & Flower, 1980), 'there has been little if any naturalistic research on the strategies students use in the of fiction, especially relating to learner-generated questions and As discussed in chapter 2, T. H. Anderson (1980) conducted an informal study of graduate students "engaged in serious study" of nonfictional works 49

PAGE 63

(p. 495); he used varied methods with eight students, such as observation; interviews; having learners place question marks in the margins of their texts; and having learners read aloud, predict content of paragraphs, summarize out loud, and relate other thoughts as they read. I have used some of his methods in the present study. For example, instead of placing questions marks in their texts, journal directions asked students to: record questions you have about an assigned story (any questions whose answers you believe will help you understand the story), and ask questions during class discussion. Students could also summarize the story in writing, and elaborate on the story in writing and during class discussion. For most of the students, surveys replaced Anderson's interview method; I also interviewed five students to obtain more information about their journal-writing strategies. 2 Because of the nature of the reading-writing assignments, I did not observe students during actual reading and writing. Back!Wlund and Setting Back&round of' the Researcher When I initiated the study, I had taught college English and literature full time for 25 years, the 23 of which were at Arapahoe Community College, a public community college in a large metropolitan area I have a BA and an MA in English, 2see below .and chapter 4 for an explanation of the criteria used to select students for interviews. 50

PAGE 64

and a BA in I have done additional graduate work in English and Education and'am a doctoral candidate in Instructional Technology at the University of Colorado, Denver Center. My special focuses are cognition and instruction and instructional design. Settincfor the Study I preliminary studies of journal writing in literature classes for three semesters at Community College (ACQ. I conducted the current study during the 15-week spring semester, 1992. Located in a white-collar suburb, the college enrolled 7,614 students spring 1992; full-time-equivalent enrollment was 3,744.3 The older students are generally more motivated and higher achievers than the recent high school graduates.4 Table 3.1 compares the college population to Group 1, the primary focus of this study. At the college, an EnglisMiterature instructor teaches five classes with a maximum of 23 students per semester. Technological support is traditional (chalk board, overhead film projector, slide projector, tape recorder, and limited access to 1: -VCR). Although the college hadextensive computer resources for its occupational vocational programs, there was essentially no computer support for the academic program until the middle of the semester during the current study. 3 Full-time equivalency is based on a 12-credit load. is based on my own observations and on conversations with faculty in many other disciplines. We base our conclusions on issues such as students' preparedness (e.g., Do they come to class prePared? Do they submit assignments on time? Do they learn what is expected or continually challenge requirements or make excuses?), their participation in class discussions, and indepth rather than sluface learning. 51

PAGE 65

T bl 3 1 C a e .. ompansono fACCP ul dG op1 atJ.on an roup 1 ACC Group 1 Number Enrolled' 7 614 15 Average Age 31.2 28 Average# CreditS Sming 1992 7.4 8.4 Employment Full time 44% 67% Part-time 18% 26%* Unemployed 19% 0% Unknown 19% 7% Marital Status Manied 42% 40% Single 56% 60% Unknown 2% 0% Sex Female 59% 80% Male 40% 20% Unknown .004% .. The mm1mum number of hours employed was 18. Within this setting, I have used journal assignments as an inexpensive strategy to respond to the needs of individual students; although journal writing is time-consuming for both students and teacher,s it does not require expensive technological support. For the previous seven years I required students in my literature classes (and most of my composition claSses) to write journals on most of their reading assignments. The primary intent.has been to engage the student with the work as much as possible within the constraints of a traditional college learning environment; that is, an instructional environment with high student-faculty ratios and little technological support. 5 Although no surveys addressed the amount of time spent, some students reported spending up to 10 hours on journal assignments. I spent as little as five minutes grading short, superficial journals, and up to an hour grading long, probing journals. I probably averaged 10 to 15 minutes per short-story journal. 52

PAGE 66

Subjects1Resppndents6 In choosirig the subjects, I used purposeful sampling of typical cases, as described by Patton (1980); Berg (1989) uses the term "purposive." The subjects of the study were students in two of my classes at the urban public community college where I teach. All stude:nts were enrolled in freshman-level "Introduction to Literature," a 3credit semester course in the study of fiction, drama and poetry, meeting 150 minutes a week. There were 157 students in the primary section (Group 1), which met from 5:30 to 6:45p.m., Mondays and Wednesdays. I selected this class because it was the largest one I had at the time of the study8 and was fully constituted on the first day of class.9 To assess typicality, I included a second section of Introduction to Literature with 10 students,lO which met from noon to 12:50 p.m., Mondays, Wednesdays, and 6Guba and Lincoln (1982) prefer the term "respondents" because they believe it more accurately describes the relationship which characterizes qualitative ("naturalistic" research). They believe the term "subjects" is more suited to experimental ("mtionalistic") inquiry. To reflect the ecological validity of my study, I generally use the term "students." ?sixteen students enrolled, but one had to drop after the first few weeks for fmancial reasons. Only 14 students completed the two journals which were the primary focus of the journal analysis. 8The small sizes of my classes was an unfortunate circumstance, since it affected the statistical significance of many of the correlations between factors of interest. Normally, my classes enroll at least 20 students, but this term started with 16, 10, and 8. Although I normally teach five classes, I taught only three at the time of the study; I was released from two classes to work on other projects. %fy second largest class, which I included as Group 2, was constituted from the "wait list" of another class. Unfortunately, the instructor did not immediately send all the students to the new section. Thus, of the students missed up to three days of the first week of class, including many of the class orientation activities. I lOI had plannect to include a third section, a sophomore-level course in world litemture. Although I did assign studentS to write journals, tested their reading levels and field dependence{mdependence, and gave them the same surveys I gave the other two classes, the class was too small to warrant inclusion in the study. 53

PAGE 67

' Fridays. Students in college-level classes at the college typically represent a broad range of ages (18-45 or older, sometimes retired senior citizens), abilities and backgrounds (e.g., veterans, housewives entering college for the first time or returning to college after 10-20 years, students who have taken developmental composition or reading, and outstanding students who ultimately earn full scholarships to distinguished colleges such as Amherst, St. John's College, Georgetown University, Ripon College, and Colorado College). At least 75% of the students in the Introduction to Literature course take it to satisfy part of the Humanities requirement; other students take the course for general-elective credit or for personal enrichment (occasionally senior citizens, most of whom already have bachelors.or masters degrees, take the courses for enrichment). Although students are advised to have college-level reading skills before enrolling in college-level courses, some students disregard the advice. During fall semester 1990, 15% of the students in my two sections of Introduction to Literature entered with reading skills below the 9th-grade level; 25-30% read at the lOthor 11th-grade level. However, that distribution was an extreme one. During the current study, only 14% (2) of the 14 students in Group 1 who took the reading test read below the 12th-grade level (vocabulary level, comprehension level, or total reading level), and 35% (5) read at or above the 16.9th-grade level. None of the students in Group 2 read below the 12th-' grade level on any measure. Materials and Resources Materials and resources included the following: 54

PAGE 68

Faculty Index of Story Difficulty (Appendix A); Nelson-Denny Reading Test, both the Vocabulary and Comprehension/Rate subtests (FormE, 1981 edition). Although the college uses the ASSET tests for academic advising, I used the Nelson-Denny test for two reasons. First, it is used by the Reading Department at the college. Second, according to Ysseldyke (1985), it is one ohhe "few measures of reading achievement designed for use with ... college-age students" and "is carefully constructed" (p. 1037). Grade equivalentsfor the test range as follows: vocabulary, 3.6-16.9; comprehension, 3.6-16.9; total reading, 3.7-16.9. Ysseldyke noted that the Vocabulary test has a test-retest reliability range of .89-.95, and the Comprehension test-retest reliability range is .75:-.82. Since the Reading Rate is not of concern in this study, its lower reliability range (.62-.82) does not affect the study. Ysseldyke (1985) observed two problems with the Nelson-Denny Reading Test which potentially impact this study. First, the college and university sample on which the test was standardized were not truly representative; colleges and two year schools with fewer than 5,000 students were over-represented, and Hispanics and Blacks were underrepresented. Since this is an exploratory, qualitative study, the representativeness of the norming sample is not of great concern (it is important only that the test yield a wide range of abilities). The second concern relates to validity and has greater significance for the study; Y sseldyke ( 1985) reported that there is "little data in support of [claims for] intended uses" of the tests, "including identification of both superior students and those who may need special help in reading" (p. 1037). However, the test should still facilitate identifying the best and 55 !.-.--------------------------

PAGE 69

poorest on a relative scale; I did not use it for academic placement purposes. Given the lack of a demonstrably more valid test, I chose this one. Hidden Figures Test (Ekstrom, French, & Harman, 1976), Part 1. I chose this test to assess the cognitive style of field dependence/independence for two reasons. First, has demonstrated an association between the Hidden Figures Test and field (Nowrozi, 1980; Jonassen & Grabowski, in press; also :see Glossary). Second, I wanted to minimize the amount of class time devoted to testing in order to maximize the amount of time for instruction. Since this test takes less than 15 minutes of class time, it was possible to administer it to Group 1 during the same class session as the Nelson-Denny Reading Test Furthermore, I didn't want testing to potentially create an environment that might suggest that the students' learning was not my primary focus. At the beginning of the course, I told students that I would be collecting information that would help me help them a,s well as future students. Although the Hidden Figures Test has been used as a measure of field dependence/independence in many published research studies, I could not locate any data on,reliability or validity, even in the Boros mental measurements yearbooks (7th through lOth editions). However, Jonassen and Grabowski (in press) report that it "has high reliability and good construct validity." an introduction to literature textbook containing all or most of the stories to be read but generally without study questions (Barnet, Berman, & Burto's [1989] An Introduction to Literature) and, "The Appointment in Samarra," reproduced from another text. Stories were selected (a) which, with two exceptions, none of the 56

PAGE 70

students had read or studied before the semester began, II and (b) for which, with one exception, 12 the text did not include adjunct questions and/or commentary, since one would expect that reading stories with which one is familiar or which have adjunct questions and/or commentary would reduce the number of questions and comments as well as affect the type of questions and comments. handouts to defme journal writing, to describe the process used by three former students "Student Accounts of Journal Writing" (Appendix H), to provide examples (Appendix 1), and to provide prompts for thinking about stories ("Catechism for Stories," Appendix G); The "Student Accounts of Journal Writing" is an optional job aid to scaffold students in developing their schemata for journals, including a flexible procedure to meet their needs. The Accounts were three former students' descriptions of how they wrote their journals. 11Dinah hadstudied "Hills Like White Elephants" and 'The Cask of Amontillado." Frances had studied 'The Story of an Hour" in my composition colll8e the prior semester. She read "The Cask of Amontillado" during the current study before I decided to assign it Lisa had read 'The Cask of Amontillado." :sheila had studied 'The Cask of Amontillado," but she did not submit a journal on it Trish had also studied 'The Cask of Amontillado." Because three students had already studied it, I initially planned not to assign "The Cask of Amontillado." However, for several reasons, I decided to assign it: (a) I felt the remaining students deserved the experience of studying it; (b) it is such a difficult story students often seem to benefit from studying it a secoQd time; and (c) many teachers often do not provide adequate insight into the story. I base my conclusions on comments prior students made when we studied the story and on my observations (as fanner chair of the English department) of English classes at the college. 12"The Story. of an Hour" had adjunct questions. I assigned this story because it provides excellent examples of dramatic and situational irony. But to eliminate the confounding variable of adjunct questions in other journals, I purposely avoided assigning other stories with accompanying questions. Although the journals on this story were not included in the primary analysis, they did suggest some relationships between students' use of adjunct questions and field dependence/ independence. See discussion in chapter 5. 57

PAGE 71

The for Stories" is an optional job aid to scaffold students in (a) writing journals and (b) developing their schemata for analyzing stories. two HyperCard tutorial programs for students to use as resources if they wished: Writing Jo'urnal Questions and Writing Journal Comments;B a HyperCard resource on fiction (definitions and examples of 54 terms; practice applying many of the terms; brief discussion of journal writing; expanded discussion of the "Catechism for Stories"); objective short-answer quizzes (Appendix E), to assess factual knowledge of the stories students have written journals but before class discussion (these ,. quizzes, of:course, also served as one means to motivate students to read the stories carefully). a multiple-choice mid-term and final examination, to assess mastery of the course objectives (the examinations also served as means to motivate students to focus on the objectives of the course as they read the stories and attended class); recall items were included to determine if students had mastered the terminology; application I items relating to stories discussed assessed near transfer; application items relating to an unfamiliar story assessed far transfer, which is widely accepted as a primary aim of problem-solving (e.g., Hayes, 1985; Newell, 1980) and of education in general (e.g., Prawat, 1989). Buntil the fQurth week of the term (when the college opened a laboratory with the college's ftrSt Macintosh computrs), these were accessible only by making an appointment with me to use them on a faculty computer, Only one student took advantage of the opportunity to use them. 58

PAGE 72

and semistandardized interview forms to gather demographic data and data about students' uses of journals and their perceptions of the value of writing journals specifically in the study of fiction (Appendices A, B, C, D, and F). Data Collection In the tradJ.tion of qualitative methodology, I triangulated (e.g., Berg, 1989), using a variety of data sources, all listed below. I piloted all of the paper-and-pencil questiolUlaires 'in the fall of 1989, then revised and piloted them on students in three literature courses in the spring of 1990. I further revised surveys of journal writing (Appendices and C) to include questions about handouts on journal writing, which I developed prinlarily in response to students' suggestions in the spring of 1990. 1. Student jouptals. Students wrote their questions about and comments on each story in a journal before class discussion. I chose an unobtrusive journal format rather than think-aloud protocols to study the types of questions and comments students have, because writing questions and conunents in journals parallels the normal study behavior of students more than do think-aloud protocols. Students had the opportunity to read the stories as many times as they felt necessary, to look back at passages, to look at their notes, etc. Intuitively, journals are ecologically valid teachinglleahting strategies; the very nature of their instructional use also reflects one function of direct quotations in qualitative methods as described by Patton (1980): quotations are a basic source ofraw data] ... revealing respondents' level of emotion, the way they have organized their experiences, and their basic p erceptions [e.g., schemata relevant to the story]. The task for the methodologist is to provide a framework within which people can 59

PAGE 73

respond in a way that represents accurately and thoroughly their points of view about the world, or that part of the world about which they are talking.... (p. 28) Although I cannot be sure that the students' journals "accurately and thoroughly represent[ ed] their points of view," use of journals would seem to allow and encourage ihis more than think-aloud protocols. It is important that the journals be confidential (between the student and me). I do not share them with other students, as is often the case with instructional methods which incorporate journal-writing (e.g., Duke, 1982; Nugent & Nugent, 1985). From personal experience, I have found that are more willing to share their comments and questions with the instructor than with other students; 14 and although no student may object in class when the instructor asks, "Does anyone not want to share his journal with other students [in groups]," at the end of one semester in which I had students share their journals in small groups, one student did indicate his objection on a survey I administered. Time constraints made it equally unrealistic to expect students to write journals in the presence of the researcher, who would be able to observe the actual activities students engage in. I chose to trade off (a) the ability to directly observe students writing journals on much shorter and much easier stories under artificial conditions for (b) ecologically more valid conditions for students' studying "reflective" :short stories of the type and length nonnally assigned in the course. 2. Surveys of students a. First day of class, a checklist of prior reading. to determine stories students had read (I used responses to select stories most students had not yet read) and 14see Laura's comment on the most helpful aspect of journal writing, in chapter 5. 60

PAGE 74

a Q.Uestionnaire to collect demowmhic data. including major, background in literatUre and number of college-level credits (see Appendix D). Orally and in print ori various surveys, I told the students that I was collecting data so that I could :learn more about how students learn to understand stories and so that I would have information to help me help them and future students. b. Mter first journal and last journal, survey of journal writing to collect data on students' procedures in writing journals and on their reactions to writing journals (see Appendices B and C). 3. Interviews to probe students' journal writing behaviors. During the term, I interviewed five students who differed across a range of dimensions (see discussion selection criteria in chapter 4). I used a semistandardized interview format (Berg. 1989; see Appendix F). 4. Paper and pencil quizzes to assess students' factual knowledge of each story; I administered a five-item objective quiz at the beginning of class after students submitted journals but before class discussion of each story (see Appendix E). The was to motivate students to read the story carefully as a basis for class discussion. 5. An objective multiple-choice mid-term and final examination to assess students' mastery of the course objectives (e.g., verbal information on concepts and principles; application of concepts such as protagonist, irony, and theme; and problem solving). The midterm (the end of the eighth week) addressed only the short-story objectives; the final (on the last day of class) was comprehensive. 61

PAGE 75

6. Readin& Test (Vocabulary and Comprehension/Rate; FormE, 1981 edition) dUring the third week of the study, to assess the typicality of the (students) and to help determine cognitive ability. 7. Hidden Figures Test (Ekstrom, French, & Harman, 1976), Part 1, during the third week of the study, to assess cognitive style. 8. Student records, to identify exact age, and semester and grade-point averages. Procedure I served as researcher and teacher, teaching the course in the same basic manner that I have used for about 20 years in all of my courses-predominantly teacher-led discussion (I have also had students write journals in most of my courses during the past seven years). Students studied fiction during the first seven weeks of the course, reading a story'and writing a journal about once a week. They wrote two additional journals on plays. Table 3.2 provides an overview of the research procedure. This overview elaborates parts of the Conceptual Framework in Figure 3.1: developing tools and indices and instruments for observation and measurement; observation and data collection; and data analysis. Following is a detailed discussion of the procedure used with Group 1, which met twice a weekfor75 minutes: 1. Established difficulty leyel of stories. Weeks 1 -6: I planned to sequence stories along a continuum of generally increasing difficulty. In determining the difficulty level of stories, I rejected readability indices although they have been used in studies of fiction (e.g., Shannon, Kameenui, & Baumann, 1988). Readability indices were derived from studies of expository prose, whose 62 ...._ _______ L,--------------------------

PAGE 76

Table 3.2. Ov.erview of the Research Procedure Established 'the difficulty level of stories potentially to be assigned Oriented stUdents to the course and to the study, and collect demographic data. Selected and sequenced stories to be assigned Administered tests of reading ability and field dependence{mdependence. Presented shema for classically-structured stories. Established ,baseline of students' journal writing activities. "Survey of Journal Writing, Mter First Journal." Conducted class and collected journals for analysis: Assigned story and journal; gave quiz and discussed answers; discussed story. Interviewed: and tutored students. Administered mid-term examination. Administered "End-of-Journal-Writing Survey." Analyzed and interpreted data. Developed taxonomy of questions and comments in students' journals. Developed model of journal writing. purpose is to transmit information as clearly as possible, typically in a top-down structure. The indices are based on the average number of syllables and average sentence length. However, research showing the importance of general world knowledgein comprehending expository prose (e.g., Brown, Campione, & Day, 1981) calls into question the usefulness of any of the currently existing readability scales even for expository prose.15 15Brown, cainpione, and Day (1981) note that the teacher may "select texts that deal with familiar material" pr "actively attempt to provide the requisite background knowledge for a particular text, but she/he cal)not always do this." For several reasons, this problem would seem to be exacerbated in the I study of literature. First, college-level literature makes extensive use of allusions. Second, many writers (e.g., Donald Hall and Alexander Solzhenitsyn) have argued that literature is a vehicle for expanding knowledge of people, history, and other cultures. Third, teaching declarative knowledge out of context (i.e., as background knowledge for a particular piece of literature) is, at best, artificial. Moreover, such an approach does not teach students to learn to identify and debug their 63

PAGE 77

In to the purpose and techniques of expository prose, fiction requires the reader to construct meaning in a way that expository prose does not The basic structure is never top-down (although the structure of individual paragraphs may be), and the author often delays or even withholds information for artistic or affective pUrposes. Thus the difficulty level of stories for a given readership (e.g., college-level readers) is more a function of the demands placed on the reader by factors such as the narrative structure; the point of view; and the use of allusion, I symbolism, and irony. Faculty ratin&s. I had planned instead to use faculty ratings of story difficulty (Appendix A) to guide my selection and sequencing of stories. During the first week of the term, I administered the survey to the other six full-time members of the college-level English program. The survey revealed no consensus, 16 and suggests the need for a future study to try to develop a tool for identifying difficulty. The range of responses provides further justification of the usefulness of joumals:in communicating students' difficulties. If teachers with so much 7 do not agree on the difficulty level, then at least some of them must have inaccurate schemata of their students' needs. comprehension problems. Learning how to learn and to debug one's own knowledge has been called the most important goal of instruction (e.g., Nonnan, Genbler, & Stevens, 1976). 16For 29% or' the 14 stories rated by more than one person, the ratings spanned the range. Including my ratings for those stories increased the percentage to 36. The faculty did not agree unanimously on the rating of any story, although 83% (5 out of 6) did rate Faulkner's '"The Bear" as "Difficult" Note: I asked respondents to focus on the difficulty of understanding a story's meaning, not on its vocabulary leveL 17Four of the:teachers had taught full-time more than 20 years; the fifth had taught full time more than 15 years; the sixth had taught full time less than 2 years. 64

PAGE 78

Since:there was no consensus, I decided to rely on my own perception of the difficulty llevels based on the journals and class discussions from prior semesters)8 I wanted to minimize assigning stories students were already familiar with, because I wanted students' journals to be as free as possible of the influence of prior exposure and study. (See Step 3 for the actual selection and sequencing.) Results based on assumptions of story difficulty are necessarily tentative. However, with many years experience reading students' journals on four of the stories, I am confident of the rankings I assigned. Furthermore, my estimations of story difficulty correlated 1.0 with the ratings assigned by an adjunct faculty member who requires her students to write more structured journals than I assign and who has many years experience teaching college English and literature. She based her ratings on criteria similar to mine. I asked her to rate the stories, without knowledge 10f my ratings, during the data analysis phase. 2. Oriented stUdents to the course and to journal Week 1, class 1: I oriented students to the course and the study. Orally during the first day of class, I told students that I was collecting data so that I could learn more about how students learn to understand stories and so that I would have information to help me help both them and future students. The syllabus also included directions for writing journals (see Step 6 below). I administered the survey of students' backgrounds in literature and the checklist of prior reading (see Appendix D). 181 had taught most of these stories to at least 30 classes in the previous six to eight years. 65

PAGE 79

I gave 'students in-class practice in writing a journal on a haiku poem19 (asking them to write any questions or make any comments that they would like to make). To introduce journal writing, I chose a haiku poem rather than a short story primarily because of time constraints. The brevity of a haiku poem allows time for thoughtful reading, writing, and discussion within one class session. That particular haiku also stimulates discussion of the problems that arise when a reader has an inadequate schema for concepts in a work of literature. I co.llected their journals, and, without reading their names, read what each student written. I noted that many students had written similar things, but that some had written quite different questions and comments. I emphasized that each journal was acceptable. Then using students' questions and comments, I helped students construct more complete understandings of the poem. 3. Selected and segyenced stories, Weeks 2-6: I drew on my prior experience in teaching stories (see Step 1) and the student checklist of prior reading (Appendix A) as a guide in selecting and sequencing stories for study. My primary goal was instructional; each story satisfied one or more major Instructional goals.20 Within that constraint I selected and sequenced stories that would also satisfy my research goals. Following is an explanation for 19"Dragon Flies," an 18th-century Japanese haiku. 2DExcept for "Cathedral," each of the stories has stood the test of time and is generally regarded as a "classic" work in the canon of short stories. Although I do not consider "Cathedral" as outstanding as the other works, it is highly regarded and is included in many anthologies for college-level fiction and introduction to literature courses. 66

PAGE 80

the selection and sequencing of specific stories; with one exception (story b), none of the stories had adjunct questions (a copy of each story is included in Appendix J): a. 'The Appointment in Samarra" is a relatively difficult one-paragraph tale. I wanted the first story to be challenging enough that all students would be likely to have some questions or points of confusion, but short enough that students would not become overwhelmed. Thus, I would have the opportunity to observe how they responded to a story they could not readily comprehend. For example, would they communicate their confusion through their journals? What types of questions, if any, would they ask?21 Would they generate self-explanations? This story involves schemata that students have difficulty comprehending (e.g., the concept of unavoidable fate, the association of death with white rather than black); furthermore, the outcome relies on irony, and the author does not follow the convention of enclosing dialogue in quotation marks. 22 b. ''The Story of an Hour." I selected this story predominantly for pedagogical, rather than research, reasons. It is a short, relatively easy story, yet its use of irony and its tum-of-the-century schema for women always pose problems for mimy students. Although this story had adjunct questions, I decided to assign, it because it is the best story in the students' text to illustrate dramatic 21They might ask questions directly or indirectly, and they might ask open or closed questions (see defmitions and examples in the Glossary). 221nstruction3Ily, the story fostered discussion of several story elements; for example, setting, protagonist, antagonist, conflict 67

PAGE 81

and situational irony. Many students find these concepts difficult to understand as well as apply, yet they are important for students to master early'in the course since so many stories exploit irony. The adjunct questions ultimately provided insight into students' journal writing, since some students apparently used them as prompts (see discussion in chapter 5). c. "Hills Like White Elephants." I chose this story to help students develop their concept of point of view. But it is, in my view, the most difficult of the five stories, not only because it uses the objective point of view but also because it withholds the specific topic of the conversation (an abortion). Generally students are unable to infer the correct topic. A short journal from a former student illustrates the reaction of many students: I don't care how renown E. Hemingway is, he has no right to write :a story which no reader can grasp. It is one thing to be intrigued by some nebulous thing in a story, but it is another to have the whole story !be about that thing and to never find out what it is. Marriage, an illness, duo suicide? There is no way of telling. And the story is pointless without knowing! lck! d. "Cathedral." I chose this story to build students' confidence in their ability to understand short stories while I also introduced symbolism and gave students the opportunity to read a story with a happy outcome.23 Placing it after the most qifficult story allowed me to examine interactions between journal items and difficulty level. e. "The Cask of Amontillado." I selected this story for two reasons. First, it allowed me to address major instructional goals which students had not yet 23students often complain that all of the stories deal with death or other sad outcomes. Indeed, most anthologies have few happy or humorous stories. 68

PAGE 82

mastered. Second, it is a masterpiece which every student should have the opportunity to study, not merely read I placed it fifth24 because it is difficult for many reasons: its vocabulary; its unfamiliar schemata (e.g., the II: "brotherhood of the masons"; coats of arms; family mottoes; and Mardi Gras, which is identified only as "carnival season"); its symbolism; its subject (murder for revenge); and the other problems it poses (e.g., Does the murderer achieve revenge as he defines it, and why does he want revenge?). Initia)ly I did not intend to assign this story because three students had studied I it in another course, but I decided that all the students would profit from studying it (see footnote 11). 4. Tested ability and field dependence/inde.pendence, Week 3, class 1 A counselor/academic advisor at the college administered the Nelson-Denny Reading Test and the Hidden Figures Test of field dependence/independence while I was attending a professional meeting. I had told students that I planned to use the test results help both them and future students. 5. Introduced classic short story schema and oriented students to outcomes, Week 1, class 2; and week 2, class 1: I read Petronius' tale "The Widow of Ephesus" in class, stopping periodically to have studeqts predict whether the story was finished and why they thought so. 241 have often taught this story first, thinking that class discussion would help students realize the types of problms one can encounter in trying to understand a story. But I now believe that placing such a difficult stqry at the beginning of the course is cotinter-productive. It de-motivates some students (they believe that all the stories will be totally beyond their grasp), and it fosters overgeneralizatiori in some others (they believe that all stories must contain esoteric elements and seem to try to find ways to justify totally unwarranted interpretations of various elements). Sequencing should try to optimize the level of cognitive dissonance (see discussion in chapter 5) and thus be responsive to students' zones of proximal development (see discussion in chapter 5). 69 __ ....., _____

PAGE 83

Then I explained the classic short story schema, applying it to the story. I also introducecl:the concept of plausibility, applying it to the story. The classic short story schema includes the following elements: Exposition Setting Time Place Protagonist Antagonist Conflict Crisis Climax Consequences Denouement Symbols Irony Verbal Dramatic Situational Point of View Atmosphere Theme For several reasons, I refmed and expanded the various story schemata used by many irivestigators (e.g., Adams & Collins, 1977; Rumelhart, 1977; Singer & Donlan, 1982; and Thorndyke, 1977). First, Singer and Donlan (1982) found the schemata of the other researchers and theorists "inadequate for comprehending complex short stories" and thus developed a new schema. Singer and Donlan emphasized that the other schemata focus on plot and do not acknowledge the importance of theme in complex short stories (or, for that matter, any adult, reflective fiction). However, as noted in chapter 2, after Singer and Donlan found facilitative effects for learner-generated questions for the other parts of their schema but not for theme, they concluded, "in retrospect we should not have included theme in our instruction because it is not a compelling aspect of story comprehension" (p. 182); they cited Adams and Collins (1977) for support Yet every college literature text with instructional apparatus (and/or teacher's manual) 70

PAGE 84

presents theme as the all-embracing element of college-level stories (apart from escapist literature). Second, none of the schemata cited above acknowledge elements that typically contribute to the theme directly or indirectly (e.g., setting, symbols, point of view, and irony). The inability or failure to recognize irony can result in a total misunderstanding of a story. I wanted to provide a schema which would encourage students to think about all relevant aspects of a story and the way they interrelate. 6. Assi!Wed stmy 1 and journal1: conducted "Survey of Journal After First Journal," Week 2, class 2: Story 1 and journal 1: For homework students read a one-paragraph story and wrote their first journal. I scaffolded students in this assignment by giving them a copy of tqe "Catechism for Stories" (Appendix G), three students' accounts of journal writing from spring 1989 (Appendix H), and a copy of a journal from spring 1989 (Appendix 1). 25 The included the following instructions for journal writing: Wiite a journal for each story, at least a half-page long-but there is no maximum length. Write your journal in ink, using complete sentences otherWise I will not be able to understand what you are asking or saying. any questions you have--except the definition of a word, unless you are unsure of the relevant definition-or any comments you would like to make.about the story. The journal is not busy work, but is intended to help 25ouring die fall, 1990, I gave students copies of these documents after they had written their frrst journal-I brut wanted them to try to discover their own approaches to journal writing before giving them accounts and models from other students. Although several students said they found the documents helpful, they, nevertheless, said they wished they had had them before they wrote their first journal. Since then, I give these to students before they write their frrst journal. Note: The HYJ>ert:aro tutorials Writing Journal Questions and Writing Journal Comments were not readily availaole when the semester began. A major question is whether examples facilitate or impede students' own journal writing. The tutorials now provide an alternative learning strategy whose benefits must be investigated. 71 _______ ...,;,,, ____________________________

PAGE 85

me kitow what you need help with in understanding the story. Write any questions you believe will help you understand the story now as well as later, for the examination-or any other questions you are just curious about Journals are worth a maximum of 10 points each. H a journal meets the minimum requirements, you will receive 7.5 points--a grade of C; you will receive additional credit, depending on how much your questions and comments indicate that you are paying attention to all the elements as you try to understand the story. There are no wrong questions or comments (but there,. are questions and comments which ignore the details of a story). JoUll)als are a tool for you to learn about what you don't know, not to show me that you already know everything. Even if you don't understand a story at all but ask questions about what you don't understand and indicate that you are paying attention to details, you can receive a 10. I will collect your journals at the beginning of the class on which they are due. :H you miss that class, you may submit your journal at the beginning of the next class you attend26 (Note: Journals counted approximately 20% of the course grade.27) "Survey Of Journal Writing. After First Journal": To preclude any effect of feedback and grade on students' evaluations of journal helpfulness, I administered the first journal survey right after I had collected the first set of journals.: Although student anonymity on surveys is generally considered desirable in experimental research, qualitative research does not impose such constraints. Furthermore, in this ecologically valid setting (the students were enrolled for credit and were entitled to as much help as I could give them), I needed to know students' identities for two reasons. First, 'I needed to know how individuals responded so that I could do follow-up intervi.ews and adapt in-class instruction or offer tutorial assistance as 26Later subm,issions defeat the pwpose of the journals-to have students think for themselves rather than merely record class comments. 27There were journals on five stories and two plays. 72 ---------------------------

PAGE 86

necessary. Second, I needed to be able to identify respondents so I could analyze the data from the various sources. 7. Assigned stories 2-5 and their journals, Weeks 4-7:28 Students read a short story for homework and wrote a journal, about one a week (see in Step 6). I photocopied each journal after I had graded it Students took a five-question objective quiz on the story (to ensure careful reading of the story). Quizzes counted five points each, totaling approximately 10% of the course grade; I dropped the lowest quiz grade. 29 I We discqssed each story (usually a week each) and applied the story schema. Discussion addressed questions and comments from the journals of current and former students and students' in-class questions and comments. Class participation counted 10% of the grade. I responded in writing to students' journal questions and comments and returned journals at the next session. If the same question or misconception appeared in several journals, I included it in the class discussion, without identifying the student(s) who wrote it. 8. Administered midterm examination, Week 8, class 1: I an objective, multiple-choice midterm examination on short stories (40 questions, 100-points). The midterm and final examination each counted about 27% ofthe.grade. 28There was no class session 2 of week 3; I was attending a professional meeting. I dismissed all students in lieu o,out-of-class interviews later in the tenn. A blizzard forced the closure of the college for session 1 of week 7. 29There were quizzes on six stories and one play. 73

PAGE 87

9. SurvS
PAGE 88

f,' permission, I also typed a copy of her fourth journal (see Appendix K) and ,, it to students as an additional example of journal writing. 32 Conduc$J interviews. I inteiViewed five students to probe their journal-writing strategies. Students' schedules and mine did not pennit me to interview more students dUring the weeks they were writing journals. (See case studies in ,. chapter4.) Tutored students who asked for help. I tutored the one student who asked for help (Lisa). 33 The purpose of tutoring was to help her develop more effective ', strategieS for reading stories and writing journals. I modeled the process of generating questions and self-explanations, and scaffolded and coached her, in keeping Y.,ith the model of cognitive apprenticeship described by Collins, Brown, and Newman (1989).34 r 11. Analyzed jpld interpreted findin&s, After all the data were collected: I deyeloped a taxonomy of journal qpestions and comments. My primary goal was to try to increase understanding about the cognitive processing of students in ., ,, 32some stu4ents (particularly in Group 2) still seemed to be having difficulty writing productive journals. Others '(e.g., Sheila) had been productive on more difficult stories but on the easiest story (4) wrote mainly or judgmental statements. By contrast, Trish understood the story well but took the opportunity to explore higher-level issues, such as the role of the setting. 33Tutoring was in my office before class, six times during a two-month period. Sessions r generally lasted at least one hour. ( 34m the past_ I have only provided advice on journal writing to the class as a whole (no one ever asked for additio.W help). As I conducted this study, I wanted to take a more proactive position. In my comments to ,the class as a whole and in my written feedback on journals of students who seemed to be having diffi,culty, I encouraged students to come in for help if they wanted to, so that I could work with them iQdividually. I also encouraged students to view the tutorials on writing journal questions and comments. As noted elsewhere, Lisa was the only student to ask for help or to use the tutorials, on the End-of-Journal-Writing Survey (Appendix C) several students said they did not understand the purpose or grading criteria and suggested that I provide more help in the future. 75

PAGE 89

their journal writing. I particularly hoped to be able to identify levels of questions and comments and to determine any relationships these had to individual learner characteristics and other variables (e.g., students' ratings of journal helpfulness). But I also wanted to develop a taxonomy that would have practical value for a literature teacher. Although I applied all but the fmal taxonomy to most of the journals, for several reasons I focused my analysis on journals 3 and 4. First, I wanted to examine writing after students had had an opportunity to develop their skills and become comfortable with the process. Injournal1, and to some extent journal 2, students were learning how to write journals. 35 Second, journals 3 and 4 were on the most difficult and the easiest stories, respectively. They allowed me to examine potential effects of difficulty level. Third, I wanted to maximize the sample size; thus I chose journal3 rather than journal 5 (also a difficult story) because more students submittedjoumal3.36 Like Chan, Burtis, Scardamalia, and Bereiter (1992), I "began by trying to understand the data [in my case, the content of journals] at an intuitive level and then ... moved toward a more objective scale that preserved as much of the intuitive understanding as possible" (p. 101). Moving toward a more objective scale progressed through three major incarnations and several minor ones. For 35 Also, in journal 2, some students were apparently influenced by adjunct text questions. See discussion in chapter 5. 36sample size for comparing journals 3 and 4 was 14; sample size for comparing journals 4 and 5 would have dropped to 12. Sample size would have dropped to 10 or 11 for some of the correlations with individual learner characteristics (e.g., reading ability and field/ dependence[mdependence). I wanted to maximize possible statistical significance of correlations. 76

PAGE 90

after several attempts to operationalize shallow versus deep processing on two or three levels, I determined that the questions and comments in students' journals were too remote from the actual cognitive processes to reliably determine the level:37 (See Appendix K for samples of students' journals.) The unit of analysis was generally a sentence; however, if a sentence seemed to have more than one purpose, I classified each one separately.38 Recognizing the importance of the inductive approach of qualitative methods, I wanted to guard against imposing an existing taxonomy on the data. 39 (See Appendix L for a list of the categories in the taxonomy and Appendix M for examples.) I correlated with indiviciuallearoer differences. I calculated Pearson product-moment correlation coefficients to detennine if there were any relationships between categories in the taxonomy and individual learner characteristics such as age, ability (vocabulary, reading comprehension, and total reading levels as measured by the Nelson-Denny Reading Test), college gradepoint average, class performance (quiz scores, journal scores, and examination scores). The sub-groups were too small to analyze relationships between other variables and sex or background in literature. I also calculated Pearson productmoment correlation coefficients to determine if there were any relationships 37Reliability problems were pronounced when I reapplied the taxonomy to particular journals; I often made contradictory classifications or found myself puzzled over which level applied. See the related discussion of "cognitive artifacts" in chapter 5. 381bis occurred so infrequently I did not statistically analyze the occurrences. 39Jnitially, I had also hoped to apply several existing taxonomies (e.g., Bloom et al., 1956; Gagne, Briggs, &:Wager, 1988) as points of comparison, but I found them impracticable for the complexity and diversity of students' questions and comments. 77

PAGE 91

between journal performance (e.g., grades and content) and other factors (e.g., student ratings of journal helpfulness and their ratings of job aids). To identify potential interaction of ability level, story-difficulty level and question frequency (see discussion of Miyake and Norman, 1979, in chapter 2), as well as frequency of self-explanations, I split Group 1 into good40 versus poorer reading levels and performed several analyses, including Pearson product moment correlations and one-way ANOV As. I developed a model of journal writing. I had hoped to develop a model such as a flowchart, which depicted a clear sequence with decision points. However, when I examined data from the student surveys, interviews, and journals, I discovered that no single model could describe journal writing either within or .across students.41 For example, the difficulty level of a story affected the process in a variety of ways (type of activities, order of activities, and number and of recursions) for almost every student 42 At least one student (Foster) varied her procedure so she wouldn't get bored. It is possible that students would develop a more stable procedure over time, as they incorporated journal writing into their repertoire of learning strategies.43 40Qnly 12 of the 15 students wrote all five journals. The top six students all scored at the 16.9 grade level on reading comprehension, vocabulary and total reading ability. The bottom six students were distributed over as many as six grade levels. 41 I tried developing several traditional models; for example, a very complex flow chart, a chart of the type used in an extended task-analysis procedure (Jonassen, Hannum, & Tessmer, 1989), and a complex circular model of the type popularized by Romiszowski (1981). 42Eric was the only one who seemed to follow the same basic procedure for each work. 43 As noted above, Duffy and Roehler (1989) found that students may require several months to incorporate new laming strategies into their repertoires. 78

PAGE 92

However, it seems more likely that, at least for college-level students, journal writing in literature courses is an ill-defined domain (see Spiro et al., 1988), responsive to the demands a particular piece of literature makes on the student. (See chapter 5 for a description and discussion of the generic model I developed.) 12. Ensured. the trustworthiness of the study, Throughout the study: In this section, I am applying the terminology Guba and Lincoln (1982) advocated for naturalistic (qualitative) research. They emphasized the need for terminology separate from that used to describe rationalistic (scientific) research because the two types of research make different assumptions about the nature of reality, the relationship between the inquirer and the object of study/respondent, the nature of "truth" statements, the nature of causal relationships, and the role of values within disciplineci inquiry. Trustworthiness addresses four issues (Guba, 1981): truth value (intetnal validity in rationalistic research versus credibility in naturalistic research), (external validity versus transferability), consistency (reliability versus dependability), and neutrality (objectivity versus confinnability ). I implemented the following strategies, which Guba and Lincoln (1982) proposed for naturalistic research, to try to ensure the trustworthiness of my study. To establish and test credibility (internal validity), I had a engagement" with the students; the short-story journal-writing segment spanned 7 weeks; total data collection spanned 15 weeks, beginning with the background survey and concluding with the last-day survey of improvement in confidence in ability to write journals; 79

PAGE 93

engaged 1n "persistent observation"; I met with the students as a group 150 minutes a week for 15 weeks, 7 of which directly involved the study of short stories al}d 9 of which involved journal writing; and interviewed five students outside qf class; engaged in "peer debriefmg"; I met with my dissertation advisor and another member 9f my committee to help gain perspective on my study as it was evolving; used of data sources (e.g., surveys; interviews; content analysis of journals;. :comments during and outside class; and objective tests of students' reading abilities and field dependence{mdependence); : used "referential adequacy materials"; I recorded field notes (more than 40,500 a word-processing file after each class session and tutoring session. I kept a separate word processing file (more than 16,000 words) of notes on of data to theory and prior research, possible significances of the data, futUre research, and recommendations for future implementation of journal writing. ] .did not audiotape or videotape class sessions because I did not want to interfere with the primary purpose of the class-to facilitate student learning in an academic course. On the first night, when I was prepared to audiotape the class session, I. asked the students, "Would you tend to participate less if I audiotaped the class?" Since a few said or nodded their heads yes,44 I said I would not audiota. any of the class sessions. I used semistandardized interviews; I printed the quest:i,ons and took notes during each interview. And I photocopied each journal Group 1 after I had graded it 44one of ttuise who said she would falk less was Frances, one of my composition students from the previous semester. See her case study in chapter 4. 80

PAGE 94

To support: transferability (external validity), I used "purposive sampling" in interviews; based on my extensive teaching experience and knowledge of the population, I selected students who represented a cross section of the population (random selection of only five students from such a small sample could have failed to provide the necessary diversity); "thick description"; I grounded interpretations in the data, taking care to provide the reader with enough "low-inference descriptions of behavior and excerpts from .. .interviews" (Donmoyer, 1990, p. 196) or, in Phillips' (1990) term, "lowlevel observations" (p. 26) that the reader will be able to evaluate the conclusions which I draw or generate his/her own. I provide thick description either in the body of the text or in footnotes, and included several journals in Appendix K. To establish dependability (reliability), I used "overlap of methods," a form of triangulation; I used surveys, interviews, analysis of journal content and of comments during and outside class, and "dependability audit"; I left an audit trail of my methodology and decision points. To establish conjirmability (objectivity), I used "triangulation" (see above), "practiced reflexivity"; I continuously examined the purpose, methods, potential biases, etc., and used a "confmnability audit"; I traced each finding to its original data source; I have descijbed many of these in the body of the text or in footnotes, and have included several journals in Appendix K. 81

PAGE 95

CHAPTER4 CASE STUDIES The purpose of this study was to develop a cognitive model of journal writing in a college introduction to literature class. In this chapter I present the case studies of five students in Group 1. In chapter 5, I systematically address each of the main research questions of the study. Selection of Interviewees I would like to have interviewed all of the students. However, students' work and academic schedules and my professional obligation to assist students during office hours before and after class limited my access to them through interviews. ln selecting students to interview, I tried to span as broad a range of variables as possible. However, my prhnary focus was on the following traits (a swnmary of those selected is in parentheses): field (2 of the most field dependent, 2 of the most field independent, 1 in the middle), reading-comprehension level (1 sub-college level, 1 college-sophomore level, 3 at least college-senior level), age (2 who were below both the median and the mean, and 3 who were above), academic experience (2 who had essentially failed in early college experiences; 1 with a mixed record who had returned to college; 2 who had first enrolled in 82

PAGE 96

,, college as: :nontraditional students but both of whom had high grade-point averages).' Secondarily, I .was interested in the following traits: sex (3 of the 12 females, 1 of the 3 males), journal skill (2 very astute but not seeming to perform up to their potential in their journ'als, 1 very imperceptive but probably perfonning up to her potential, 2 seeming to perform up to their potential), and class participation (the 2 most actively involved in class discussions, the 1least actively it1::volved, and 2 somewhat-to-moderately involved). I was able to interview students who satisfied all the primary traits except age, and all the secondary traits. I would like to have interviewed at least one traditional age student Todd or Bonnie, who were also the most field dependent in either group and who carried the heaviest total loads); however, our schedules did not permit it. Because Todd was extremely field dependent, in chapter 5 I draw many examples from his journals to contrast the journals of the most field independent students. I interviewed Lisa and Colleen after they had written their third journal on a story, and Trish after she had written her fourth. Because of the time constraints, I was not able to interview Frances and Eric while they were writing journals on stories; I interviewed them after they had written their sixth journal, on the play 'night, Mother. For brief biographies of all the students in Groups 1 and 2, see AppendixN. 83

PAGE 97

These five cases illustrate that although instructional implementation could I benefit by ensiJ.ring that all students understand the purpose and criteria of journal writing, journil writing can be a flexible cognitive tool. These students all found journal writing helpful. With the exception of Eric, who never read the "Catechism ;, for Stories," ttiey relied more on job aids, such as the Catechism, at the beginning of the term than the end, gradually becoming more responsive to their internal learning and turning to the Catechism only occasionally to "get their wheels turning." I ... At the end of page 245 "It's just to let the air in." Something soon after that comment leads me to believe there is one particular thing wrong in [the relationship but again Hemingway doesn't tell us directly what it might be. If: I am correct in saying this is the objective point of view. It does make this .:giece ofwork hard to put all the pieces of the story together. Because the author does not give us any thoughts or feelings regarding what they are to do .... i Exce.rpt from Lisa's third journal Personal Backigound I At the age. of 36, Lisa had returned to college after 12 years. Because she often came to my office for help before class, I came to know more about her personally than I did most of the other students. I share some of the more personal details of her life to i!Juminate characteristics about her reading and journal writing strategies !; later in this and in chapter 5 and to demonstrate that being frank about one's 't personal life does not necessarily mean that a student is totally frank about his/her understanding of course material. 84

PAGE 98

Lisa is a person who always seemed sensitive to other people's feelings, but generally l;acked self confidence. She is in her second marriage, and about midterm contemplated leaving her husband, particularly because she felt their six year-old daughter would suffer in a marriage filled with strife. Her husband is a professional who "doesn't talk enough" and who has been unwilling to I; listen to any o( her anxieties and problems with her school work. She worked 18 hours a week secretary. Being a mother, working, studying and requiring extra rest to combat :a serious case of lupus did not leave much time for her with her husband. she said their marriage had been one loss or other problem after another: they one child, they had had financial problems which resulted in the loss of the hou;se they were buying; she has lupus, etc. Apparently our discussions helped Lisa her marital problems at least for the remainder of the semester. ,, Lisa was for six credits (Introduction to Literature and Beginning Algebra); altho.ugh she struggled with both, she earned C's. Her educational goals fluctuated durffi.g the term. At the beginning of the term, she said that her goal was a doctorate in psrchology. Later in the term, she focused on becoming an elementary I school teacher:. She had 68 cumulative credits, with a 2.8 GP A. She was field !; dependent, a score of 4 (out of 16) on the Hidden Figures Test. On the NelsonI. Denny Readin9 Test, she had a total reading level of 11.1 (vocabulary level, 12.7; comprehensiOIJ.' level, 9.4 ). She had perfect attendance in class, was almost always i. cheerful and participated moderately, although she was probably the least perceptive student in the class. 85

PAGE 99

Although she is not academically gifted, she was probably one of the most highly motivated and diligent students I have ever taught; her journals were the longest in the class (515;: 1,332; 1,159; 1,722; and 1,173 words-all but the last were at least 2.3 standard deviations longer than the respective averages). About midterm she said that the most important thing in her life was to complete her education and to become a teacher. She tried to relate whatever she was reading to her own life so that it would be more meanfugful and easier to remember; however, this sometimes caused her to impose her life schema on the work, not attending to the differences between her life and the story. She generally tried to apply whatever advice was given and to use the various aids available to her. She was the only one in either group to view the HyperCard tutorials Writing Journal Questions and Writing Journal Comments and the Introduction to Fiction program, and one day in class she told the students how helpful she had found the programs. However, she refused to accept the fact that she really needed to take a developmental reading course; I recommended that she take such a course before taking any other college courses because her reading problems were obviously interfering with her ability to learn. She was often unable to connect pronouns with their correct antecedents and to make other very low-level inferences. She generally .Pad a much larger percentage of incorrect journal items than did other students; for example, in journal3, 42.9% of her items were incorrect, 3.3 times the standard deviaP.on for journal 3. (But at least she was trying to make sense of the story; her 63 items were 3.6 times the standard deviation.) 86

PAGE 100

i In to staying after class occasionally to talk or to work on the HyperCard I' programs in tlie English lab, she came for help outside class eight times, 30 to 60 i minutes each. ','My extra help included asking her to elaborate on elements from the works we reaq' in class as well as reading and discussing other works. For example, ., as she read "Tlle Chrysanthemums" (a short story by John Steinbeck) out loud, I asked her queS,tions about setting, motivation, meaning of specific passages, theme, and We also read and analyzed several poems. But her high motivation to succeed, coupled with her lack of self confidence t might have interfered with her optimal performance (see Yerkes & Dodson's 1908 r. law on the between arousal and performance). For example, she said she suffered from test anxiety and discussed her academic problems with an academic advisor at least twice during the semester. She spent up to ten hours working on her journals for class; she even took off work one day to write her journal because she had been ill th day before, her day off (she lost $40 pay). She felt the suggestions I ,, ,' gave her during. the interview about journal writing were so helpful she asked if she could share with the class. ,, Although :she seemed comfortable talking with me, she did not give me a complete description of her efforts to understand assignments, either in personal r'' conversations pr in her journals. She was so anxious about understanding the stories when she her journals that she discussed at least some of them with others. She did tell me her.husband refused to discuss her school work any more, but she did not say that she discussed any of the assignments with other students while she was i: writing her In journal 5, she did not give credit to anyone for insight into the 87 ,I '''----------------------------

PAGE 101

masonic reference in "The Cask of Amontillado"; she wrote, "After I read this about 3 times and was writing my first journal it dawned on me that mason in this conversation means two things .... Yet during her interview, Frances said that Lisa had called because she did not understand the reference to masons. (Frances felt responsible because Lisa misunderstood her explanation of the ironic use.) Rather than using the journal as only one of the tools the class used in creating an understanding of a story, she wanted to get the "right" meaning before we ever discussed the story in class. By not operating within her own zone of proximal development (Vygotsky, 1978; see Glossary), she cheated herself, because her journals provided little insight into the specific points at which she became confused. Furthermore, she tended to become more confused as she tried to map much more advanced understandings onto her own misunderstandings. If she had used her journals to monitor her comprehension (as most students did), more of the meaning of a story would have unfolded for her, at the very least, I would have had a clearer understanding of the real sources of her confusion and been able to have provided more appropriate feedback and remediation. Since some of the other students also tried to over-reach or ignored their real difficulties, this problem suggests the need to refine future implementations of journal writing to try to make sure that students understand the functions of the journals. Responses to Questions (Fifth Week) I interviewed Lisa after she had written her third journal (on "Hills Like White Elephants"). Her responses reveal her high level of motivation, including her willingness to try to take advantage of resources and to invest as much time as she 88

PAGE 102

needed to try to assimilate the material meaningfully. Her answers seem to reflect her field dependence; that is, she does not attend to details and apparently prefers to have organization provided rather than to impose her own organization. Nevertheless, like several of the other students, she was initially uncomfortable submitting unpolished journals a carry-over from writing essays). Following are excerpts from the interview. (Note: At the beginning of each interview, I emphasized that the student's responses would in no way affect the student's grade. I urged each student to be as honest as so that his/her answers would give me insights which would help me help him/her as well as future students. Whenever a response indicated the need for an instructional intervention, I provided some help on the spot. If the student needed additional help, I arranged to meet with the student at another time or indicated that we would address the issue in class.) How did you feel about writing the journal on "The Story of an Hour"? .. 1 think I've taken a long road to journal writing. I put 5-8 hours on these journals .... It takes a long time for things to sink in my mind. I miss details .... Do you feel that journal writing is helping you? Yes. But it's time-consuming. It takes me about45 minutes to get into the flow ... it makes me think. Did you find writing the journal helpful? Yes In what way? The Catechism helps me think. And "How to Read a Story" in the computer program to Fiction]. That's excellent .... I enjoyed that a lot That's why I told everyone they should see it .... 89

PAGE 103

How did write your journal? (while reading, after reading?) When I have more time [writing journals on the weekends rather than during the week], I re(ld through once without writing, let it sit, come back, try to get the keywords, .... Then I go through and pick it apart. Then I go through and really pick it apart. I write in my study area, with complete quiet. [Note: An analysis of her journals confirms that she does indeed pick each story apart, generally applying questions from the Catechism. Her journals were the longest in the class; her frrst four were 2.3 to 2.8 standard deviations longer than the class averages; even the fifth, by which time many students had significantly increased the length of their journals, was 1.6 standard deviation longer than the average.] Why did you choose that approach? [She said she needs complete quiet and needs time to get into the flow and be able to concentrate.] How many times did you read the story? At least6. Did you up words you didn't know? Immediately. I write them on separate papers. Yes [I suggested that she write the definitions in the margin of the story so that the ', meanings are immediately available when she rereads. She asked about how to learn new vocabulary words. I told her that when I look up words I like to look at the derivation-root, preflX., etc.-and that that often helps me remember the meaning.] 90

PAGE 104

Did you try to answer your own questions? I Yes, but I don't get anywhere .... You have to remember that I've been out of school [college] a long time, 12 years. Did you ask how the story made you feel? Yes I I But with this story I felt like I could relate because I'm sick all the time [During a 1,' previous meeting, she had told me about her problems with lupus, and that she had been menstrually for two months]. Did you try to predict what would happen in the story? No Why I didn't take enough time. [I said she's spending a lot of time.] I teach first-grade religion (Catholic Church) and when I read to them I stop and ask what do you think is going to happen. I ask tMm. but not myself. Maybe I should ask. myself too. [I asked why she hadn't. She said it never dawned on her. I suggested it might be because she was busy doing so much other thinking.] Did you trr to identify the elements in the story schema? Yes I" I want to tfy to apply the terminology. Did you what meaning you got from the story? No Why not?:. I think I'm! so wrapped up in trying to figure the story out. I Did you write all your questions? I 91

PAGE 105

I don't think I put enough in. I think I think about them in my head and then try to answer [without writing the questions down]. I go back [after I've written my journal] and ask myself questions that I think you'// ask on the quizzes. [I suggested that if she didn't take time to enter her journal on the computer that she would have more time to add the questions she had as she studied for the quizzes.] Did you change the way you wrote this journal compared to your journal on "The Appointmep.t in Samarra"? No But I think I added computer [Introduction to Fiction program] gave me more ideas on how to read a story. Did you change the way you wrote your journal on "Hills Like White Elephants" [due for that evening's class]? I took off work today so I could do the journal 9:30-11:45, then went over to my neighbor's to type it. But today' s [journal] was pressured because of time [she said she was unable to work or study yesterday because she had had an endometrial biopsy and had been in great pain]. [I asked why' she typed it?] Because/' d be too embarrassed to turn it in. [I advised her not to waste time copying or typing. But she said she changed her mind on some comments as she typed.] [I advised her to just write in the margin "I misread. See. below." Then she'd be able to track her changed views. I also said maybe the first view was right. She said she tries. to put it in good structure, e.g., all the thoughts on a given topic in a single place. I advised her just to deal with things as they occur to her.] 92 --------nw.-t-. _________________________

PAGE 106

Summazy During the semester, Lisa's strategies for understanding stories and her confidence in her ability to write journals changed. In the Survey of Background on the first day of class, she checked that she was "Somewhat" to "Very Unconfident" about her ability to write a journal on an assigned short story, and described mainly a brute force, externally directed method for making sure she understood what she read for a class: 1. Read it through once w/full concentration 2. Read again, & highlite what I feel is important to me-that I will remember 3. Review my highlights & corresponde45 with what the instructor has been teaching 4. To study both high-lighting & Instructor notes I took in class. Mter her last journal, she checked that her confidence had improved 4 on a scale of 5, but on the last day of the semester, she checked that her confidence in her ability to write a journal had improved "Very Much" (5). Even though she failed the midterm examination, in the final survey of journal writing Lisa rated journal writing as "Very Helpful." She focused on journal writing as a tool for helping her understand the material. Her journal average (9.3) reflected her efforts to ask question and understand the reading, rather than the perceptiveness of her questions. and comments. She had nothing negative to say about journal writing and, in response to suggestions for making journals more helpful, wrote: "I felt the software package [HyperCard tutorials on writing journal questions and 45To ground the data in the students' character, unless the meaning would be unclear, I have retained the students' spelling, grammar, and mechanics. With few exceptions, I have not marked the students' errors with "sic." 93

PAGE 107

comments] was of great assistance to me." However, like many field dependent learners, she at least initially seemed to want a recipe for reading stories and writing journals; her behavior was also reminiscent of Putnam's (1991) description of novice teachers: First, attention focuses on the recipe itself, and there is little understanding of how it fits into a larger sequence. Thus the novice quickly gets stuck and, quite appropriately, lacks confidence in his or her ability to follow through. (p. 151) Lisa seemed to view reading and understanding a short story almost as if it were an algebraic equation, with each part of the equation neatly enclosed in one part of the story-rather than a dynamic interaction of all of the story elements as the story unfolds. She wanted a rigid procedure, similar to doing simple math. Rather than identifying and communicating the questions that arose from reading each story, she turned to the Catechism. Her journals rarely contained questions such as "What does the passage about the masons mean?" (the question she asked Frances but did not write in her journal). Moreover, when she used the Catechism, she applied only the lowest levels, answering factual questions (e.g., What is the point of view?), rather than exploring the high-level issues (e.g., What did that point of view contribute to the meaning?). About the third week of class, during class discussion, she asked why it is important for the reader to try to change the various story elements as the reader thinks about a story; she said this had confused her while she was studying the discussion of the Catechism in the Introduction to Fiction HyperCard program. Although the program provides several illustrations and during class we frequently explored the significance of such changes in assigned stories, she never seemed to fully 94

PAGE 108

She often seemed to be reading as though all meaning is contained in the words and all she had to do was know the definition of each word-the meaning of the story would automatically follow. For most of the course, she seemed not to understand the role the reader plays in creating meaning from a knowledge of word meanings (text based) and world knowledge. At times, she imposed her experiences on the text situation (e.g., imposing her occasional death wishes because of her lupus onto the death wish of a housewife who has "heart trouble" but is also unhappy in her marriage and wishes that life might be short). She was very much a novice at reading literature, not adapting her strategies to the demands of the situation. She always used the same strategy, unlike some of the other students, such as Trish, who varied her strategies. By the end of the term, her reading strategies had begun to include more questioning, but in her journals she was still attending to external questions rather than to her internally defined needs. Thus, she continued to apply the Catechism before making sure she had a basic understanding of the text. For example, in journal 3, on "Hills Like White Elephants," she never once mentioned the operation (although she quoted passages that refer to it) and said that the reader is never told the subject of the story (although the story clearly indicates that the immediate subject is whether to have, the operation). And she explained what she thought the denouement of "Hills" is, yet said she did not know what the subject of the story is. Overgeneralizing a student's frankness can have important implications for instruction. In Lisa's case, because I was initially unaware of her unwillingness to share all of her points of confusion, I was unable to address appropriately. 95

PAGE 109

If her continues to develop, I believe she will attend more to her own questions, rather than to those provided by an external source. But this shift will probably depend on her participating in a course on assertiveness training. She said that as soon as she gets her AA degree (December 1992), she intends to take such a course, in spite of the cost (about $150). I encouraged her to take it sooner, noting that developing her ability to assert herself would help her in her academic work as well as in her personal life. Addendum. During the second week of the following term (fall 1992), Lisa to visit during my office hours. She said she wanted to thank me for helping her so much during spring semester .. She said she was taking philosophy (and loving it), and was writing journals in the class.46 She said writing journals really helped her understand what she was reading in that course just as it had in the literature course. In the tenth week, she told me that she was earning an A in her philosophy course. Colleen What type of "operation" was the girl going to have? What is meant by "just lt:?t the air in"? Is this an actual physical operation? I'm completely baffled by this statement made by Jig when she said, "And we could have all this, and we could everything and every day we make it more impossible." Excerpt from Colleen's third journal 461 talked with the instructor later. He said Lisa is "a good student" He said he gives the students extra credit for writing journals but that the amount of extra credit influences only a few students with borderline grades. He said the students who write journals are the good students anyway. Nevertheless, I believe that students' use of journals in the course, as well as their perceptions of the helpfulness of journal writing, deserves careful examination. 96

PAGE 110

Personal At the age of 36, Colleen had returned to college to change careers. She had studied in the legal assistant program at ACC for three years (1986-89) and worked three years in law, then decided she wanted to become a teacher. She was married, with two children (a 17-year-old from her frrst marriage, and a 7-year-old), and worked full time (35 hours a week) with special-needs children. She had 34 cumulative credits, (mostly in the legal assistant program), with a 3.8 GPA, and was registered only for the literature course, in which she earned an A. She was moderately field dependent (5), with a vocabulary level of 15.1 and comprehension and total reading levels of 16.9. She participated in class regularly (with only one absence) and was very perceptive. Like Lisa, who was also field dependent, she found the Catechism helpful. But, unlike Lisa, she seemed to use it more as a set of heuristics than as a recipe. This interpretation is supported by the change in her rating of the Catechism during the semester; her rating dropped from 5 on the first survey to 4 on the second, when she elaborated: "I probably did not utilize it to the fullest-but it did help to give me a sense of direction for my journals." Like Lisa, was gentle and sensitive, but she was not as anxious as Lisa about expressing her questions and comments. On the first day of class she checked "Undecided" about her initial confidence in journal writing, and elaborated: "I'm not afraid to express my questions and comments, but I'm always a little 'unsure' of myself at frrst." As she gained confidence and expertise, her journals got longer (61, 329, 435, 682, and 885 words). Again, like Lisa, her confidence continued to 97

PAGE 111

improve even after she had written her last journal. Mter the last journal she checked that her confidence had improved "Somewhat" (3), but on the last day of class, she checked that her confidence had improved "A Lot" (4), and elaborated: "I have learned to organize my thoughts more, and express them in writing." Also like Lisa, she consistently rated journals as "Very Helpful." But unlike Lisa, who always seemed anxious about her journal writing, Colleen wrote on both surveys of journal writing that she "enjoyed" writing journals. For her, the most helpful aspect of journal writing was "Knowing that I could be honest about my opinions-feelings. Then I would generally use the Catechism as a rough 'outline' of areas of importance." Her journals (average, 9.4) contained probing questions and insights. Remonses to Intecyiew Questions (Fifth Week) I Colleen after she had written her third journal (on "Hills Like White Elephants"). Her answers revealed her to be more metacognitively aware than Lisa was; for example, she was aware of whether she had questions (on some stories she did, and on others she did not), and she made a conscious decision not merely to tell what happened in a story but to explore the reasons why. Following are excerpts from the interview: Did you find writing the journal on "The Story of an Hour" [the previous story, whose journal I had returned] helpful? Yes In what way? I like to write the journal. It helps me clarify my thoughts when I put them in writing. But I feel I have to guard against saying too much. 98

PAGE 112

[I asked wliy.] Because you lose your idea if you keep rambling. [I suggested, "Like adding an extra can of water to the chicken soup?" She said, yes.] How did you write your journal? (while reading, after reading?) First I read it through completely. Then I go back (I have an idea of) what it's about. I jot down notes, and then I type it on computer) because it's easier to change my thoughts. [I suggested not changing, just adding so she could track her thinking.] How many ,times did you read the story? 4 Did you try to answer your own questions? I didn't have any [questions on "The Story of an Hour"]. On "Hills [Like White Elephants]" I tried. Did you ask how the story made you feel? Yes Did you try to predict what would happen in the story? No Why not? It flowed so well,[/] didn't need to. Did you ask what meaning you got from the story? Yes Did you write all your questions? I didn't have any. [But she said she did ask her questions for "Hills Like White Elephants," which she found totally confusing.] Did you change the way you wrote this journal compared to your journal on "The Appointment in Samarra"? Yes 99 !::..... __________________________

PAGE 113

In "The Appointment in Samarra" I mainly just put what happened, but in this one I put more of my personal feelings and said why I interpreted the way I did Why? I felt I needed to write more to explain my ideas. Did you change the way you wrote your journal on "Hills Like White Elephants"? Yes Please explain. [She said she had read the story at least five times.] I had a very difficult time. I had so many questions. There was so much I didn't understand. I even had my husband read it .... [She said her husband knows a lot and always thinks he knows what things mean.] He said, "Hell, I don't know what that means." Summazy While Lisa remained primarily at a novice level in both reading literature and writing journals, Colleen operated at the journeyman level or beyond for most of the semester. The excerpts from their third journals (above) illustrate the difference; Lisa superficially applied the Catechism, ignoring the fact that she apparently did not understand the phrase "let the air in," while Colleen confronted her confusion by asking questions. Although Colleen rarely addressed issues such as why a given point of view was most appropriate for a story, she used journals effectively to monitor her comprehension. 100

PAGE 114

... What kind of operation are they talking about? More than likely they are discussing an abortion. There are a number of key phrases which I believe support this supposition. The American says, "I'll go with you and I'll stay with you all the time. They just let the air in and then it's perfectly natural." He also assures her they will be fine afterward, as if they aren't already. He tells her they'll be just like they were before (before what?-the pregnancy?) He also says, "That's the only thing that bothers us. It's the only thing that's made us unhappy." Excerpt from Trish' s third journal Personal Background Trish seemed in many ways typical of very bright students who are unsuccessful on their first efforts in college. Single, aged 25, she had failed or withdrawn from six of the eight colirses she took between the ages of 18-22 (1984-88). As she said, she has been "in and out of college for years; it took me a long time to be ready." She lived with her family, worked full time as a secretary/paralegal, and attended college at night. Her tuition was paid by her employer, who she said has been very supportive of her pursuing her education. She said he believes she is a good role model for fellow employees because she is such a hard worker. She would like to study foreign languages, probably in international business, something that involves travel. She is fluent in Spanish, would like to learn Japanese and/or Chinese or Korean (her brother is stationed in Korea and she was looking forward to the opportunity to visit in two months). She had six cumulative credits, with a 1.7 GPA. She took six credits, including Introduction to Literature because her younger sister, who will graduate with a BA in creative writing this year, persuaded her that she needed to take the course in order to develop a different facet of her mind. She 101

PAGE 115

earned an A in Macro Economics and a B in literature, but was certainly intellectually capable of earning an A. Her Hidden Figures Test score (8) indicated that she was balanced in field dependence/independence, although she said that she is very rational and that if she hadn't had a migraine headache the night that the class took the Hidden Figures Test she would have got all the answers correct Her vocabulary, comprehension and total reading levels were all16.9. She had perfect attendance, was the second most active participant in class discussions and was very perceptive. Like Colleen, Trish seemed to use the Catechism more as a set of heuristics than as a recipe, but she consistently rated the Catechism as "Very Helpful." On the first journal survey, she wrote: "It helped me to look for things I may not have known or remembered existed." The Catechism continued to serve as a prompt for "working through the stories and plays." She was the only one who fairly consistently explored the higher-level questions on the Catechism, such as what would be gained or lost if the setting were changed. With the exception of her journal on the fifth story, which she had studied in another class, her journals progressively got longer: 360, 472, 923, 1,326, and 815 words. She was one of the most self-confident students in class, but not at all cocky. On the first day of class she checked "Somewhat Confident" about her initial confidence in journal writing, compared to Lisa's "Very Unconfident" to "Somewhat Unconfident" and Colleen's "Undecided." Nevertheless, like Lisa and Colleen, she indicated that her confidence improved significantly during the semester; it remained constant at "A Lot" on the two surveys at the end of the term. 102

PAGE 116

Unlike Lisa's and Colleen's consistent ratings of journal writing as "Very Helpful," Trish's ratings increased from 3 ("Somewhat") to 4. Her journals (average, 9.9) were probing and insightful. Res.ponses to Interview Questions (Sixth Week) I interviewed Trish after she had written her fourth journal (on "Cathedral"). Her answers revealed her to be probably even more metacognitively aware than Colleen was; for example, she was consciously selective about which elements of the Catechism she focused on. Although she intuitively changed her strategy when she encountered a story that she couldn't understand, she was aware of her strategy changes. Follo:wing are excerpts from the interview: How did you feel about writing the journal on "Hills Like White Elephants"? That one I felt very unsure about. Probably the third reading I started taking notes ... Normally I take notes right away, but I found this story very confusing .... Did you find writing the journal helpful? Yes definitely. In what way? It helped me break it up and decide, "What's going on here?" What in the story is making it this way? I'm using the Catechism. How did you write your journal? (while reading, after reading?) [She said she read it three times before taking any notes; then jotted down notes.] "I must had a full page of cat scratch. I put it down and said it was time to go to sleep. But I still had voices talking to me. I kept adding things .... [Then she 103

PAGE 117

had to stop! and go to sleep.] The next day I went to work a little early and put my thoughts into writing. I elaborated .... Why did you choose that approach? It wasn't an intended approach. It was the one that came easiest to me. How many times did you read the story? 10 [times, altogether]-3 times before [taking] any notes, then several times while taking notes. At work before I began putting thoughts into writing, I read it once. I read it again to study for the quiz. Did you look up words you didn't know? Yes [She said she knew what absinthe is, that she had tasted it, but] I looked up absinthe because I wanted to apply the direct meaning; I look up [unknown] words the first time I read them. Did you try to answer your own questions? To some extent I think I tried. But I can't say I gave it a devoted effort-probably because the answers just weren't there. Did you ask how the story made you feel? Yes and No In my mind I did, but not in my journal. Did you try to predict what would happen in the story? No Why not? Because I dldn' t know where the story was coming from, I couldn't say where it was going. Did you try to identify the elements in the story schema? I tried to pick the ones that were most relevant to find the meaning of the story. 104

PAGE 118

Did you ask how the story would change if any of the elements in the story were changed or: how important a given element is to the meaning of this story? Yes Especially with the setting. Did you ask what meaning you got from the story? Yes I don't think I asked it directly that way so much as applying it to human experience. Did you write all your questions? I probably did not. I probably answered some as I asked them in my head. [I commented that often people's journal comments are responses to implied questions, i.e., questions they did not actually write in the journal.] Did you change the way you wrote this journal compared to your journal on "The Appointment in Samarra" _and "The Story of an Hour"? Yes Why? In comparison to those two, I kinda used the Catechism more ... because I didn't feel I was using all the resources available to me in the first two. But I tend to take notes and elaborate later and that was the same for all three. I waited longer before I started taking notes because the story was overwhelming. I didn't know what direction he was coming from or going to. And then I began to pick up on the meaning. It's kinda like walking in on the conversation oftwo people you don't know .... Did you change the way you wrote your journal on "Cathedral"? Yes 105

PAGE 119

This time I just started asking questions and writing. Because I wanted to see if there was something I had missed in the story. It's a very direct story and I think logically and I wanted to see if I was missing something the writer was trying to imply. Is there anything else you would like me to know? It's real helpful. I don't mind doing it at all. There are some classic stories of Hemingway I've read. I was reading [this story] but I wasn't getting the meaning. This [journal writing and using the Catechism] is helping me get it. What Lisa said about keywords [when she gave students suggestions about journal writing after her interview February 19]-keywords are very important. If a word jumps out on the page, I know it's significant. [I asked how it jumps out.] It just jumps out. Like "absinthe" [in this story]. I wanted to know why he used that. Summazy Her fmal comments suggest that she intuitively perceives significant elements of stories, but then, as her earlier comment about looking up the definition suggests, she exercises her rational powers to try to discover the exact significance. She seemed to have more expertise in both reading and journal-writing than Colleen. Her description of her reading strategies and her performance on the Nelson-Denny Reading Test confirm the former (Trish placed in the top percentile-grade level 16.9---on comprehension, vocabulary and total reading, while Colleen placed in the top percentile only in vocabulary and total reading; she placed at the 15.1 grade level 106

PAGE 120

in comprehension). Trish's strategy for journal writing confinns the latter; her journals explored not only the meaning of a story but at least some of the principles involved in its structure; for example, what would be lost by changing the point of view from the husband to that of the wife or the wife's blind friend in "Cathedral"? But she had not developed to her full potential; for example, she did not explain why the story would not be more effectively told from the limited oiriniscient or omniscient point of view. Nevertheless, she was the most competent journal writer in Group 1 or2. Frances This story reminds me of a dance people did during the 18th century (I think) where the man held the tips of the woman's fingers and they walked side by side; then the pattern of the dance took them apart from each other for a time. They come back together during the dance, but never really touch each other. I feel the man has talked her into an operation (abortion?), and she really doesn't want to do it. He thinks after this operation things will be the same as before, and they will be happy. She is wiser and knows things will never be the same. She knows they will never again "own the world." Excerpt from Frances's third journal Personal Background At the age of 33, Frances started college for the first time, two years earlier, in January 1990. She is married, with three children, aged nine and under. She averaged 18 hours per week at three part-time jobs (all of whose work she did at home, sometimes working 50 hours per week, sometimes not at all). Her educational goals are complex, a mixture of self-development and economic necessity. In a note to me, she wrote, 107

PAGE 121

I have three career choices to make. My first love is philosophy, I'll keep that as a hobby because I can always read. Psychology is interesting and I have a personal interest in helping people from dysfunctional families. But the time it will take me to complete my education is prohibitive .... Accounting is also fun for me and the degree requirements are much less .... Of course if I am earning a good income as a Public Accountant then I may decide to get a degree in philosophy for fun. Maybe I'll just go to school for the of my life .... But whatever she does, she wants to be her own boss. As a child, she and her sisters were abused by their grandfather and mother; she still suffers from the emotional scars, most noticeable in her tendency, as she said, to want to keep hidden. She is very astute, but still fairly reticent about participating in the class discussions. Nevertheless, she has made great progress. In my composition class the previous semester, she had essentially not participated at all in class discussions, but by the middle of this semester she participated once or twice during each class. And often she stayed after class to talk or attached letters to her journals. She had a reading level of 16.9, and is a voracious reader, with, as she reports, more than 2,000 books in her personal library. She wrote in the first survey, "I read so much that I read reminds me of something else." Her intellectual curiosity sometimes leads her to research topics in fiction; as she explained in one journal, Some times if I read a piece of fiction it makes me curious and I do research on a subject. For instance, I once read a book called The White Horse Goddess about a Goddess of the Celts in prehistoric Europe. It was fiction, but because of that book I spent a considerable amount of time researching Celts. She had 18 cumulative credits, with a 3.0 GP A. She registered for nine credits, including accounting and statistics, but dropped the latter in the twelfth week. She was field independent (12), with vocabulary, comprehension and total reading levels 108

PAGE 122

at 16.9. Her one absence was due to scarlet fever. She earned an A in both literature and accounting. More field independent than Lisa, Colleen, and Trish, she used the Catechism less often: "If I drew a blank[,] referring to the Catechism helped 'get the wheels turning."' Her journals rarely systematically addressed items in the Catechism. Nevertheless, she rated the Catechism 4 ("Somewhat Helpful") on both surveys. Her response to the Catechism suggests that it is a flexible instructional/learning tool; it can be an effective resource even if a student doesn't use it all the time. On the first day of class she checked "Undecided" (3) about her initial confidence in journal writing, and explained that that was only because "I have trouble deciding if my opinion is valid as far as knowing what the author had in mind. I don't always question what I read, but just accept it at face value." Like Lisa and Colleen, her. confidence continued to improve even after she had written her last journal. After the last journal she checked that her confidence had improved "4" (out of 5), but on the last day of class, she checked that her confidence had improved "A Lot" (5), and elaborated: "I have learned a journal is a method for me to focus my thoughts." The positive shift in her appraisal of journal writing was significant. She only rated journal writing as "Somewhat Helpful" after her first journal, commenting in response to the most helpful aspect, "I like this particular story so its easier to write about. If I don't like a story I fmd it harder to think of questions or make comments." Mter the last journal she rated journal writing as "Very Helpful," commenting that the most helpful aspect of journal writing was that it "forces [me] to think about what I'm reading on a conscious level. Instead of passively absorbing something I have to 109

PAGE 123

find the words that express what the story means to me. That is still a laborious process." But like Colleen, she enjoyed journal writing; she noted, "It's fun!" One interpretation of her response is that she found in journal writing a discipline by which she could safely begin to liberate (or was even "forced" to discover) her thoughts. Her journals (average 9.5) became increasingly probing and insightful. Res.ponses to Interview Questions (Eleventh Week) I interviewed Frances after she had written her sixth journal (on the play 'night, Mother). Her answers revealed (a) a high level of metacognitive awareness, including the realization that some understanding is tacit and difficult to tap; (b) an awareness of the power of a journal to probe intuitive understanding and to objectively examine what one does not like; (c) a reluctance to submit a journal that does not satisfy the writer (and thus throwing it out and starting over); and (d) Frances's flexibility in fmding approaches to the different stories and plays that were interesting to her. Following are excerpts from the interview: How did you feel about writing the journal? I was stressed for time [on this journal] so I was kind of in a hurry. Stories like that interest me because people do that (get that hopeless feeling) all the time. I liked writing it ... because I had definite opinions about it. Overall I've. enjoyed writing them ... If I don't like the story I have to try a lot harder. But/ don't think there were any I didn't like this semester. Did you find writing the journal helpful? Yes In what way? 110

PAGE 124

Because I did have some questions about it [the play]. I'm not possessive of my kids and I didn't understand why she said, "I thought you were mine." It helped to put into words that she [Mama] might feel either possessive or that their personalities were similar. I think when you're not sympathetic with a character it helps writing it down, maybe. It helps to clarify your thoughts. And ifyou have to [are required to] write something you have to look for something and it may make it click. It helps me understand the characters. How did you write your journal? (while reading, after reading?) Usually after reading the second time. I don't read it as an assignment the first time. [I asked what she meant.] If I read it as an assignment, I lose it. It goes in one ear and out the other. I have to take notes. If I read it for pleasure, I can sometimes remember it almost word for word. How many times did you read the play? Twice: read, watched movie, read and watched together. [Note: Students were required to read the play as well as to watch the movie and to compare them.] Did you look up words you didn't know? No Why not? [She said she did not recall not knowing any of the words. She said she usually gets the meaning from the context, but if a word is totally foreign, she does look it up later.] "Stopping to look up words destroys the flow." 111

PAGE 125

Did you try to answer your own questions? Yes Did you ask how the play made you feel? Yes A lot of times /' m interested in my own reaction because I don't always know why I react the way I do. I didn't understand why I reacted the way I did to the play and wanted to know why. [See below. She explained that she discovered that her reaction to this play was tied to her associating Mama with her grandmother. But she did not write that in her journal. I didn't think to ask her why she hadn't. See follow-up below.] Did you try to predict what would happen in the play? Not consciously, but I knew intuitively .... That's why /liked "The Unicorn in the Garden" [a story on the mid-term]-it surprised me. /' m not surprised by outcomes in stories often. [See below.] Did you try :to identify the elements in the play schema? (setting, protagonist, antagonist, conflict, etc.) Yes But I got about half-way through and decided it was too easy. But I still have a problem with the conflict and climax [in the play]. [We stopped and discussed these briefly. Our discussion made me realize we needed to discuss these more in the next class session. I told her we would. She did not ask about the conflict and climax in her journal. I did not think to ask her why not. See follow-up below.] Did you ask how the play would change if any of the elements in the play were changed or how important a given element is to the meaning of this play ? No Why not? 112

PAGE 126

Maybe I'm too accepting of what the author says. Did you ask what meaning you got from the play? Yes But it took me a long time. I went to bed wondering what I thought of it ... 1 think it took me until after I'd watched the video and then read it as I watched it. It helped me when I realized that I associated Mama with my grandmother ... it would be hard to live with her. [In her journal she did not say she associated Mama with her grandmother. See follow-up below.] Did you write all your questions? Yes Did you change the way you wrote journals from one story to another? Yes Why? It doesn't seem I do any two the same .... With "The Cask of Amontillado" I was playing with the method of following the Catechism. With this one, it got boring, so I just wrote what I understood. [Nancy apparently threw out at least one journal draft, rather than submitting her first attempt and then her later approach. I have tried to encourage students to submit everything that they write, but they bring a very strong mental set about the quality and purpose of their assignments.] Is there anything else you would like me to know? The hardest part of these journals is trying to put into words my impressions of the stories, because it's all intuitive. A lot of times I really don't have any questions ..... 113

PAGE 127

Follow-up: A few days later, I asked Frances to explain a couple of her answers. Why hadn'.t she submitted the Catechism she started on 'night, Mother? She said she thought what she was saying was "obvious" and just decided to shift gears. I probed to try to see if there were another reason (I thought she might have been reluctant to express her thoughts in the journal). I said that she could have shifted gears and observed in the journal the reason why. She said she tends to ramble. Still not convinced that that was the reason, I probed some more. Although I am still not convinced that we got at the real reason, we decided that probably she didn't want to saddle me with extra reading, even though she said that she knew that I had said that was not a problem. Since she wrote in one note that, having been abused as a child, she is still reluctant to draw attention to herself by participating in class discussion, it is possible that she also is anxious about drawing too much attention to herself in her journals, particularly if, in her view, she "rambles." Nevertheless, the fact that her journals significantly increased in word length as the term progressed (135, 197, 343, 473, 993 for the five journals on stories) suggests that she became less anxious about writing them as she gained experience .. Although it is also possible that she (merely) had more to say, this does not seem likely since she was prolific in her essays in the previous semester, in her journals in the following semester, and in personal notes which she appended to various assignments. In her journal, why hadn't she mentioned the connection she made between Mama and her grandmother? 114

PAGE 128

She said she wasn't aware that she was making the connection until I drew a parallel with "A Good Man Is Hard to Find" during the discussion of the play. Summary Frances flexibly adapted her approaches to reading and to writing journals to achieve learning that was both meaningful and enjoyable. She began as a novice at journal writing. and became very good. Although she had read extensively, her metacognitive skills were poorly developed; by the end of the term, she had developed much more powerful skills in critical reading. In journalS, she wrote: Yes, these journals are a help to me .... As much as I read I've never learned to read critically .... The point of all this is that I never learned (bothered) to take a story apart and look at it I've always accidentally picked out a phrase or image that was especially appealing, but now I'm doing it more often, and I'm spending more time contemplating the author's motive, even in my "fun" reading. Although she tried to apply the Catechism systematically, she gave up that approach because it "bored" her and generally focused on the hwnan relationships revealed in a work. Eric I think they are talking about an abortion, but I'm not sure. The American Man said that it was a "simple operation" and then states that "It's not really an operation at all." The only thing that I can think of that would not really be an operation, that would relieve the tension or unhappiness that they both apparently feel, would be an abortion .... Excerpt from Eric's third journal 115

PAGE 129

Personal Backuound Eric was the prototype of the bright male who fails out of college-having joined a fraternity and invested most of his time in drinking and socializing-but then realizes the importance of an education. He had actually failed out of two universities (the first was a large public university). When he saw himself in the same pattern at a small private university, he quit during the first term to get his life together. Working construction jobs during the summers and operating his own ski tune-up shop during the winters, he ftad saved enough money to return to college; he had enrolled at this community college at the age of 23 so he could raise his GP A enough to qualify for admission to a baccalaureate school. He was single and worked full time at his shop. His goal is a BA in business. He had approximately 165 cumulative credits, with an A in his single prior course this community college; he was registered for six credits, including Introduction to Business. He was very field independent (16 out of 16), with vocabulary, comprehension and total reading levels of 16.9. He enrolled in the class, with my permission, at the beginning of the second week, and had two absences. He w.as one of the brightest, most articulate, and most likable students I have had in 25 years of full-time teaching, always cheerful and eager to participate; his questions and comments were insightful and stimulated much class discussion. (I believe students.liked him more than students have ever liked any other fellow student I have ever had). He earned an A in literature and business. Although he rated the Catechism as "Somewhat" helpful (3) on the f'rrst survey, on the second he rated it "Not at all Helpful" (1) and added that he didn't think he had read it. Given his extreme field independence and his statement (see below) that he 116

PAGE 130

had no idea what I wanted on the first journal, it is likely that he hadn,t consulted any of the written instructions or models. Helping students like him understand the criteria for journal writing may require forced training, such as requiring them to view the HyperCard tutorials on writing journals and discussing in class what they learned from the tutorials. There is no record of his initial confidence in his ability to write journals (through some oversight, he apparently did not complete a Survey of his Background), but after the last journal he said that his confidence had not improved much (1) and added the comment: "Apart from the first one when I had no idea what you wanted." But his confidence rose to "A Little,, (2) by the last day of the term, when he added: "I didn,t really change the way I wrote them and never really improved on grades, so my confidence never rose either." Although he said that he had no idea what I wanted when he wrote his first journal, he found journal writing "Very Helpfur, and explained the most helpful aspect as follows: "I found that by writing my questions down on paper, I had a chance to really think it through and formulate some of my own opinions about it, instead of just becoming aggrevated and pushing the story aside." Instead of identifying any "least helpfur, aspect, he wrote: "I found journal writing very helpful." After the last journal his rating of journal writing had dropped from 5 to 4, but his view of the most helpful aspect remained basically the same: "The fact that I had to really think about the piece before I got to class." Although he said the least helpful aspect was "That it took more time to get the homework done," in the open ended comment on journal writing, he wrote: "Keep it." 117

PAGE 131

He was the only student who indicated on any of the surveys or during interviews that he read each of the works only once; he apparently chose not to take the time to reread a work in order to delve more deeply into it. His self-description in the interview suggests that "lack of self-discipline" rather than lack of time was the main reason for not rereading works. Although he was very bright, it is likely that he would have benefited from reading the works more than once-students who found journal writing most helpful tended to connect multiple readings to deeper insights. For example, Frances said she always read a work at least twice, first "just for pleasure." Trish said she read each story several times but that she varied her approach, depending on the difficulty of the story. Unless a work totally puzzled her, she took notes during the first reading; but she ultimately used the Catechism to prompt her to consider various aspects of each story. But because Eric was fixated in his approach and unwilling to invest more time in thinking about it, he seemed to have reached a plateau in his journal writing and was unable to raise his journal average above 8.3 (his journal grades ranged from 8 to 9). When he thought he understood a work, he did not explore the significance of individual elements (e.g., How does the setting contribute to the meaning? What, if anything, would change if the setting were changed?). In spite of their incisiveness, the length of his journals remained relatively constant and typically in the bottom quartile of the class: 140, 158, 198, 179, and 187 words. Furthermore, in his interview, he said, "It just seems pretty clear, so I never really stop to think about it. In stories I never before realized there could be different points of view, so I do think about that." Thus for students who are reluctant to use highly 118

PAGE 132

structured job aids (as field independent learners tend to be) it may be necessary to implement other strategies, perhaps providing specific questions to think about or providing alternate assignments (e.g., "Rewrite passage X in at least two different points of view and explain what is gained and/or lost"). Remonses to Interview Questions
PAGE 133

How did you write your journal? (while reading, after reading?) After reading. Why did you choose that approach? I like to read it and get a feeling for the flow. Sometimes if you have a question while you're reading, it's answered by the time you get to the end. I do circle sentences or parts of paragraphs. How many times did you read the play? I read it once and watched the movie once, at the same time. During the silences I'd go back and forth between the text and the movie. [Note: Students were required to read the play as well as to watch the movie and to compare them. Although I asked them to read it first so they would develop a schema for the stage play, Eric was apparently interested in saving time.] Did you look up words you didn't know? No Why not? I could get the meaning out of the context. Usually, if there is a word I don't know, somebody in class always asks. [I commented that a person can get a misinterpretation that is difficult to change later. He said, yeah, he ran into that problem with "The Cask of Amontillado." He didn't remember the exact problem, "something to do with where the wine was stored in the catacombs."] Did you try to answer your own questions? Yes Did you ask how the play made you feel? No Why not? 120

PAGE 134

Maybe subconsciously, but I never consciously thought about it. I knew how it made [He said this was also true of the stories but] I've changed my mind in class a couple of times. [He didn't remember which stories this occurred with.] Did you try to predict what would happen in the play? No Why not? I was thinking about skipping ahead to see what would happen, but I felt she would kill herself It never crossed Jessie's mind that it wasn't going to happen. With stories, I always run different scenarios of what might happen through my mind [as I rf!ad]. Did you try to identify the elements in the play schema? (setting, protagonist, antagonist, conflict, etc.) No Why not? It just seems pretty clear, so I never really stop to think about it. In stories I never before realized there could be different points of view, so I do think about that. Did you ask how the play would change if any of the elements in the play were changed or how important a given element is to the meaning of this play? No Why not? I try to understand why it's the way it is. I never understood why you had us do this [in class]. [I gave some more examples of how changing elements could change the meaning and also give insight into the current version. For example, I asked what would be the difference if 'night, Mother were set in New York City or the deep South-would 121

PAGE 135

it tend to iq.voke stereotypes? And I asked why a writer might choose to write a play rather'than a short story to convey the "story" and what would be gained or lost.] Did you as]c what meaning you got from the play? No Why not? [Eric said he did phrase the theme in his head.] I guess I should come up with a theme in my journal. I asked whether or not I agreed with the actions [of the characters]. Did you write all your questions? Yes Did you change the way you wrote journals from one story to another? No [I used the same process:] Reading and underlining, and writing journal. Did you change the way you wrote this journal compared to your journal on any of the stories? No Is there anything else you would like me to know? It [journal writing] seems to be working the way it's set up now. Summazy Eric was a classic example of a very bright student who does not voluntarily engage in study strategies, yet who appreciates their benefits when he is forced to do them. Although he didn't initially like doing the journal writing, he found it helpful from the start, and said he had even "learned to enjoy it," just as Colleen and Frances did. 122

PAGE 136

Summary In spite of their individual differences, from levels of field dependence/ independence levels of self-confidence and perceptiveness, each of these students reported journal writing as helpful. However, their experiences and corrunents, like those of the other students, suggest ways to refme the implementation of journal writing to make it even more helpful in the future. Lisa's, Colleen's, Trish' s, and Frances's responses to the Catechism suggest that it is a flexible learning tool, capable of providing scaffolding to a wide range of students. It was more consciously a part of both Lisa's and Trish's journal writing, although Lisa used it as a recipe, usually engaging in low-level cognitive processing. Trish typically engaged in high-level cognitive processing, writing the best journals in the class. On the other hand, Colleen used the Catechism loosely as a guide, and Frances used it just to get her "wheels turning"; field independent, she apparently preferred to impose her own structure on thinking about the stories, but lacked much of the self-confidence ofTrish and Eric. By contrast! Eric (who scored a perfect 16 on the field independence test) didn't even use the Catechism to get his wheels turning. He apparently just dove into writing his joUI'lj.als, confidently imposing his own structure. The performances of all five students call to mind Putnam's (1991) differentiation of competent and novice performances: Competent performance requires following up any particular move by noticing its impact, what Schon (1983) would call the 'backtalk' of the situatiQn, and shifting to any of several possible further moves. Someone who is learning will tend to 'fixate' on the move that is at his or her learning edge. One way of understanding this is to realize that the move at the learning edge requires so much conscious attention, so much of the limited 123

PAGE 137

available cognitive capacity, that even well-learned moves may not be With further practice, a move that was once at the learning edge becomes more skillful, freeing up attention for noticing other aspects of the situation. (pp. 154-55) Lisa remaired primarily at a novice level in both reading literatur.e and writing I journals. She qften had difficulty integrating widely separated passages, and sometimes had.:difficulty integrating consecutive sentences. She fixated on individual segments of a at a time and on low-level questions in the Catechism, rarely understanding the terminology when she tried to apply it. She applied a rigid procedure in trY,ing to understand easy as well as difficult stories. On the other hand, Eric and Trish seemed to be the most competent among these five as well as ih the class as a whole, in both reading and journal writing. Whatever Eric wrote was astute and he seemingly paid little conscious attention to his "moves." As he noted in qis interview, as he read, he "always [ran] different scenarios of what might happen [his] mind." This behavior seemed automatized. Whatever recipe he used was so well composed, he did not seem to be consciously aware of it. While she was also very competent, Trish seemed to be more conscious than Eric of her "moves." Her journals provide evidence of more direct backtalk (e.g., considering the impact of changes in the a story). Frances and Colleen fall somewhere on the scale between Lisa on the one hand and Eric and Trish on the other. 124

PAGE 138

CHAPTERS ANALYSIS OF RESEARCH QUESTIONS The pwpose of this study was to develop a cognitive model of journal writing in a college introduction to literature class. In chapter 4 I presented the case studies of five students in Group 1. In this chapter I systematically address each of the main research questions of the study. I focus on the students in the case studies, but also incorporate information from surveys, journals, and class discussion about other students in Group 1. I include results of surveys in Group 2 mainly to provide insight into implementation problems which affected students'. journal-writing.47 Research Question 1 What Metacognitive and Activities Do Students Engage in W1len They Write Journals? The surveys, case studies, and contents of journals indicate that students engaged in diverse metacognitive and cognitive behaviors; these varied both within and across students. Throu:ghout the term, all of the students seemed to try to understand what they read, although they differed greatly in their preparation and execution strategies, in the types of things they tried to understand, and in the amount of effort they seemed willing to invest Furthermore, the difficulty level of the story influenced the strategies of at least some students. 4 7 Fifty percent of the students in Group 2 missed one to three days of the first week of the course. During that time, instructional strategies included trying to establish an environment in which students felt comfortable to express their honest thoughts arid feelings, defining journal writing, and providing in-class practice and feedback on journal writing. Surveys and class comments later in the term indicated that many of the students had not understood the pmpose of journal writing. 125

PAGE 139

Metacognitiv(and Cognitive Strategies Students employed a variety of metacognitive and cognitive strategies in reading stories and writing their journals. and execution strategies. The survey responses and interviews suggest that 12'out of 13 students (all but Eric) generally planned from the outset to read each work. more than once, but the number of times depended on the difficulty of the work. All of the interviewees generally used at least two preparation and execution strategies (see Table 5.1).48 Table 5.1. Preparation and Execution Strategies Interviewees Generally Used Strategy_ Lisa Colleen Trish Frances Read First J ustto Get the Flow X Reread X X X X Circled Text Elements Diagrammed Events Took Notes X X X Identified Key Words X X Used the Catechism X x* X x* Looked Up Unknown Words** X X X Tried to Answer Own Questions X X X X Consulted Othet:s X x*** Colleen usM the Catechism only to stimulate her initial thinking. Frances generally used it only to "get her wheels turning." ** Eric always. inferred from context; Frances generally did. *** Trish consulted others only if she found a work very difficult. Eric X X X X For example; all five used a strategy such as highlighting or note-taking to focus their attention on various elements of the story, but the difficulty of the work influenced some. students' note-taking strategy. For example, on easier works, Trish took notes during the first reading, but on the most difficult works she waited until 48The interviewees may have used other strategies which they did not identify. 126

PAGE 140

later readings. Although Eric read each work only once, he circled sentences or parts of paragraphs as he read. Timing of writing. Before they began to write their journals, most students apparently preferred to get an idea of what a story was about,49 "a feeling for the flow of a story."50 In the End-of-Journal Survey,51 54% (7 of 13) said they generally wrote their journals after they had read a story at least once; another 31% (4) waited until they had totally finished reading; 8% (1) wrote during the initial reading as well as during additional readings; 8% (1) wrote journals only while reading. Interviews confirmed the delayed-journal-writing approach for the five students in the case studies. Using resources. Most students said they used the dictionary but not critical resources (e.g.,. reference books or published literary interpretations). Sixty-two percent (8 out of 13) said that when they encountered an unfamiliar word, they usually looked up the word immediately; 23% (3) said they looked up words later; one student (8%) looked up words later if she couldn't figure out the meaning from context; only one student (8%) said, in general, he never looked up words. 52 Again, interviews commned the survey responses of the five students. One interview also 49see Colleen's interview (in chapter 4). 50Jrus is Eric's phrase, but Frances used almost the same phrase in her interview response to that question. 5 1 Thirteen of the 15 students completed the survey. Rhona and Todd were not present for the survey. 52This was Eric, whose reading comprehension level was abOve grade 16.9. His astute observations in journals and during class discussions suggest that he was generally able to infer correct meanings from context But in his interview, he said that he did encounter a problem in ''The Cask of Amontillado" because he incorrectly inferred the meaning of a word. 127

PAGE 141

shed light on qualitative differences in the use of the dictionary. While most students only looked up unfamiliar words, Trish said she tried to identify the significance of key words in a story. When she read "Hills Like White Elephants," she already knew the meaning of absinthe (and had even tasted it). But she sensed that it played a larger role in the meaning of the story, so she looked it up in the dictionary to see what additional insight she could gain from the definition. 53 However, she did not always check the dictionary for additional meanings of words that she tagged with questions in her journal (see later discussion in this chapter about her questions about "white elephant," in journal 3). Fifty-four percent (7 of 13) said they never consulted reference works or critical articles; another 31% (4) did so rarely or only "Occasionally."54 The survey did not ask if students consulted family, friends, or other students; however, the Background Survey, journal comments and oral comments indicated that at least seven students55 occasionally turned to others when they had difficulty understanding works. Because 53 Based on the infonnation she obtained from the dictionary, she concluded that Hemingway referred to absinthe to foreshadow the bitter outcome of a long-awaited event. Lisa also tried to identify key words, but her reading comprehension skills were so poor (grade level of 9.4), she was rarely able to identify all of the key words or to draw appropriate inferences when she did. 54The consulting appeared to have been related to two works. A couple students indicated that they had used resources to try to locate "Samarra" in the first story on which they wrote a journal. The second work, Oedipus the King, was not directly included in this study; however, the End-of-Journal Writing Survey was conducted after students had completed all their journal assignments. 550n the Background Survey on the first night of class, Bonnie, Dinah, Todd, Trish, and Wanda all identified this as one of the strategies they used to make sure they understood reading assignments. In at least two journals, Todd and Wanda also commented that they had discussed the stories with others. Colleen and Lisa did not identify discussion as a strategy in the survey or in their journals; however, during casual conversation they told me that they had asked their husbands to discuss stories, but that their husbands refused after the frrst or second story. Frances told me that Lisa had asked her for help while she was writing her journal on the fifth story. 128

PAGE 142

such discussions can lead to story misconceptions, 56 future studies should investigate students' use of such resources. Students' use of the Catechism varied greatly. Some students used the Catechism more as the course progressed; some used it less, as they increasingly used the journal to respond to their unique needs; and some used it only when they needed help. getting "the wheels turning." The five interviews and contents of journals confirmed this diversity. (See detailed discussion later.) Usin&journals as tools. In their comments on the "End-of-Journal-Writing Survey," 69% (9 of 13) of the students described journal writing in terms of a (cognitive) tool, helping them understand a story, forcing them to think, or stimulating them to ask (and to try to answer) their own questions. This view of journals was confirmed in the five interviews. However, survey comments and journal contents revealed that two students often did not use their journals as a tool in discovering meaning; nor did they try to integrate the various elements of the story on their own. They turned, instead, to discussing the story with others (i.e., other people became tools for discovering meaning). 57 A couple students viewed journals 5 6 For example, during a discussion just before class, Todd apparently accepted the conclusion of a couple other students that the girl in "Hills Like White Elephants" was deaf and that the operation was to restore her hearing. He then wrote that interpretation in his journal. Similarly, when Lisa asked Frances for help in understanding the reference to masons in "The Cask of Amontillado," she misunderstood Frances's explanation and imposed an incorrect schema on the story which made it even more difficult for her to assimilate other details. 5 7 For example, in journal 3, Todd and Wanda both mentioned discussing "Hills Like White Elephants" with family or friends. But because they did not share the substance of such discussions, it is impossible to know what metacognitive strategies or cognitive activities, if any, they engaged in, and whether they understood the basis of other people's interpretations of the stories. Often, however, discussions did not seem to provide any additional help; for example, injournal3, Wanda noted that both she and her family were anxious to find out "what the story was about" Furthermore, students who relied on out-of-class discussions in writing their journals did not seem to understand that it is possible to develop' skills in understanding difficult stories. In journal 5, Todd said that although he 129

PAGE 143

primarily as a place to record their opinions of stories. (In Group 2, 20% [2 of 10] students viewed journals primarily as a tool to help them prepare for quizzes.) Selectin& content for journals. Survey comments, interviews, and students' comments during class discussions indicate that students did not always record all of their questions or comments. 58 Some students apparently invoked metacognitive strategies to what or how much to write in their journals. For example, two of the students whose journals were among the longest nevertheless said they constrained their journals because they did not want to burden me, even though I encouraged all students to write as much as they wanted. (But what strategy guided their selection of elements to question or explain? Although there is some evidence that the two students began by addressing the elements most troubling and/or interesting and stopped writing when they began to feel they were exceeding some preconceived limit, this issue must be explored further in future studies.) Some students at occasionally discarded journals they were not satisfied with and began anew; for example, Frances got bored in applying the Catechism in journal 6 and started over. Some did not add questions that arose after they had written their journals but before they submitted them. Students' comments during class discussions also made it apparent that at .least some students had not included all of their self-explanations or questions on major elements of the story.59 found the story "neat," "when I have questions even after re-reading a work, I get flustered .... I guess simplistic stories are more my style." 581 can only infer the possible reasons for their not writing all their questions; for example, lack of time, laziness, inadequate motivation, etc. 59por example, having applied the catechism in journalS, Frances said she got bored using it in journal6, so she threw that journal out and started over. Lisa revised her journals on at least one 130

PAGE 144

Cognitive Activities The goal of maintaining the natural environment in the study made it impossible to directly monitor students' journal-writing behavior through think-aloud protocols or other intrusive means. 60 Thus, I was forced to infer students' cognitive activities primarily from the content of their journals. Surveys and interviews provided evidence that students did engage in certain types of cognitive activities before they began to write .their journals. For example, all but one of the students said they read stories at least once before beginning to write their journals, and many took notes or highlighted elements in the stories. Most said they looked up unknown words; some inferred meanings from context. Some said they consciously tried to predict what would happen as they read; one said he always ran several scenarios through his head as he read. Some discussed the stories with others. Thus it is clear that journal writing was the culmination of a series of cognitive activities. Nevertheless, identifying what we might term "cognitive artifacts" in the journals provides valuable information about students' attempts to understand stories. I coin this metaphor from archaeology, in which researchers draw inferences about a past culture, based on limited material evidence. This study did not directly address the question of whether students would have engaged in the same cognitive activities if the course had not included journal writing. My experience suggests they would not. Eric's response seemed fairly occasion. During class discussions, Patrick's comments were almost always more insightful than his journals. 60some students said they spent up to 10 hours on a story-journal assignment. Several said they spent time on a journal both at home and at work. 131

PAGE 145

typical. Although he said he found journal writing helpful and recommended keeping journal writing as a requirement in the class, he said that, unless he had been required to write the journals, he would not really have thought about each story until specific issues were during class discussion. He said he was basically "lazy." Several others noted that journal writing stimulated or facilitated their thinking, or even, forced them to think about what they were reading on a more conscious level. (See Research Question 6.) The literature on graded journal writing generally supports these students'. perceptions (e.g., Bauso, 1988; Robinson-Armstrong, 1991; Roth, 1985; Wilson, 1989). However, the literature does not include any studies comparing the use of journals to other graded learning strategies which actively involve the learner. Analysis of journals 3 and 4 revealed that students' generally engaged in three types of cognitive activities: reviewing content (summarizing and paraphrasing), evaluating (making judgments about character actions, author actions, or the story),. and problem solving (asking questions and generating self-explanations). The elaborated classification is reminiscent of the cognitive strategies Brown and Palincsar (1989) say "good students routinely bring to the task of studying texts": questioning, clarifying, summarizing, and predicting (p. 414). The classification is also suggestive of Chan, Burtis, Scardamalia, and Bereiter's (1992) three "response subtypes" in their levels of constructive activity in learning from text: declarative, interrogatory, and evaluative (p. 103). However, neither of those classification schemes is directly applicable in this study. They relate to what Ng and Bereiter (1991) have called "extracting knowledge from text," while the current study focuses 132

PAGE 146

on constructing meaning from short stories, which requires more interpretation than extraction. Although the taxonomy I developed includes several other categories (see later discussion), mr attempts to classify the level of cognitive processing (e.g., shallow versus deep) and level of outcomes (e.g., concept learning or problem solving) were counterproductive for several reasons. First, as noted above, journal questions and comments are cognitive artifacts, and thus do not provide adequate insight into whether a student was, for example, applying a procedure or problem solving. Thus, if a student wrote: "What difference would it make if the setting were changed?," it was impossible to determine if the student was rotely writing a question in the "Catechism for Stories," seriously asking the question while lacking the confidence to venture an answer, or sincerely trying to problem solve while lacking confidence to write his or her tentative answer. Comparing the best reader (probably Eric) with the poorest (Lisa, who was probably the mpst highly motivated in the class) clarifies the problem. Eric's journal and class comments always reflected a very good grasp of each story, yet he never read a story more than once. Moreover, his interview suggested that his reading skills were highly automatized. For example, he said that he didn't really predict outcomes, but instead automatically ran several possible scenarios through his mind as he read. (Apparently he automatically rejected scenarios as new details made them untenable.) On the other hand, Lisa read each story up to ten times to try to comprehend each story, laboriously applying a procedure which was automatized only at the macro level. As she said in her interview, 133

PAGE 147

I read through once without writing, let it sit, come back, try to get the keywords, .... Then I go through and pick it apart. Then I go through and really pick it apart. But at the micr.o level of sentence and paragraph comprehension, many of her activities were not automatized. Generating low-level inferences seemed to constitute problem-solving rather than application of rules. For example, in "The Cask of Amontillado" the narrator tells us that Fortunato's weakness was his pride in his knowledge of wines. The narrator appeals to that pride, telling Fortunato that "some fools would have it that his taste for wines is a match for your own," and that he (the narrator) will ask Luchesi to authenticate a cask of Amontillado since the narrator can see that Fortunato is busy. Yet Lisa struggled with this in her journal: "Luchesi must be a rival of some sort to Fortunato. Because/ think Luchesi is a wine conissouor [sic] also" (emphasis added).61 The second reason that attempts to classify levels of processing and levels of outcomes were counterproductive is closely related to the first: semantic ambiguity. In the process of developing the taxonomy, I applied given versions several times, sometimes several days apart, to some of the journals. Although my classification of items was generally consistent, I was distressed to find that I was not always consistent in classifying some of the comments. Semantic difficulties became even more apparent when I asked a part-time literature teacher who requires her students to write structured journals to apply the fmal taxonomy to two journals. 61When I tutored her in my office, I saw her struggle several times with pronominal reference (not always successful in the outcome), again problem-solving rather than applying rules of grammar. 134

PAGE 148

The following item from Bonnie's journal on "Hills Like White Elephants" illustrates the problem: "I think that the man wants her to have the operation but doesn't want to feel that he's responsible for the decision." The classification problem relates to the words "I think." I considered them a matter of style, a way of softening an assertion (with the student's sentence being equivalent to "The man wants her to have the operation but doesn't want to feel that he's responsible for the decision"). However, the other teacher thought they were tagging an indirect question "I wonder if the man wants ... "). There is, of course, no way to reliably determine which interpretation, if either, is correct. Even if one were to ask the student to clarify her intent, the very act of asking would potentially affect the student's response. 62 Reviewing content. All but one student (Eric) reviewed some of the content at least once in journals 3 or 4. Although 79% (11 of 14) reviewed content at least once in both journals, it was not a major activity of most students. Nevertheless, reviewing content apparently helped some students monitor their comprehension. Although the mechanism is npt clear, reviewing content also apparently scaffolded some students in generating self-explanations. 63 Evaluating. The second category of cognitive artifacts was students' evaluations of character actions, author actions, and stories. Evaluative behavior could be predicted to a limited extent from one journal to the next. Seventy-one percent (10 of 14) of the students engaged in some type of evaluation. Seven students (50%) judged 62This problem is reminiscent of Heisenburg' s principle of uncertainty. 63see later discussion of Ann's discovering meaning through summarizing. 135

PAGE 149

characters' actions in one or both journals. Five (36%) evaluated the story or the author's actions in one or both journals. As Table 5.2 indicates, the mean numbers of comments were relatively low, but there was a small positive correlation between the numbers of evaluative comments students made on characters in journals 3 and 4. Furthermore, there was a moderate positive correlation between the total numbers of evaluative comments in the two journals. Table 5.2. Statistical Summary of Evaluative Comments in Journals 3 and 4* Element Evaluated M!iliiD # Qf Commjilnts t-test r Probability Journal3 Journal4 ofr Characters' .36 .86 ns .60 .01 Author/Story .50 .43 ns .48 ns Total .86 1.29 ns .74 .01 *n= 14 Problem solving. The frequency and specific type of students' problem-solving activities-the third cognitive artifact-varied across journals, but there were no consistent differences associated with story difficulty level. Regardless of the difficulty of the story, on average students used their journals more to explain than they did to ask questions, except in their first journal (see Table 5.3). T bl 5 3 M N be f C a e .. ean urn rs o ogruti.ve A .... J ctivtti.es m 1 ournas Problem Solving Self -explanations Questions Content Reviews 1 3.58 3.58 1.08 2 10.75 1.83 .25 3 10.58 2.83 3.17 4 12.92 2.33 6.83 5 14.25 3.25 8.58 n = 12 (Only students who completed ailS JOurnals were mcluded.) 136

PAGE 150

There were significantly more self-explanations than questions in journals 2 through 5 (p <.05) and significantly more self-explanations than content-review items in all five journals (p < .05). Since most students described journal writing as a cognitive tool to help them understand stories (see discussion later), these patterns are not surprising. The small number of self-explanations compared to questions in journal! is probably related to two factors. First, it was a very difficult story.64 Second, several students used the Catechism as a direct prompt, sometimes drawing on questions without offering answers. (For example, some asked if a given character were the protagonist, rather than asserting that they thought the character was the protagonist.) As the course progressed, students' journals did not reflect the Catechism as (see discussions elsewhere). Statistical analysis revealed a strong positive correlation (r = .944, p = .001, n = 14) between the numbers of self-explanations students generated in journals 3 and 4. I, therefore, calculated correlations of self-explanations between each pairing of all five journals and found small to moderately strong positive correlations (Table 5.4). Table 5.4. Correlations of Numbers of Self-Explanations Between Journals* Journal 2 3 4 1 .868*** .806*** .671** 2 .742** .576** 3 .944*** 4 n = 12 (Only students who completed all 5 JOurnals were mcluded.) ** p< .OS *** p< .01 5 .638** .649** .666** .693** 64 As I explained in chapter 3, I purposely selected a story that I knew would stimulate questions. 137

PAGE 151

The correlations suggest that certain students habitually engaged in generating self, explanations-:regardless of the level of difficulty-more than other students did. The drop in correlations with journal 3 from .944 for journal4 to .666 to journal 5 is difficult to explain. The high correlation between 3 and 4 may be an aberration. For example, students who were able to generate self-explanations in the most difficult story continued to do so, while those who were unable to make many selfexplanations that story may have been temporarily demotivated. The fact that the correlations were generally not stronger suggests that some students did vary their behavior from to story. Several students wrote increasingly longer journals, with relatively more self-explanations; 65 consequently, the ordinal position of other students dropped while the frequency of their self-explanations may have remained relatively constant. For example, Eric generated two self-explanations in journals 1 and 2 and one self-explanation in journal 5. Todd generated none in journal1 and one in journal 5; the much higher frequency in journal 2 (nine self-explanations) was apparently stimulated by the adjunct questions (see related discussion later). While generation of self-explanations could be predicted to some extent from one journal to another, questioning and reviewing content generally could not be. The only significant correlation of question frequency was between journal 4---on the easiest story-.andjoumal5----on a relatively difficult story (r = .613, p < .05). This relationship might be explained by the fact that I gave students a copy of Trish' s joumal4 (see Appendix K), since some students still did not seem to understand how 65From joumall to journalS, Frances moved from 9.5 (tied with Grace) to 2 in relative frequency of self-explanations; Grace moved from 9.5 to 6; Colleen and Ann from 4.5 to 3.5. 138

PAGE 152

to use journals for their own benefit. The frequency and high level of questions in her journal may have served as a model for other students, as I intended. The only significant correlations of content review were between journals 1 and 5 (r = .623, p < .05), and 4 and 5 (r = .596, p < .05). Since these journals were on the easiest story (4) and two relatively clifficult ones (1 and 5), the correlations suggest that factors other than story difficulty level and habit tend to mediate the use of these two cognitive activities. Journal content clearly indicates that students differed in the purposes for which they reviewed content. Ann seemed to use content review in difficult stories to scaffold her discovery of important insights. For example, in journal 3 after apologizing for summarizing much of "Hills Like White Elephants," she wrote: The point of view seems to me to be objective. Now I am very amateurishly speculating that the author chose this point of view to highlight the division between these two people who seems to find communicating about feelings clifficul t. And again, in journal 5, on "The Cask of Amontillado," summary led to insight: Montresor, as he begins to guide Fortunato to his palazzo's catacombs, and Fortunato has had a fit of coughing, states that Fortunato should go back; he would be missed. He was rich, respected, admired, etc., his health was precious. With this passage it has suddenly dawned on me that Fortunato is a blessed scion and Montresor symbolizes his encroaching ruin! In contrast, several students seemed to use content review as a transitional device unless they found the story particularly clifficult
PAGE 153

research (R. C. Anderson, 1977; Rumelhart & Ortony, 1977). According to this theory, when existing schemata readily accommodated story elements, students would not experience any cognitive dissonance66 and therefore would not usually mention those elements in their journals. However, the interviews provided evidence that when their existing schemata did not readily accommodate story elements, most students summarized, 67 generated self-explanations, or asked questions. Furthermore, when students experienced too much cognitive dissonance (as some appeared to do in trying to understand "Hills Like White Elephants"), they resorted to general strategies such as stating their general confusion or blaming the author. Mayer's (1975, 1979, 1980) assimilation theory provides additional insight into the students' use of schemata in problem solving. The theory suggests that meaningful learning (or, in this case, understanding) depends on three conditions being met: [T]he material to be learned must be received (reception), the learner must have a; meaningful set of past experiences that can be used as an assimilative context (availability), and the learner must actively use an assimilative context during learning (activation). (p. 772) However, students did not always receive the material (see later discussion: "Omission of Significant Elements"). Nor did students always have an assimilative context. 6 8 This was particularly evident in their journals on the most difficult stories; 6 6 For a discussion of cognitive dissonance, see Wilson and Cole (1991). 6 7 Although summarizing seems to be related logically to an absence of cognitive conflict, this was not necessarily true in students' journals on literature. As noted earlier, summarizing seemed to scaffold some students in discovering meaning. 68For example, most students do not have a context for assimilating the operation, which is described as "letting the air in." This is an old technique for an abortion, which even older doctors cannot necessarily identify. Two years ago, when she was writing her journal, one of my students asked her uncle, a doctor, if he could identify the purpose of the operation, but he was unable to. 140

PAGE 154

some students mainly asked questions or expressed confusion instead of generating self-explanations. Journal questions and comments did, however, suggest that students were generally trying to use an assimilative context during reading and writing. When readers must activate schemata, they run the risk of imposing inappropriate schemata (Bransford & Johnson, 1972). This was particularly evident in the third journal (on the most difficult story, "Hills Like White Elephants"). Apparently unable to make sense of some key details, several students (e.g., Ann, Patrick, and Todd) attended to details that fit a schema they could understand; that is, they concluded the girl was deaf because she could not understand the waitress and asked the man to tell her what the waitress said. 69 (See related discussion later: "Omission of Significant Details.") In a slightly different vein, all the students apparently activated an incomplete schema for "white elephants." None of them had additional schemata for a white elephant as something precious which is difficult and expensive to maintain or an endeavor that is a failure. Research Question 2 How Are These Activities Related to Individual Learner Characteristics? Calculation of Pearson product-moment correlation coefficients revealed few significant correlations between cognitive strategies/activities and individual learner characteristics measured in the study (Table 5.5). As would be predicted, students 69 Actually, the girl did not understand the Spanish-speaking waitress and the man did, so she had to rely on him to translate. The students miss the fact that the girl has no difficulty understanding the man. And even though they note that she gets so distressed by the man's talking that she asks him to stop talking, they fail to note the inconsistency in concluding that she is deaf. 141

PAGE 155

who wrote more self-explanations in one or both journals tended to have higher journal averagt(s. Furthermore, students' initial ratings of the helpfulness of the Catechism were a small predictor of students' use of two cognitive activities (self explanations and reviewing content) regardless of story difficulty. Table 5.5. Significant Correlations Between Individual Leamer Characteristics and Cognitive Activities in Journals 3 and 4 1. number of items supporting self-explanations, journa14 Average Grade. Journals on Stories 2. number of items with details supporting questions, journal4 3. number of self-explanations, journal 3 4. number of self-explanations, journal4 5. number of questions, journal4 6. number of direct closed questions, joumal4 Confidence in Ability to Write Journals. First Day 7. number of items reviewing content, journal4 Improvement in Conficlence in Ability to Write Journals