Evaluating the impact of Colorado's higher education admission requirements

Material Information

Evaluating the impact of Colorado's higher education admission requirements the unintended consequences of increased standards on marginalized populations
Schaible-Brandon, Sonia
Publication Date:
Physical Description:
xii, 248 leaves : ; 28 cm

Thesis/Dissertation Information

Doctorate ( Doctor of Philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
School of Education and Human Development, CU Denver
Degree Disciplines:
Educational Leadership and Innovation
Committee Chair:
Goodwin, Laura
Committee Co-Chair:
Sands, Deanna
Committee Members:
Leech, Nancy
Becker, Madeline


Subjects / Keywords:
Universities and colleges -- Entrance requirements -- Colorado ( lcsh )
State universities and colleges -- Entrance requirements -- Colorado ( lcsh )
Universities and colleges -- Admission -- Colorado ( lcsh )
State universities and colleges -- Admission -- Colorado ( lcsh )
Minorities -- Education (Higher) -- Colorado ( lcsh )
Minorities -- Education (Higher) ( fast )
State universities and colleges -- Admission ( fast )
State universities and colleges -- Entrance requirements ( fast )
Universities and colleges -- Admission ( fast )
Universities and colleges -- Entrance requirements ( fast )
Colorado ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 225-248).
General Note:
School of Education and Human Development
Statement of Responsibility:
by Sonia Schaible-Brandon.

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:
268773096 ( OCLC )
LD1193.E3 2008d S33 ( lcc )

Full Text
Sonia Schaible-Brandon
B.A. University of Colorado at Colorado Springs 1992
M.A. University of Colorado at Colorado Springs 1997
A thesis submitted to the
University of Colorado Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Educational Leadership and Innovation

This thesis for the Doctor of Philosophy
degree by
Sonia Schaible-Brandon
has been approved
Dr. Laura Goodwin
Dr. Deanna Sands
Dr. Nancy Leech
Dr. Madeline Becker

Schaible-Brandon, Sonia (Ph.D., Educational Leadership and Innovation)
Evaluating the Impact of Colorados Higher Education Admission Requirements: The
Unintended Consequences of Increased Standards on Marginalized Populations.
Thesis directed by Professor Laura Goodwin
This research is a sequential quantitative to qualitative mixed-method study
which examined the Colorado Higher Education Admission Requirements. The
model and its components were studied to investigate potential adverse affects on
marginalized populations. Phase I replicated the path model of the current policy. The
population was incoming baccalaureate students at one less selective 4-year
institution (N=783) who graduated from high school in the spring of 2005. Phase II
employed a phenomenological study of 3 African American men predicted by the
model to fail, yet remained in good standing after 2 years. Results showed access to
opportunities need to be addressed before more standards are enacted.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
Laura Goodwin

I dedicate this thesis to my daughter, Sydni, who gave me the resolve to finish this
thesis with her own drive and determination. She is a great example for all who
pursue a dream. To my loyal son, Jordan, for all the love and support he has shown
me, I also dedicate this to him.

I would like to thank my advisor, Dr. Laura Goodwin for her continued support of my
research. I also wish to thank all the members of my committee for their valuable
participation and insights.

1. Introduction......................................................1
1.1 Purpose of the Research..........................................1
1.2 Problem Statement................................................3
1.3 Conceptual Framework.............................................8
1.4 Research Questions..............................................17
1.5 Overview of Methodology.........................................17
1.5.1 Phase I of the Research Quantitative Analysis..............18
1.5.2 Phase II of the Research Qualitative Analysis...............20
1.6 Structure of the Study..........................................21
2. Literature Review................................................22
2.1 Overview and Limitations of Colorado Admissions Policy..........25
2.2 The Social Reproduction of African-American Males in the
Education System Related Studies...............................33
2.3 Overview of High School Curriculum..............................37
2.4 History of the High School Curriculum...........................39
2.4.1 Progressivism.................................................42
2.4.2 Combining Rigorous Curriculum and Comprehensive Ideals........45

2.5 Present-Day K-16 Alignment.......................................46
3. Methodology.......................................................55
3.1 Design ..........................................................56
3.1.1 Phase I Quantitative Analysis..............................57 Subjects and Sampling Phase 1.............................59 Setting and Materials Phase I.............................59 Independent Variables Phase 1.............................60 Dependent Variables Phase 1...............................61 Data Collection Procedures Phase 1........................62 Data Analysis Procedures Phase I..........................62
3.1.2 Phase II Qualitative Analysis...............................66 Subjects and Sampling Phase II.............................66 Setting and Materials Phase II............................68 Selection of Interviewees Phase II........................68 Data Collection Procedures Phase II.......................69 Data Analysis Procedures Phase II.........................73 Constant-Comparison Analysis and Emergent Themes..........75 Content Analysis...........................................78
3.2 Summary of Methodology.....................................78
4. Results Phase 1...................................................80

4.1 Results
4.1.1 Significance of the Model......................................84 Equation One.................................................85 Equation Two.................................................86 Equation Three...............................................87
4.2 Total Effect Size of Each Equation..............................88
4.3 Examination of Parameters.......................................90
4.4 Power Analysis..................................................92
4.5 Examination of Assumptions of Path Analysis.....................93
4.5.1 Proper Specification of the Model..............................94
4.5.2 Sample Size....................................................95
4.5.3 Linear Relationship............................................95
4.5.4 Independence and Normal Distribution..........................100
4.5.5 Errors of Prediction..........................................102
4.5.6 Multicollinearity.............................................104
4.6 Discussion....................................................106
5. Results Phase II................................................109
5.1 Students........................................................109
5.1.1 Kyle .........................................................110
5.1.2 Rahim

5.1.3 Jamal .......................................................113
5.2 Barriers in Education..........................................115
5.2.1 Lack of High School Guidance.................................115
5.2.2 Mathematics Deficiencies.....................................120
5.2.3 Family Situation.............................................123
5.3 Success in Higher Education..................................124
5.3.1 College Connections..........................................125
5.3.2 Student Agency and Resilience................................127
5.4 Content Analysis...............................................129
5.4.1 Coding Scheme................................................129
5.4.2 Content Analysis Results.....................................129 Barriers to Education (Content Analysis)...................132 Success in Higher Education (Content Analysis).............133
5.5 Student Opinion of HEAR........................................134
5.5 Discussion.....................................................136
6. Conclusion.....................................................141
6.1 Question One To What Extent Do the Components in the CCHE
Admission Policy (HEAR) predict first-semester college GPA for
students at one less-selective Liberal Arts Institution.........142
6.2 Question Two Which Components Are the Best Predictors........142
6.3 Question Three Why Do Some Students Defy the Prediction
Model by Performing at a Much Higher Level than Predicted.......143

6.4 Study Limitations
6.5 Research and the Conceptual Framework......................148
6.6 Implications for Policy....................................152
6.7 Suggestions for Future Research............................154
6.8 Summary....................................................156
Appendix A ....................................................158
Appendix B ....................................................213
Appendix C ....................................................224
References ....................................................225

1.1 Conceptual Framework...................................................16
1.2 Path Diagram of Colorado Admission Policy..............................19
3.1 Path Diagram of Colorado Admission Policy..............................58
4.1 Path Diagram of Colorado Admission Policy.............................81
4.2 Normal Probability Plot for Equation One.............................101
4.3 Normal Probability Plot for Equation Two.............................101
4.4 Normal Probability Plot for Equation Three...........................102
4.5 Scatterplot of Residuals for Equation One............................103
4.6 Scatterplot of Residuals for Equation Two............................103
4.7 Scatterplot of Residuals for Equation Three..........................104
6.1 Original Conceptual Framework.........................................149
6.2 Revised Conceptual Framework.........................................151

4.1 Descriptives of Model Variables........................................82
4.2 Indices of Fit .......................................................83
4.3 Initial Regression Analyses...........................................85
4.4 Regression Weights for Equation One...................................86
4.5 Regression Weights for Equation Two...................................87
4.6 Regression Weights for Equation Three.................................88
4.7 Standardized Total Effect Size........................................91
4.8 ANOVA test for Linearity..............................................96
4.9 Collinearity Diagnostics.............................................105
5.1 Content Analysis Counts...............................................131

1. Introduction
1.1 Purpose of the Research
The purpose of this research was to examine the predictive validity of the
new Higher Education Admission Requirements (HEAR), established by the
Colorado Commission on Higher Education (CCHE) in October 2003, and
develop a comprehensive understanding of the effects of these standards on
incoming populations of students at one liberal-arts institution in the state.
Specifically, CCHE implemented minimum curriculum thresholds in each of four
academic areas including mathematics, social science, natural science, and
English for admission into any of Colorados twelve 4-year institutions of higher
education. This requirement went into effect beginning with the high-school
graduating class of 2008 (Colorado Commission on Higher Education [CCHE],
2005). The focus of this research explored the impact of the changes in order to
investigate negative consequences on students whose high school curricular
preparation may have been impeded.
Between academic years 2002 and 2005, the state realized a continuous
increase in the number of recent high-school graduates who were identified as
needing remedial education once admitted to college (Futhey & Brandon, 2003;
Gianneschi, 2006). This trend was seen across the country, not just in Colorado. A

report by the Alliance for Excellent Education (2006) indicated that nationally,
the crisis in American high schools resulted in economic losses of more than 3
billion dollars annually. In a Report to the Legislature in 2004, Futhey & Brandon
noted that approximately two-thirds of the students attending state institutions
were found to be lacking many of the traditional courses deemed necessary to
succeed in college. For example, in some cases, students were found to have taken
as little as one year of high-school mathematics. As a result, many of the
competencies considered necessary for success in college appeared to be absent.
At the March meeting of the CCHE in 2003, at which the author was present as
the analyst of these data, the Commissioners confronted members of the K-12
State Board of Education in order to address apparent insufficiencies found in
high-school graduation requirements. The State Board, disagreeing with the
CCHE assumptions as well as those of the author, refused to acknowledge the
increase in remedial rates was linked to deficient standards in the primary and
secondary system. As a result, the CCHE began a six-month effort to increase
college admission standards in order to attempt to curtail the number of students
entering 4-year colleges without sufficient preparation (Adleman, 1999; Adelman,
2006; American Diploma Project, 2002; Kirst & Venezia, 2004; Venezia, Kirst, &
Antonio, 2002).
In October of 2003, the CCHE initiated the implementation of the HEAR,
a mandatory high-school core curriculum requirement to be phased in with the

graduating class of 2008, which included (a) three years of mathematics, (b) three
years of social science, (c) three years of natural science, and (d) four years of
English. For 2010, CCHE mandated an additional phase that will include an
additional year in mathematics and one year of foreign language as a response to
ever-increasing remedial rates. The HEAR, along with other components of high-
school performance, was mandated to be used in determining student eligibility at
Colorado 4-year public institutions of higher education. Completion of the core
was restricted only to those students applying for admittance into one of
Colorados baccalaureate institutions. Community college entrance requirements
were not affected (Carnahan, Leal, & McKeever, 2007).
1.2 Problem Statement
The CCHEs HEAR was developed solely on a quantitative model, using
numbers of courses, standardized test scores, and high-school performance as
measured by high-school grade-point average or class rank to predict success in
college (CCHE, 2005). Although each baccalaureate institution was allowed to
have 10% to 20% of admitted freshmen below the standard, this percentage did
not appear to be sufficient when applied to current populations of students,
particularly at those schools that traditionally have accepted students with less
preparation (Lestina, 2006). As the HEAR requires a rigorous high-school
curriculum for first-time freshmen, initial findings showed these changes would
most affect those schools with less selective admission requirements (Lestina,

2006). Prior to the policy change, the University of Colorado already had a
rigorous curriculum requirement in place, the Minimum Academic Preparation
Standards (MAPS), which began with the high-school graduating class of 1988
(University of Colorado, 2006). As a result, HEAR would have less of an impact
on the populations entering The University of Colorado at Boulder, The
University of Colorado at Denver, and The University of Colorado at Colorado
Springs. However, for institutions that have historically served incoming students
with less rigorous preparation, the implications of this new policy needed to be
explored in order to understand which populations would be most adversely
Examining current populations, one less selective institution recently
reported that, if these standards had been applied to the entering class of 2005,
only 35% of the entire entering class would have been able to meet the new
standards (Lestina, 2006). One year later, the proportion of applicants who met
the standards had increased to 44% of the students for the same institution
(Carnahan et al., 2007). However, the CCHE analyses lacked thorough impact
studies. As the analyst for this policy, the author acknowledged that any impact
studies conducted by CCHE were merely superficial. CCHE did not have access
to identify the students potentially affected and, as a result, could not assess the
true impact of the standards on a deeper basis. Who would be most affected by
these changes? Could these students be given a voice? With approximately 55-

65% of the affected population at one less selective institution not meeting the
standards, this policy needed to be further examined, focusing on student voice.
Failure to evaluate thoroughly the effects of this policy could be disastrous, not
only to students who would not be allowed to enroll at 4-year public institutions
but to the institutions as well. Losing more than half of the student enrollment
could financially devastate an institution. In regards to the affected students, the
population who did not meet the HEAR were disproportionately minority students
(Carnahan et al., 2007). This policy may have the potential of severely reducing
the proportion of entering freshmen minority and other marginalized populations
such as first-generation students at 4-year institutions of higher education in direct
opposition to the states position of increasing diversity at these schools (CCHE,
1998). In addition, the documented K through 12 practices resulting in the over-
identification of African American males in exclusionary discipline and
placement into special education programs were of great concern in terms of the
potential impact of this policy (Gonzalez & Szecsy, 2004; Skiba, Michael, Nardo,
& Peterson, 2000; Skiba & Peterson, 1999; Skiba & Rausch, 2006).
Furthermore, opponents to HEAR became more vocal as the 2008
deadline approached. Challenges to the policy included proposed legislation
prohibiting CCHE from establishing high-school courses as a part of college or
university admission requirements (Colorado Senate Bill 06-026, 2006). Also,
rural schools and districts expressed concerns over the inability to attract teachers

in the math, science, and foreign language fields, making it difficult to comply
with the CCHEs requirements, particularly Phase II (Colorado Association of
School Boards, 2006; CCHE, 2006). Other school district leaders challenged the
overall composition of the core curriculum, stating that this policy may cause an
increase in high-school dropouts in the future, particularly by those students with
no intention of going to college (Colorado Association of School Executives,
2006). Although many stakeholders have voiced their opinions, those that would
be most affected had not been given an opportunity to express their views prior to
this study.
The CCHE policy grew out of a positivistic view that, ideally, it is
important to be able to generalize a single standard across the population of 4-
year public higher education students. Built on the states student-unit database,
the data used were a general reflection of student success (Futhey & Brandon,
2003). However, demographics were not included as part of the model upon
which the HEAR was based. The data used were not controlled for demographic
differences, such as race and ethnicity, gender, or socioeconomic status. Nearly
78% of recent high-school graduates enrolling in Colorado 4-year institutions of
higher education classified themselves as Caucasian (Futhey & Brandon, 2003).
Thus, the variables in the model are strongly representative of the majority society
(Delgado & Stefancic, 2001). Exacerbating these factors, Colorados minority
high-school dropout rates have exceeded those of Caucasian students by more

than twenty points (Manhattan Institute for Policy Research, 2003). As a result,
knowledge used in the creation of this model was too general to be applied to
specific populations of interest and was potentially not representative of diverse
populations, running contrary to the states proclaimed interest in diversity
(CCHE, 1998; Johnson & Onwuegbuzie, 2004).
In order to evaluate fully the CCHE admission policy, the model needed to
be examined in the context in which it was built and outliers, or students who
defied the model, identified. By studying these students, one could begin to
assess who would not be allowed to enroll in a 4-year public institution when the
HEAR was fully implemented. Additionally, factors other than high-school
preparation could be identified when measuring student success in college.
Weaknesses in the educational experiences of these students could assist in
leading to further reforms. Of most interest were those students who remained in
good standing at the college and continued into their second spring semester
despite the prediction by the admission standards that they would fail in their first
semester, due to their predicted first-semester college grade-point average (GPA).
Once identified, a deeper study of these students characteristics enriched the
understanding of the impact of the policy and offered explanation to this
phenomenon (Johnson & Onwuegbuzie, 2004).

1.3 Conceptual Framework
As a lens for this study, a socio-political framework served as the basis for
the conceptual framework of this study. By using this framework, an
understanding of systemic issues aided in evaluating the effectiveness and
efficacy of the HEAR and examined whether these requirements would result in
unintended inequalities. Nieto (1994) stated that, in order to begin to remove
unnecessary barriers to education, policies and practices must be challenged.
Because policies in education usually derive from the social viewpoint of the
dominant culture, oppression could easily become embedded in the institutional
structure (Greenman & Kimmel, 1995). Often, reforms implemented with good
intentions have served to reproduce or even worsen inequalities for already
oppressed groups (Apple, 2006).
Successful implementation of the HEAR required a full understanding of
political beliefs in Colorado and nationally at the time of development. The
impact of No Child Left Behind legislation cannot be understated. The formalized
implementation of standards and the competency-based selection processes may
have negated the importance of family background and social reproduction to
educational attainment (Chunling, 2006). Standards-based education has been
used to mask the influence that family and social reproduction have on student
attainment. The tension caused by a government which has attempted to have
high standards, yet, at the same time desired to support all students without

marginalizing certain populations has not been addressed. Thus, a socio-political
framework assisted in identifying weaknesses in the assumptions of current
political thought.
A common tactic used to justify policy changes is the portrayal of
marginalized groups as deficient (Stein, 2004). Because only those groups in
power are given the authority to make policy decisions, labels such as deprived,
deficient, or even culturally handicapped are often institutionalized (Stein, 2004).
It is through this type of labeling that policymakers have more easily served
special interest groups as well as economic concerns by naming a problem to be
solved (Stein, 2004). Additionally, many times, policy reforms have been
further impugned because they are fragmented and not aligned with other policies
and mandates in effect (Fullan, 1996). Because Colorado has one state department
for K-12 education and another for postsecondary education, the possibility of
policies fragmenting the learning experiences of students, thus further
exacerbating inequalities in the educational system, is high. Historically, these
policy processes have been driven by positivistic mathematical models, such as
the one used in the creation of the HEAR. These techniques by themselves were
not specific enough to focus on the structural realities or to consider all of the
social actors (Stein, 2004). Processes that consider interpretive, cultural, or
anthropological examination of the narrative elements of a policy should have
also been examined as they are more thorough. Cultural policy interpretations

involve all social actors in the policy process and pay attention to both the
historical moment and the structural realities of the situation (Stein, 2004). Such
analysis would have provided an important component for questioning the
assumptions built into policy goals as well as examining the true effects of certain
elements of HEAR outcomes.
According to Nieto (2000), several factors need to be considered when
examining the experiences of students from a socio-political perspective. These
factors include racism and discrimination, cultural and language differences
among students, and school structures, such as educational policies and practices.
School structures are particularly insidious when they reinforce social inequalities
and create barriers to the opportunity structures. Additional focus on literacy may
be one means of assisting the removal of barriers seen in todays culture (hooks,
2003). Without adequate preparation in literacy, students are less able to focus on
comprehension as they struggle with the basics of the structure of grammar
(hooks, 2003; Gutierrez, 2001). Within the educational system, students from
multicultural backgrounds are predisposed to practices such as tracking,
differential curriculum and pedagogy, and disciplinary policies (Nieto, 2000).
Shame caused by these practices often plays a role in preventing excellence
within marginalized groups as students become disengaged (hooks, 2003).
Use of a socio-political framework was important for this study in order to
look at both the political and social forces. For example, examining Californias

English Only policies and statutes, what has happened to English Language
Learners (ELL) begins with the increase in exclusionary behaviors (Gutierrez,
2001; Ladsen-Billings, 2006). Mediation efforts have been stripped in California
and teachers with bilingual skills basically de-skilled as they are not allowed to
teach in any language other than English (Gutierrez, 2001). Developing English
skills in this manner compromises a childs ability to acquire a deeper
understanding of literacy, and often comprehension suffers (Gutierrez, 2001).
Students cannot move on to more advanced skills such as comprehension if the
basics of grammatical structure and reading skills are not in place (hooks, 2003).
Student behaviors must be examined in terms of their cultural systems, and
teachers must realize that differences are not necessarily disorders (Zamora-Duran
& Artiles, 1997).
An additional hurdle faced by ELL students is the current standards
movement (Sleeter & Stillman, 2005). The results of these efforts have produced
a more standardized curriculum that affirms disciplinary boundaries and those that
surround English knowledge. This makes higher order thinking skills more
accessible to English-speaking students with at least an average level of reading
skills than they are to ELL students (Sleeter & Stillman, 2005). Curriculum is
being organized scientifically for efficiency, based on the decisions of the
policymakers, not the teachers in the classroom. According to Sleeter & Stillman
(2005), this approach to curriculum development is not about improving student

learning, but asserting who has a right to define what schools are for, whose
knowledge has the most legitimacy, and how the next generation should think
about the social order and its place within it.
Racism is alive and well in the United States of America (Ladsen-Billings,
2006). Continuous exposure to microagressionsthrough which progress like
desegregation of schools has been compromised, counseling in schools is not
equitable, and the education given to oppressed groups continues to lag far behind
those in the middle- and upper-class schoolsdemands that policies that continue
to subvert children must be examined (Ladsen-Billings, 2006). Barriers to
education and opportunity come in multiple forms.
To sufficiently understand the impact of the implementation of the CCHE
admission policy (HEAR), it was examined in terms of its socio-political context.
For students who have not taken a college preparatory curriculum in high school
and yet have gone on to college, it is important to understand the reasons as to
why they did not pursue a college preparatory curriculum and yet ended up in
college. To what extent did these students not achieve in high school, and for
what reasons? Further, it is important to understand the factors that seem to
support student success despite having taken an alternative high school
To study differential achievement, Nieto (2002/2003) outlined four
questions that must be addressed in order to fully understand the equity issues

associated with the lack of attainment: (a) whos taking Calculus, (b) which
classes meet in the basement, (c) whos teaching the children, and (d) how much
are children worth. Calculus, for Nieto, is used as a metaphor for access. Adelman
(2006) noted the differences in access by examining the offerings of high level
math courses in high schools that serve minorities and low income students.
When examining Calculus on a national level, 19% of white students were
enrolled in Calculus compared to 14% of Hispanic students and 8% of black
students (National Center for Education Statistics, 2005). Without access to
coursework such as this, the negative effects on students could be lifelong and
dramatic (Nieto 2002/2003).
For Nieto (2002/2003), segregation of students has also served to enforce
inequalities in the educational system. Oftentimes, courses for special needs
students are hidden away from public view, or in the basement (Nieto
2002/2003). Special needs can include anything from developmental coursework
to English language learning (Nieto, 2000). The effects of marginalizing students
in this manner can result in greater rifts between these students and those in the
regular classes. In addition, not only are schools being segregated in this
manner, but much of the progress made in the effort to desegregate schools prior
to 1988 has also been lost (Nieto 2002/2003; Orfield, 2001). By 2000, the
proportion of African-American children in predominantly minority schools had
risen from 62% in 1988 to 70% (Orfield, 2001). Segregated (or separated) is not

equal in terms of quality or resources. Segregation is a quiet phenomenon that
continues to steadily erode the benefits of integration and inclusion (Nieto,
Another issue associated with equality in education as it relates to the
socio-political framework is who is teaching the children. Although the
qualifications of teachers in lower socioeconomic schools tend to be significantly
less than those in affluent white neighborhoods (Lankford, Loeb, & Wyckoff,
2002), perhaps the more important problem is the lack of diversity among the
teachers themselves (Nieto, 2002/2003), or even the racist attitudes some teachers
bring (Howard, 2006). Studies have consistently shown that it is critically
important that students have teachers that look like them (Nieto, 2002/2003;
Zirkel, 2002; Clewell, Puma, & McKay, 2001). Even one race-matched role
model can make a significant impact on the academic achievement of students,
including marked increases on student test scores (Zirkel, 2002; Clewell et al.,
2001). A lack of diversity among teachers means that white students consistently
have teachers of the same race, and that students of color, more often than not, do
not have race role models among the teachers and administrators of their schools.
Nietos (2002/2003) final question relates to resources allocated to schools
and how this allocation affects students chances for success. Kozol (1991) aptly
describes the differences in resources between upper and lower class schools and
the significant disparity in the quality of education that results. In New York, for

example, differences in average per-pupil expenditures are stark, with the more
affluent public students receiving nearly three times the funding as those in low-
income schools (Kozol, 1991). An educational system that teaches all students
equally depends on an equitable finance system (Adams, 2000).
Combining these factors into a conceptual framework of student experience as
they related to educational policy (see Figure 1.1), the socio-political framework
used in this research showed the dominant culture and resulting racism have
driven governmental policy as well as affected the influence of social
reproduction through the family. Governmental policy and law created by the
dominant culture has exerted control on the following factors: curriculum, school
resources, student integration, and student role models and mentors. Standards of
success have also been defined by these same policy-makers. The resulting
system, along with family influence, has contributed to the students high-school
experience. Without systemic reform and a complete understanding of the system
to insure equity in access throughout the educational system, the CCHE admission
policy is at risk of creating further societal inequalities in the Colorado
educational system by increasing reliance on hegemonic standards and not
acknowledging the influence (or lack thereof) of family and social capital in
social reproduction.

Figure 1.1 Conceptual Framework

1.4 Research Questions
The purpose of this research was to evaluate the CCHE Admission Policy
(HEAR) implemented in the summer and fall of 2008. The study developed
findings and responses to the following research questions:
1. To what extent do the components in the CCHE admission policy
(HEAR) predict first-semester college GPA for students at one less selective
liberal-arts institution?
2. Which components are the best predictors?
3. Why do some students defy the prediction model by performing at a
much higher level than is predicted?
Examples of sub-questions for Research Question 3 include:
1. Do these students have family support structures that assist them?
2. Does the college provide support structures that have proven effective
for these students?
3. Do these students have role models in their school who are
demographically similar to them?
4. Have these students become integrated socially and/or academically at
the college?
5. How do the students perceive their high-school experience?
1.5 Overview of Methodology
The research was conducted in two phases as a sequential quantitative-

qualitative mixed-methods study (Tashakkori & Teddlie, 1998). The purpose of
using this strategy was to identify extreme cases in order to refine the
understanding of the usefulness and practicality of the HEAR (Caracelli &
Greene, 1997). Contradictions that existed in the data were further explored
through the use of this methodology (Greene, Caracelli, & Graham, 1989).
1.5.1 Phase I of the Research Quantitative
During the first phase of the research, the model of the CCHE policy was
reconstructed through the use of a path model to predict students first semester
college GPA (see Figure 1.2). The following data were collected from student
high-school transcripts for each aspect of the model: the CCHE entrance-
examination performance index as noted in the Technical Appendix to the
Admission Policy (2003), the CCHE high-school performance index as noted in
the Technical Appendix to the Admission policy (2003), and the categorization of
all courses taken in high school. With the longitudinal set-up, the model displayed
the relationship of each variable to success as defined by the college GPA. More
specifically, the number of courses taken in each of the HEAR areas theoretically
predicted a students high-school performance index as well as a students college
entrance-examination index. The college entrance-examination index also served
to predict the high-school performance index because each student took the
standardized test prior to completing the high-school performance index.

Figure 1.2. Path Diagram of Colorado Admission Policy
Theoretically, the high-school performance index and college entrance-
examination index predicted the students first-semester college GPA. The
relationship of high-school course work to first-semester college GPA was shown
to be indirect. Specific coding used for high-school courses is available in
Appendix A. A more thorough discussion of how the indices were calculated will
be presented in Chapter 2.
The population of interest included all first-time, baccalaureate-seeking
students enrolled at the liberal-arts college during the fall of 2005 who also
graduated high school in the spring of 2005. Regression equations constructed
from the path analysis enabled the calculation of the predicted first semester
college GPAs. The results of the prediction model, along with the CCHE

requirements, were examined to identify students whose college performance did
not support the model. Students who had been predicted to fail during their first
semester but were subsequently retained through their first year in good standing
and continued into the spring of their sophomore year made up the population of
students whose performance countered the predictions of CCHEs admission
1.5.2 Phase II of the Research Qualitative
In Phase II of the research, a sample of the outlier students identified in
Phase I was selected as subjects for the interviews (Yin, 2003; Merriam, 1998).
The students were purposefully selected based on being first-generation, African
American men who were predicted by the Phase I model to fail in their first
college semester, yet were successful through their second spring semester. Using
tiered interviews, archived student records, demographic information about the
high schools, and observations, data were collected from the selected cases to
provide information on and give voice to the students who would be adversely
affected by the implementation of this policy (Yin, 2003; Merriam, 1998). The
role of the researcher was integral, having previously been the analyst of the
policy at the CCHE and an institutional researcher at the time the study was
conducted. The resulting bias and limitations of the study will be further
discussed in Chapter 6.

1.6 Structure of the Study
The first chapter of the study contains an outline of the issue, arguments
for a thorough impact analysis of the HEAR, a discussion of the conceptual
framework used in the research, and an overview of the methodology. The second
chapter examines the literature on the development of rigor in American high-
school curricula as critical to success in college, social reproduction as it pertains
to African American males, and a discussion of the history of Colorado higher
education admissions policy. Chapter 3 expands on the methodology of the study,
including sampling, measurement, and data collection for each phase of the
research. Chapter 4 presents an analysis of the results and findings of the
quantitative phase of the research as they align with the research questions that
were posed in the first chapter. Chapter 5 discusses the results of the qualitative
phase of the research and identifies the emergent themes of the study. Finally,
Chapter 6 concludes the research by summarizing and interpreting the findings.
The research as it relates to the conceptual framework is explored, and the
relationship between the quantitative and qualitative phases of the study
discussed. Conclusions and implications for current practice and future research
are also provided.

2. Literature Review
The purpose of this research was to study the predictive validity and
possible adverse impact of the HEAR requirements implemented by CCHE,
beginning with the high-school graduating class of 2008. Because concerns have
been raised by rural high schools and districts in the state regarding the
practicality of implementation (CCHE, 2006; Colorado Association of School
Boards, 2006), as well as by other special interest groups (Colorado Association
of School Executives, 2006), this study examined the extent to which the
Colorado Admission Policy, including the HEAR, predicted success as measured
by first semester college GPA.
Also, the study identified those students who would have been adversely
affected by the additional requirements and explored the means each of them used
to defy the models predictions. The research specifically looked at the effects of
this policy on African American males. Prior studies have repeatedly indicated the
underachievement of African American males at all levels of education from
primary to secondary to postsecondary (Monroe, 2005; Ferguson, 2003;
Hrabowski, Maton, & Grief, 1998). Hegemonous policy has often been identified
as one source of the problem (Fenning & Rose, 2007). Other researchers point to
centuries of systemic discrimination that have affected the abilities of African

American children to achieve at levels comparable to Caucasian children
(Mandara & Murray, 2007). This study investigated the impact HEAR will have
on this population.
This chapter features a comprehensive review of the theoretical,
methodological, and practical evidence that correlates to the problem. Literature
critical to the study is discussed. The scope of this review reflects the complexity
of factors associated with the problem, the social considerations that need to be
addressed, and the challenges associated with the design of a model system that
addresses all of these factors.
The key questions and challenges are highlighted through this discussion
of the literature. The review includes relevant findings from three domains: the
Colorado higher education admission policy, high-school curriculum structures,
and the resultant impact of the policy and curriculum structures on African
American males. Important questions in the minds of key constituencies and
stakeholderssuch as researchers, faculty, students, policy-makers, and
consumersare addressed. The research is cited, reviewed, summarized, and
critiqued in order to provide a clear picture of the evidence as it relates to the
design, potential, and validity of the Colorado higher education admission policy.
The review of the literature for this chapter is organized into five sections.
The first section focuses on the historical overview of the Colorado admission
policy and resultant policy issues. The model first employed with the high-school

graduating class of 2008 is defined, and the limitations of the data used to build
this model explained.
The second section is a review of the literature that is related to the
barriers faced by African American men in the educational system. Racism,
exclusionary discipline, placement through tracking, and issues associated with
family structures are discussed. Social reproduction of the resultant inequalities is
also outlined.
An overview of the link between high-school curriculum and college
success is presented in the third section. Exploring the development of high-
school curricula in this country assists in determining which courses are truly
necessary to implement from a policy perspective. If the majority of students need
or want a college education, it stands to reason that high-school competencies
necessary for success in college should be made available. However, if other
factors can also be linked to success, these should also be explored. In addition to
the high-school curriculum, this section examines the literature as it relates to
other factors tied to college success.
Section four features a discussion of the historical development of the
American comprehensive high school. This section is relevant in that it explains
how and why high-school curricula expanded to include nontraditional
coursework. It considers the theory behind the development of the dichotomous
nature of the comprehensive high school. A summary of the literature is presented

in the final section, which includes a discussion of the American high-school
cycle of evolution. The movements from traditional to comprehensive and back to
traditional high schools are delineated.
2.1 Overview and Limitations of Colorado Admissions Policy
The structure of the Colorado admission standards was initially
implemented in May 1985, when the Colorado legislature passed into law the
determination of an entrance requirement that would entail the use of college
entrance-exam scores (ACT or SAT), high-school grade-point averages, and/or
high-school ranks for public 4-year institutions of higher education in the state
(CCHE, 2005). A committee of higher education governing board representatives,
in conjunction with CCHE staff, worked to develop an admission index that
standardized these variables in order to attempt to apply them fairly across
students and institutions. Application data were collected in 1985 and updated in
2003 in order to equate the ACT and the SAT as part of a college entrance-
examination index. The same was done for high-school grade-point average and
high-school rank, creating a high-school performance index (CCHE, 2003). Using
an equipercentile methodology, as described by Kolen and Brennan (1995),
CCHE staff compared data. In order to insure homogeneity of populations, only
2002 high-school graduates were used in the 2003 update. Once the data were
equated, standardized t-score distributions were calculated for each index. The
overall freshman admission index was then calculated by summing the high-

school performance index with the standardized test index. An index cut-score for
admissions purposes was established for each 4-year public institution of higher
education in Colorado with cut-scores ranging from a low of 76 to a high of 110
(CCHE, 2003).
When exploring how other states were dealing with minimum high-school
requirements, more than 40 states were found to have some sort of minimum
high-school graduation requirement (National Center on Education Statistics,
2002). As the analyst for the CCHE, the author began to look for guidance on
how to implement a curricular-admission standard, and Georgia was one of a few
states that appeared to have dealt with similar circumstances. In interviewing the
Associate Vice Chancellor for P-16 Initiatives from the Georgia Board of
Regents, the author learned that in the early 1990s, Georgia dealt with divergent
high-school graduation requirements and also implemented an admission index
based on standardized-test scores and high-school performance. Georgia also
decided that the limitations of using only an admission index needed to be
examined in conjunction with high-school curricula. In 1993, Georgia
implemented a minimum curriculum for entrance into public 4-year institutions.
CCHE staff began conversing with the implementers of Georgias requirements,
finding that such requirements did not damage Georgias college enrollment,
retention, or graduation numbers. On the contrary, since implementing a rigorous
high-school curriculum in 1993, enrollment, retention, and six-year graduation

rates had all increased. As a result of the success of Georgias program, Georgia
has subsequently implemented high-school graduation requirements as part of its
P-16 Initiative (Georgia Board of Regents, 2006).
Using Georgias apparent successes as a model, the author began
analyzing data and literature in order to determine the best course-taking pattern
to implement as a minimal requirement for entry into one of Colorados 4-year
public institutions. As a result of these analyses and a literature review conducted
by CCHE staff, the CCHE approved the implementation of a core curriculum that
consisted of the following: (a) four years of English, to include two years of
composition and grammar; (b) three years of social studies; (c) three years of
natural science, of which two must contain laboratory work; (d) three years of
mathematics at the Algebra I level or higher; and (e) two years of academic
electives. The combined coursework resulted in a total requirement of 15
academic units that began with the high-school graduating class of 2008. By
2010, these requirements will increase to four units of mathematics at the Algebra
I level or higher and an additional one unit of foreign language (CCHE, 2005).
In examining the coursework options, CCHE staff examined much of the
literature surrounding the current standards-based movement, beginning with the
much-publicized 1983 report, A Nation at Risk. During the same year, the College
Board published Academic Preparation for College: What Students Need to Know
and Be Able to Do, stating that coursework should include English, mathematics,

social studies, natural science, and foreign languages. Several subsequent studies,
including work by Adleman (1999), Kirst & Venezia (2001), and Rose & Betts
(2001), were also reviewed by CCHE. The foundation for the current Colorado
structure was based upon this review of the literature along with analyses of the
student data.
The addition of the HEAR as a component in the Colorado admission
policy was completed under a positivistic methodology supported with research
conducted by the author of this study. An important piece that CCHE and the
author neglected to analyze thoroughly was who this policy would adversely
affect and whether or not the data could justify these results. The long-term
effects of the Georgia implementation raise concerns that Georgias high-school
dropout rate has soared (Manhattan Institute for Policy Research, 2003). Although
no direct link to the revised Georgia admission policy was indicated, since
implementation of the increased standards, high-school graduation rates in
Georgia have declined significantly, with the numbers of students dropping out
steadily increasing. Prior to the policy, in 1991, Georgias rate of high-school
graduation was 68% (Manhattan Institute for Policy Research, 2005). Since the
implementation, the rate fell to a low of 56% for the class of 2003 (Manhattan
Institute for Policy Research, 2006). Therefore, the possibility of high-school
graduation rates falling in Colorado as a result of these policy changes should be

considered. The students potentially adversely affected by the implementation of
this policy should be identified and studied.
In addition to the practical limitations of the CCHE Admission Standards
Policy, several theoretical arguments exist against the type of positivistic model
constructed by CCHE. Largely based on Marxist thought, two of these
frameworks, social reproduction and critical theory, provide important further
background to support the necessity for this study.
In direct opposition to CCHEs positivistic model and the notion that the
schools should merely offer the coursework to the students is the construct of
social reproduction, a facet of sociological theory rooted in Marxian tradition.
Based on the notion of social reproduction one could argue that, although it
appears that merely providing the coursework for all students would begin to
address the achievement gaps seen in society, this is a very naive position to take.
Reproductionists believe that inequalities in society are much more deeply rooted
and that the K-12 system is incisive in the ways in which it reproduces those
inequalities (Oakes, 1982; Bowles & Gintis, 1976). Reproductionists are much
more cynical than many, suggesting that the societal structure is at fault,
maintaining biases and inequities merely by the way in which the institutions
within society are structured. However, proponents of social reproduction theory
do realize that the institution of education is critical if underserved populations are
going to permeate upwards into the stratified system and reduce the gaps that

have seemingly been inherent in our society (Bowles & Gintis, 1976; Featherman
& Hauser, 1978; Hout & DiPrete, 2004).
Social reproductionists maintain that the educational institutions in our
society perpetuate inequalities through tracking mechanisms whereby those
students who can better conform to the curriculum and pedagogical methods of
the hegemenous society will advance while others who need to spend time
acclimating to traditional and dominant culture teaching styles and materials will
fall further and further behind (Oakes, 1982; Freire, 1972/2000; Bowles & Gintis,
1976). Most often, this has led to the instantiation of ability tracking systems
which in turn, have been shown to reflect the relative social position of their
families (Oakes, 1982). Oakes (2005) also found that students in lower tracks tend
to be more passive regarding their status and more disruptive towards each other.
Disciplinary actions against students who find themselves in low ability tracks
tend to be more severe than for those in higher tracks. In examining course
sequences within tracks, however, evidence indicates that following sequences,
particularly in math, is what makes the difference in college preparation, not
whether the student was in a high or low track (Stevenson, Schiller, & Schneider,
1994). The number of upper-level offerings along with concise guidance in
curriculum within a high school was also found to be important in providing
opportunities for success in higher education (Spade, Columba, & Vanfossen,
1997). As noted by Adelman (2006), 76.9% of high schools attended by white

students offer mathematics curriculum above the Algebra II level, compared with
only 59.9% of high schools attended by Latino students. How do these schools
accommodate their high-achieving students if they are not offering the courses?
The opportunity to take higher level math coursework must be given to high-
school students if a single standard of college admission is to be applied.
A similar theory, also bom from Marxism, is critical theory. From a
critical theory perspective, the problem with assuming that all students should
successfully complete a core curriculum in high school is that not all students are
offered the opportunity to do so nor have they been prepared previously to be
successful (Gramsci, 1931/2000; Adelman, 2006). Within society, a power
differential exists. According to Gramsci (1931/2000), the hegemenous society
dictates what is important to know and what is important to learn in school.
Marginalized students whose cultural values or experiences lie outside of the
dominant culture are at an extreme disadvantage from day one and are often the
students who end up in the lower tracks in school (Delpit, 1995; Hale, 2001;
Kunjufu, 2004). Although these students may know more in areas deemed
unimportant by the dominant culture, they cannot use what they have learned to
their advantage in school.
Additional disadvantages occur when cultural behaviors impede a
students learning in a hegemenous setting. Behaviors accepted as normal at home
may be labeled disruptive in school, and as such, the child is punished for a

cultural difference (Delpit, 1995; Hale, 2001; Kunjufu, 2004). As these students
embark on the journey of completing a high-school education, every incident of
punishment and labeling pulls them further and further behind in the system. In a
structured sequential discipline like mathematics, this type of behavioral cycle can
be devastating to a students progress.
Furthermore, teachers have a manipulative form of power over students in
their classrooms. Part of the reason that manipulation is so insidious is that
exploitation of the oppressed becomes more psychological (Mills, 1948/2001;
Mills 1951; Mills, 1956). Mills (1948/2001) believed that manipulation, because
it is hidden, makes the identification of the oppressor more difficult, and in doing
so, makes it easier for oppressors to continue being more oppressive. Unrestrained
manipulation can strip all meaning from an education, making it easy for some
students simply to walk away, dropping out in many cases.
In light of the evidence that most students at some point in their careers
intend to pursue a college education (Conley, 2005) and that success in college
appears to be greatly enhanced by a solid academic curriculum in high school, K-
12 educators have two choices: (a) recognize that different pedagogical strategies
may be necessary to teach diverse student bodies and look for multiple ways of
assisting diverse students in attaining the competencies necessary for continuing
and completing their educational goals; (b) maintain the status quo. Many
theorists, like McWhorter (2001) of the Manhattan Institute, believed that failure

becomes more acceptable when different cultural explanations are allowed.
Subsequently, an attitude of victimology is created. At some point, each person
needs to take responsibility for his or her own situation (McWhorter, 2001).
Although the need for personal responsibility is essential, if an educator can save
a child simply by teaching the skills in a more meaningful way, does that not
become an educators personal responsibility, as well? If not, what is the
educators responsibility?
2.2 The Social Reproduction of African American Males in the Educational
SystemRelated Studies
African American males in the educational system are often identified as
being at-risk and, as a population, continue to underachieve at a disproportionate
rate (Jackson & Moore, 2006; Davis, 2003; Bailey & Moore, 2004; Fenning &
Rose, 2007; Monroe, 2005; Ferguson, 2003; Hrabowski et al., 1998). These
students are more likely than any other group to be suspended or expelled from
school (Fenning & Rose, 2007). Compromising the ability of African American
males to complete their education, the overrepresentation of African American
males in the exclusionary discipline practices of suspension and expulsion is
profound (Fenning & Rose, 2007; Gonzalez & Szecsy, 2004; Skiba et al., 2000;
Skiba & Peterson, 1999; Skiba & Rausch, 2006). Statistics have shown African
American males are suspended up to five times more often than white males
(Monroe, 2005). Not only are they punished more often, but African American

males are consistently punished more harshly for subjective offenses for which
white males are often not punished at all (Skiba, 2000). Teachers often interpret
African American behaviors as inappropriate when the students intent is not to be
improper (Monroe, 2005; Weinstein, Curran, & Tomlinson-Clarke, 2004;
Weinstein, Tomlinson-Clarke, & Curran, 2003). The students are punished
without knowing the consequences of their actions, and often the punishment
includes removal from educational opportunities.
Educators fears, rather than actual threats, can explain why many of the
students who are labeled at-risk end up restricted through exclusionary
discipline (Skiba & Peterson, 1999). Those seen as not fitting in are most often
targeted for removal and placed on the fast track to the prison system (Wald &
Losen, 2003; Fenning & Rose, 2007). These factors serve to criminalize young
African American males rather than educate them (Monroe, 2005), perhaps
conditioning students for their future in the criminal justice system. Although
African American males comprise only 6% of the countrys population, they
represent over 50% of the penal population (Kunjufu, 2001).
Fenning and Rose (2007) argue further that the No Child Left Behind
legislation, along with other accountability programs, has heightened the pressure
for removal of students who do not fit into the hidden behavioral curriculum
known to the dominant student population. In addition, when behaviors are
questioned, it is the teachers subjective perception that determines what the level

of infraction should be, possibly leading to the removal of the student from the
classroom (Vavrus & Cole, 2002). No tolerance policies have also been directly
linked to increases in suspension of African American males from school
(Ogletree, 2007). In some states, black suspension rates are as high as 25 percent
(Ogletree, 2007).
Along with these practices, these students are more likely to be in low-
performing schools with less qualified teachers (Lankford et al., 2002). Not being
challenged, being underrepresented in gifted and talented programs as well as
advanced placement courses (Adelman, 1999; Adelman, 2006; Grantham, 2004;
Ford, Grantham, & Whiting, 2008; Hrabowski et al., 1998), and being tracked
into remedial course work and special education (Oakes, 2005) are all reasons
these students may underachieve or disengage from school (Ford, 1996). Family
also plays a key role in the underachievement of these young men. African
Americans are more often living in poverty than their white counterparts and
more often in single-parent situations with parents having less education
(Mandara, 2006).
Advanced level mathematics courses are an important factor related to
continued success in school (Adelman, 2006; Hallinan & Kubitschek 1999; Lucas
1999). Unfortunately, research has historically suggested that mathematics course
patterns are biased, especially regarding African American and Latino students
who tend to take fewer advanced mathematics courses compared with White

students (Adelman, 2006; Ladson-Billings, 1997; Lucas 1999; Oakes 2005).
Suggestions for why this underrepresentation occurs range from lack of offerings
in the schools (Adleman, 1999; Adleman, 2006) to the students choosing not to
participate, even at prominent high schools with high academic achievement
(Ogbu, 2003).
The tiered structure of mathematics also creates a disadvantage for those
students who are not prepared or for those not given the option to begin their
high-school coursework at the Algebra I level by the time they become freshman
students (Schneider, Swanson, & Reigle-Crumb, 1998). The other HEAR
disciplines are not as strictly connected and are more open to students of all levels
and abilities (Lucas, 1999). If students cannot begin their high-school
mathematics curriculum with at least Algebra I, the chances of completing the
HEAR, particularly for 2010, is seriously compromised. In addition, those
students have very little opportunity to access the advanced math courses that
most strongly predict college attendance and success (Adelman, 1999; Adleman,
2006; Riegle-Crumb, 2006).
Further, Riegle-Crumb (2006) found that, while taking Algebra I will, on
average, allow a student to successfully reach an advanced course position by the
end of school, this finding does not hold true for African American males. These
students may feel uncomfortable and unsupported, particularly if the environment
contains predominantly white students, although they may still respond to high

teacher expectations (Gutierrez 2000; Riegle-Crumb, 2006; Yonezawa, Wells &
Serna, 2002; Ogbu, 2003). Alternatively, the students may be high ability but
placed with teachers who hold biased opinions regarding the abilities of minority
students (Oakes & Guiton 1995; Ogbu, 2003). A lack of a culturally relevant
curriculum also impedes students from maximizing opportunities (Gutstein,
Lipman, Hernandez, & de los Reys, 1997; Ladson-Billings, 1997; Riegle-Crumb,
2.3 Overview of High-School Curriculum
The link between high-school preparation and college success has been
well documented. Advocates of a rigorous high-school curriculum can point to a
number of studies that support this relationship; a strong math curriculum has
been shown to have a direct relationship to success in college: Students who take
Algebra II or higher in high school more than double their chances of completing
a baccalaureate degree (Adelman, 1999; Adelman, 2006; The College Board,
1983a; Pelavin & Kane, 1990; Rose & Betts, 2001). Advanced math in high
school not only relates to a significantly higher probability of college graduation
but appears to be a strong indicator of earnings after completion of college (Rose
& Betts, 2001); furthermore, students who complete two consecutive years of
foreign language are also more likely to attend college (The College Board,
1983b). The relationship between a rigorous high-school curriculum and college

success appears to be profound. With greater rigor comes a greater likelihood that
students will not only attend college but also persist to degree completion.
As the need for a college education increased, college attendance rates
rose to nearly 51% of adults by 2000 (Kirst & Venezia, 2004). Several scholars
proposed this number could still be higher if barriers to access were reduced.
Ingels, Curtain, Kaufman, Alt, and Chen (2002) found that, although more than
65% of eighth graders in 1988 stated their desire to attend college, less than 30%
did so in 2000. Conley (2005) reported that nearly 90% of ninth graders stated
their intention to attend college, of which approximately 67% went directly on to
some form of postsecondary education.
In spite of these trends, several studies have examined factors in addition
to rigorous curriculum that appear to be related to college attendance and
completion (Martinez & Klopott, 2005; Adelman 2006; St. John, 2004; Cabrera &
La Nasa, 2000; Horn & Kojaku, 2001; Rose & Betts, 2001; Heller, 2001;
McDonough, 1997). These studies suggested that the following factors have
shown to be the best predictors: (a) rigorous high-school preparation, (b) social
support networks, (c) access to information regarding the college process, (d)
parental involvement and knowledge about college, and (e) financial aid.
Although a rigorous high-school curriculum has been shown to be the best
predictor of college success (Adelman 1999; Adelman 2006; Checkley, 2001),
preparation alone is not sufficient to increase college-enrollment rates (Martinez

& Klopott, 2005). Social support networks consisting of family members,
teachers, counselors, and peers have also been shown to be highly related to
college attainment (Oakes, 1983; Gandara, 1999; Croninger & Lee, 2001; Lee &
Burkham, 2000; Cabrera & La Nasa, 2000; McDonough, 1997).
Several high-school reform programs that focus on developing support
systems for students have been implemented in various areas of the country
(Martinez & Klopott, 2005). Programs vary from implementing small learning
communities to enlisting student family structures as support systems.
Repeatedly, schools that build strong support networks for their students, where
meaningful relationships are developed, show student gains in school, including
increasing college-enrollment rates (Martinez & Klopott, 2005).
2.4 History of the High-School Curriculum
In examining the link between high-school curriculum and college
success, one would be remiss to ignore the significance of historical precedent. In
preparing a student for college, high schools must be cognizant of admission
expectations at various types of institutions of higher education. As advocates of
more progressive types of curricular activities continue to discover, straying away
from traditional core curricula in favor of a more expansive and experimental type
of course offering structure may be controversial as well as damaging to the
future of our children (Johanek, 2001). This section will focus on the historical
aspects of the alignment of K-12 graduation requirements and college preparation,

beginning with the National Education Associations (1894) report of the
Committee of Ten followed by the establishment of college entrance reports
through A Nation at Risk (National Commission on Excellence in Education,
1983), and ending with the No Child Left Behind legislation (Elementary and
Secondary Education Act, 2001).
In order to understand fully the evolution of the high-school core
curriculum in American education and how it relates to success in college today,
one must begin with an examination of its development in this country. The report
that came out of the Committee of Ten'in 1893 is often cited as the first major
attempt at standardizing high-school curriculum and is still recognized as
important in the development of curricular activity today (Center for the Study of
Mathematics Curriculum, 2004). Convened in 1892 by the National Education
Association (NEA), Harvard President Charles Eliot led the committee in its
charge to examine the alignment of high-school programs with college admission
requirements (National Education Association, 1894). At that time, nine subjects
including Latin, Greek, English, modem languages, mathematics, science, natural
history, history and government, and geography were examined in order to
determine the extent to which each should be taught as part of the optimal
curriculum (Center for the Study of Mathematics Curriculum, 2004). Although
subject-matter experts all espoused the benefits of their particular disciplines as
being of primary importance, the report also established guiding principles. First

of all, the Committee unanimously agreed that, regardless of whether a student
was pursuing college, scientific school, or other studies, students should take the
same core curriculum. No differentiation was advised (Center for the Study of
Mathematics Curriculum, 2004). Second, the Committee, although recommending
a core curriculum, recognized the importance of flexibility in student choices
(Center for the Study of Mathematics Curriculum, 2004). Finally, the Committee
recommended following a uniform standard so that students graduating from high
school would be able to meet admission requirements at a variety of institutions
(Center for the Study of Mathematics Curriculum, 2004).
Within two years of the Committee of Tens findings, however, the NEA
commissioned another committee to examine college entrance requirements
(Wechsler, 2001). Some accreditation agencies like the North Central Association
of Colleges and Secondary Schools began to refer to the recommendations of the
Committee on College Entrance Requirements not only in establishing standards
of college admission but also in certifying high-school curricula as college
preparatory (Wechsler, 2001).
However, not all accrediting agencies agreed to use the certification of
high schools, believing that the standards for college admission were more
important (Rudolph, 1962/1990). The Association of Colleges and Secondary
Schools of the Middle Atlantic States decided the better route was to create a
board of examiners. As a result, the College Board was created in 1900, with its

first set of exams offered shortly thereafter (Rudolph, 1962/1990). The subject
matter consisted of the nine areas examined by the Committee of Ten. The
immediate issue that appeared with regard to this testing was its exclusion of
alternative coursework as a means to enter into higher education.
2.4.1 Progressivism
Beginning with Dewey, the progressive movement in education ran
counter to the work done by the Committee of Ten (Blocher, 2000).
Progressivism had begun during the latter part of the nineteenth century, and
Deweys advocacy of the production of knowledge as opposed to more traditional
views of training the mind are well recognized as the basis upon which much
educational reform was developed (Westbrook, 1993; Dewey, 1910/1997). In
regards to the Committee of Ten report, Dewey (1896/1983) saw the dichotomous
position in which high schools were placed as unstable. The societal expectations
of preparing some students for college while educating others for community
demands were problematic. Aligning K-12 and higher education was far more
complicated than it seemed, given that access to primary education came out of a
democratic school of thought while higher education stemmed from a more elitist
notion of education (Dewey, 1906/2002).
Growing out of the progressive movement, the desire to make the
educational system more practical gave rise to the beginnings of vocational
curricula in high schools. In 1918, the Commission on the Reorganization of

Secondary Education published a report entitled Cardinal Principals of
Secondary Education, which aligned itself with the current progressive thought
(Commission on the Reorganization of Secondary Education, 1918). The
principles set forward included such educational objectives as vocation,
citizenship, health, worthy use of leisure, command of fundamental processes,
worthy home membership, and ethical character. The Cardinal Principles are
often recognized as the foundation of the comprehensive high school,
recommending a broad curriculum for a larger number of students (Hammack,
2004; Wraga, 1994).
In 1919, the Progressive Educational Association was founded (Kridel,
Bullough, & Goodlad, 2007), and the predominant movement in education was
based on a hands-on, leaming-by-doing approach that supported the growth of
vocational programs (Butts, 1955). Opponents of adding vocational education to
high-school curricula, including social reproductionists, saw this as another way
to keep people in their places (Bowles & Gintis, 1976). However, advocates of
this type of education saw it as a means to enhance access to education (Wraga,
In addition to opponent critiques, the problem with implementing these
more comprehensive high schools was almost immediately recognized. Colleges
were not convinced to change admission standards as a result of curricular
changes in the high schools. Several concerns were voiced, and research was

conducted to try to determine the impact on college admission if a student were to
choose to take vocational courses in high school (Proctor, 1927). The National
Committee on Research in Secondary Education determined that as many as
seven vocational courses could be taken in high school, providing that the student
also took at least two years of English and a total of 15 Carnegie units of
traditional academic courses (Proctor, 1927). The Progressive Education
Association expressed its concerns regarding this issue at a 1930 meeting, stating
that, although many of the desired changes could be made and perhaps even
should be made, it was necessary to recognize that this type of curricular change
could very well impair a students ability to attain college admission (Aikin,
1942; Kridel et al., 2007). Although other studies were conducted to try to
quantify the relationship between a high-school curriculum and success in
college, the research of the time indicated that overall grades and class rank were
more than predictive enough to indicate whether a student would succeed in
college, regardless of course-taking patterns (Aikin, 1942; Kridel et ah, 2007).
As the need for attaining a baccalaureate degree increased, the movement
towards a more progressive high-school curriculum, in conjunction with several
historic government interventions, allowed more and more access to students.
Beginning with the G.I. Bill in 1944, which gave veterans of World War II tuition
benefits, colleges began to see enrollments that may not have occurred otherwise
(Bowen, Kurzweil, & Tobin, 2005). It has been estimated that more than 20% of

the nearly two million veterans who took advantage of the college benefits of the
G.I. Bill would not have attended without the tuition afforded them by this
legislation (Bowen et al., 2005).
2.4.2 Combining Rigorous Curriculum and Comprehensive Ideals
Soon after the G.I. Bill was created, the 1947 Truman Commission
recommended the conception of community colleges, stating in its report that the
first two years of college should be available as part of a common experience for
all educated men and women (Presidents Commission on Higher Education,
1947). The creation of a general education component was established in order
to shift the authorities away from the elitist notions of education and towards a
more service-oriented and democratic one (Presidents Commission on Higher
Education, 1947).
One of the unintended consequences of these federal actions, according to
critics, was that the substantial investment in our system of higher education that
enabled access also was a major component in reproducing social inequalities.
These separate forms of education (vocational, practical) were not seen as equal
to the traditional scholarly education (Conley, 1995). After the Russian launch of
Sputnik in 1957 revealed problems associated with the United States deficiency in
pursuing rigorous math and science curricula, advocates of vocational and
practical higher education were effectively silenced (Hammack, 2004; Resnick &

Resnick, 1985). Subsequently, federal dollars were invested in pursuing a
redevelopment of math and science.
The resultant comprehensive high school contained elements of the
progressive movement as outlined by Dewey, traditional scholarly rigor, and
vocational education in an attempt to serve as many people as efficiently as
possible (Hammack, 2004; Snedden & Dewey, 1915/1977; Wraga, 1994). Former
Harvard University President James Conant believed that a successful
comprehensive high school should include not only a general program for all
students but a vocational program for most and a college preparatory program for
the few high achievers. In addition, strong leadership was determined necessary to
sustain such schools. Critics noted that the college preparatory curriculum under
Conants plans would be limited, at best (Hammack, 2004).
2.5 Present-Day K-16 Alignment
Alignment of the K-16 transition must begin with an understanding of
how the two are related if total alignment is to be accomplished. A high-school
diploma does not necessarily mean that a student is college-ready. Policy-makers
and legislators first need to decide to what extent alignment between K-12 and
higher education is desirable or necessary, and what subsequently needs to be
done to achieve such an alignment. Advocating rigid alignment through a fourth
year of college necessarily supports a curriculum such as the type recommended

by the Committee of Ten. The question of to what extent alignment is desirable or
necessary is one that must be answered.
Proponents of K-16 alignment provide strong arguments for their point of
view. The Stanford Bridge Project (2002) found that nearly 90% of ninth graders
expressed the intention to go on to college and that, within five years of high
school, 75% of them actually did enter higher education. According to some
researchers, one key to success is to convey clear expectations to students as to
what it will take to be successful in college (Kirst & Venezia, 2001; Venezia,
Kirst & Antonio, 2003; Kirst & Bracco, 2004; Conley, 2005). Alignment of
expectations, including those of the students, is important in reducing levels of
remediation in college and increasing persistence to degree. Increased remedial
needs in higher education result from a lack of communication of expectations
well before the transition from high school to college (Kirst & Venezia, 2004).
In addition to the number of high-school students that expressed the intent
to continue on to college, other factors have supported the need for a rigorous
curriculum and K-16 alignment. Since 1973, the proportion of factory jobs held
by workers with at least some college coursework tripled, and wages remained
consistent. At the same time, wages for workers with a high-school diploma or
less declined (Barth, 2003). Evidence suggests that our economy is structured to
employ people with the high-school preparation necessary for college-level work
(Camevale & Desrochers, 2003; Kendall & Williams, 2004). These indicators

support the position of standardizing the high-school curriculum so that students
receive enough of the necessary competencies to be successful in college or the
workplace. Indeed, many vocational students would also be hurt if not prepared in
the coursework necessary for college-level work.
As this study has alluded to unintended consequences of implementing
standards, one of the most profound has resulted in critics calling for the
elimination of tracking (Oakes, 1992) in favor of a more rigorous curriculum for
all students in order that options are not closed for students who are not initially
considered to be high achievers. In considering the argument that tracking
students is detrimental, one must begin to examine which academic courses
appear to be most linked to college success. Is there a threshold of coursework
that can be determined to be vital to such success? Which areas of the curriculum
can be substantiated as most important? Is the entire HEAR curriculum essential
to providing students necessary tools in their quest to become college educated, or
are some portions of the HEAR not significantly critical to student success? These
questions must be addressed in order to examine fully the relationship between
high-school preparation and subsequent success in college.
In addition to literature on appropriate coursework necessary to develop
college readiness, Powell, Farrar, and Cohen (1985) posited that one of the major
issues in student readiness for college may be the fact that students have had too
many options available with which to build their high-school schedules. Case

studies found that those students who suffer when given an overabundance of
choices in curriculum are those in the middle, those who are not performing
poorly but are not high achievers either. Without the delineation of a strong and
rigorous curriculum, many of these students would just do enough to get by. They
would not achieve their full potential (Powell et al., 1985). Additionally, work by
Lee, Croninger, and Smith (1997) suggested that schools with narrow curricula
with fewer elective options that were based on the delivery of academic
coursework actually improved student standardized test scores. Clune and White
(1992) examined the problem of low-achieving schools as compared to those with
high academic standards. These would be the schools which would have had the
hardest time complying with the new admission requirements. Schools with
higher graduation requirements were found to have already had the operational
mechanisms in place to accommodate the higher standards.
Implementing higher graduation standards, whether at the level of high-
school graduation or college admission requirements, appears to affect student
achievement in a positive way through an increase in core-course enrollment
patterns (Chaney, Burgdorf, & Atash, 1997; Clune & White, 1992), although
dropout patterns in states such as Georgia are of great concern (Manhattan
Institute for Policy Research, 2003). As students recognize the need to take more
courses to fulfill requirements, more core courses are taken. Subsequently, as
more courses are taken, student achievement increases. Chaney et al. (1997) also

examined the possibility that the implementation of more rigorous coursework
would impose negative consequences on students and found that the vast majority
of students do not fail but rather rise to the challenge. Could this be due in part to
dropouts no longer being part of the cohort?
As course-taking patterns approximate academic preparation, examining
the coursework taken by a student in high school does not only assist college
admission officers, for research indicates that the more a student achieves in high
school, the more likely the student will apply to college in the first place (The
College Board, 1983b). High achieving students are then more likely to apply to,
enroll in, and be successful in college (Chaney et al, 1997). In addition to college
preparation, rigorous coursework in high school prepares a student for many of
the technically oriented vocational fields (American Diploma Project, 2002; ACT,
Opponents to the movement towards a standardized curriculum argue that
this format limits many students opportunities by not providing comprehensive
instruction that would allow certain students to learn valuable vocational skills,
making them employable after graduation (Wraga, 1994). The standardization
within the comprehensive high schools, in an ideal setting, would be attained by a
general curriculum that would be required of all students. While historically there
may have been a point at which both vocational and college preparatory tracks
could have existed in one high school, this dual track is no longer easily

attainable, given the current movement towards standardization (Wraga, 1994;
Colorado Association of School Boards, 2006).
However, if rigorous curricula provide skills necessary for both college
and vocational fields, the need for differentiation may not be as profound.
Theoretically, by taking these courses, the students will be better prepared and, in
general, score higher on college entrance exams (ACT, 2006b). Many of the
practical courses that are flooding high-school offerings will be unnecessary
and would be eliminated as a result of their inability to act as a college or
vocational preparatory courses. The author reviewed transcripts of recent high-
school graduates enrolling in one liberal-arts college and found that more than
65% of entering freshmen were lacking in college preparatory coursework when
compared to the minimum HEAR requirements. Inordinate numbers of students
had filled their high-school schedules with non-academic coursework. In one
instance, for example, a student brought in only one math course in four years of
high school. In addition, courses that were taken instead of math were not
academic or vocational in nature. Multiple sections of physical education, teen-
living, keyboarding, health, and state-testing preparation courses were selected by
many students. Not only were students who did come to college not prepared for
college level work as indicated by more than 54% of the incoming students being
assessed as needing remedial-level courses (Gianneschi, 2006), but these same
students did not acquire vocational trade skills either.

When examining the literature on high-school course-taking patterns and
college success, many researchers have agreed that math coursework has a
significant impact on college attendance and successive achievement (Pelavin &
Kane, 1990; Adelman, 2006, Rose & Betts, 2001; Brandon, 2005). Findings from
the various studies suggested that, once a student has completed a certain level of
math education (ranging from geometry to trigonometry), the chances of
succeeding in college increase extraordinarily, regardless of race, socioeconomic
status, or gender. These studies support restricting curriculum, particularly for
those students planning on attending college. If key coursework like Algebra II
can be identified, and if 90% of students indicate a desire to attend college
(Conley, 2005; Stanford Bridge Project, 2002), why would these courses not be
mandated for high-school graduation? These studies suggest that high schools can
work to close gaps in college attendance and success by increasing requirements
for graduation from high school.
One of the more disturbing responses regarding the HEAR came from one
of Colorados school districts. The district indicated that its schools cannot deliver
the necessary coursework (CCHE, 2006). In fact, legislation was introduced into
the Colorado House of Representatives in January 2006, which subsequently
failed, to remove the requirements implemented by CCHE (Colorado Senate Bill
06-026, 2006). How have the needs of their advanced students been

accommodated if schools have failed to offer the suggested minimum number of
courses necessary to succeed in college?
In its examination and review of the admission standards in 2003, CCHE
found that, although high-school GPAs and entrance-exam scores had been
identified as primary measures used by many admission offices (Conley, 1995),
course-taking patterns explained most of the variance for Colorado students when
predicting first-semester college GPAs. However, comprehensive review of
transcripts is much more time consuming and staff intensive than examining a
single GPA or entrance exam score. Is the added effort necessary for review of
transcripts for CCHE purposes worth the additional costs accrued in staff time?
The question is valid considering one Colorado institution found a strong
relationship between high-school coursework and scores on the Colorado index,
the other measure mandated by CCHE (Brandon, 2005).
A substantial body of literature exists to inform the study concerning the
structures of high-school curriculum, both from a practical and theoretical
perspective (Wechsler, 2001; Rose & Betts, 2001; Adelman, 2006; National
Education Association, 1894; Hammack, 2004; Wraga, 1994). As the high-school
educational system in this country evolved into a dichotomous system during the
early part of the last century (providing vocational training for some and college
preparation for others), it became increasingly more complicated to align K-12
with higher education (Dewey, 1906/2002). The comprehensive high school that

resulted from a desire to provide education to the masses is now challenged again.
The current economy demands a more standardized education for its workers than
ever before (Camevale & Desrochers, 2003; Kendall & Williams, 2004).
Additionally, more students than ever acknowledge a desire to achieve a college
education (Conley, 2005).

3. Methodology
An examination of the current literature revealed a reliance on statistical
models in understanding which factors have been identified as important in
student success in college (Tinto, 2003; Braxton, Hirschy, & McClendon, 2004;
Bean & Eaton, 2000; Baird, 2000). Absent from this research have been in-depth
analyses that holistically examine the assumptions resulting from years worth of
studies. The research questions in this study called for a broad validation of the
current model, as well as a deeper understanding of resultant issues; therefore, a
mixed-method approach was employed (Kroc, 2007). The purpose of this research
was to analyze the implementation of the upcoming CCHE admission policy and
develop answers to the following research questions:
1. To what extent do the components in the CCHE admission policy
(HEAR) predict first-semester college GPA for students at one less selective
liberal-arts institution?
2. Which components are the best predictors?
3. Why do some students defy the prediction model by performing at a
much higher level than is predicted?
Examples of sub-questions for Research Question 3 include:
1. Do these students have family support structures that assist them?

2. Does the college provide support structures that have proven effective
for these students?
3. Do these students have role models in their school who are
demographically similar to them?
4. Have these students become integrated socially and/or academically at
the college?
5. How do the students perceive their high-school experience?
Based on the perspectives presented in Chapter 2, this chapter explains the
two-phased research method used to generate the data for this study. The first part
of the chapter provides an overview, describing the necessity for using a mixed-
method, sequential, quantitative-to-qualitative design. This is followed by details
of the procedures for the quantitative phase (Phase I), including the identification
of the population, the method of data collection, a discussion of the predicted and
predictor variables, and a description of the data analysis. Subsequently, details
about the procedure for the qualitative phase (Phase II) will include the
identification of the sample and the multiple methods of data collection. A
description of the data analysis for this phase will follow. Limitations of the study
will be discussed in Chapter 6.
3.1 Design
In order to analyze the validity of the components of the CCHE Admission
Policy and understand how they have affected students, the study employed a

sequential, mixed-method design (Creswell, 2003; Greene & Caracelli, 1997).
The mixed design was necessitated by the fact that the policy was built using a
statistical model, and to reproduce the effects and validate the components, the
model needed to be replicated. The quantitative phase addressed the relationship
of the models variables with subsequent success in college. It assisted in
establishing the extent to which the policy predicted success at the college, as well
as offering insight into which aspects of the model were the best predictors. The
model also served to identify students who succeeded in college despite having
been predicted to fail. In the second phase, qualitative interviews were used to
explore significant deviations in the data. The phenomenon of students
succeeding in college in spite of the model predicting failure was examined for
first-generation African American male students at one less selective 4-year
public institution.
3.1.1 Phase I Quantitative Analysis
Initially, a quantitative path analysis model was constructed using the
components of the Colorado admission policy (see Figure 3.1) in order to
reconstruct the prediction of first semester college grade-point average (GPA) in
order to study the validity of the components of the Colorado HEAR. The model
reflected the construction of the CCHE Admission Policy.

Figure 3.1. Path Diagram of Colorado Admission Policy
The primary dependent variable for the path model was the first-semester
college GPA. The college GPA was calculated on a traditional scale from 0.00 to
4.00. Secondary dependent variables consisted of the two individual indices as
described in the Technical Appendix to the Admission Standards Policy (CCHE,
2003). The high-school performance index and the college entrance-examination
index were standardized, each with a mean of 50 and a standard deviation of 10.
The high-school performance index was based on a students high-school grade-
point average or high-school class rank, whichever was higher once indexed. The
college entrance-examination index was based on a students ACT Composite
score or a sum of SAT Verbal and SAT Math scores, whichever was higher once
indexed. The independent variables consisted of the number of courses in each of

the academic areas defined by the HEAR (English, mathematics, social science,
and natural science) that students took and passed while in high school. Subjects and Sampling Phase I
The sample for Phase I included the entire population of first-time,
baccalaureate-seeking, entering college students who graduated from high school
in the spring of 2005 and who entered into one liberal-arts college in the fall of
2005. The initial population consisted of 783 students. 19 students were removed
due to missing data, leaving the analysis population at 764 students. As all of the
data for the admission requirements had already been collected by the college, no
incentives for participation were necessary. Data were extracted from the student
information system of the college. Setting and Materials Phase I
All research for this study was conducted at one less selective liberal-arts
institution in Colorado. Because this type of institution was identified by CCHE
as most affected by the implementation of the HEAR (Lestina, 2006), it became
critical that policy-makers understand the potential adverse impact on the
diversity of the institution. Data were subsequently obtained from the colleges
student unit database for purposes of analyzing the effects of the HEAR. This data
extraction was accomplished through the use of ODBC connections into SPSS
and AMOS software.
59 Independent variables Phase I
The independent variables in the path analysis included the number of
courses in each of the four academic areas identified by HEAR that were taken
and passed by each student during high school. For English, these included the
number of courses taken by a student in English composition, grammar, and/or
literature. Courses addressing such topics as journalism and yearbook were not
considered viable by CCHE for this category (CCHE, 2006). In mathematics, all
coursework counting towards mathematics HEAR credit was at or above the
Algebra I level. Courses counting towards the science component of the HEAR
were standards based and did not include offerings such as general science or
outdoor science (CCHE, 2006). Social science courses were also aligned with
standards and did not include such courses as those addressing teen living and
consumer education.
Coding structures were constructed prior to the evaluation of transcripts,
and the subsequent data entry into the institutional student information system
allowed easy extraction from the database. All courses were coded for all
students, even those who were not in the specific academic subject areas. In
addition to the four HEAR areas, additional coursework was coded into academic
elective or supplemental categories. Courses in the fine arts, business, journalism,
and other academic areas were coded into the elective category. The coding for
the supplemental category included courses such as teen living, physical

education, test-taking strategies, and other work not identified as academic in
nature. The complete list of coding for courses is contained in Appendix A. Dependent variables Phase I
The primary dependent variable for the path analysis was the first-
semester college GPA. The GPA was derived by summing the number of quality
points earned by the student: four for each hour with a grade of A, three for each
with a B, two for each credit hour with a C, one for each credit hour with a D, and
zero for each credit hour with an F. The quality points were then divided by the
total number of credit hours attempted. The resultant GPA was thus calculated on
a 0.00 to 4.00 scale.
The secondary dependent variables included the components of the
Colorado admission index. The high-school performance index was calculated by
using CCHEs converted high-school GPA index with a mean of 50 and a
standard deviation of 10. For students whose high schools did not provide GPAs,
equated high-school GPAs from transcripted class ranks were used, as instructed
by the Technical Appendix to the Admission Standards Policy (CCHE, 2003).
The college entrance-examination index was calculated using an equated index
score, depending on whether the student provided ACT or SAT scores to the
college. In cases in which the student provided both sets of scores, the higher
index score was used. As was done with the high-school performance index, the
ACT and/or SAT were standardized on a t-scale with a mean of 50 and a standard

deviation of 10 (CCHE, 2003). Data for the dependent variables were obtained
from the colleges student unit database. Data Collection Procedures Phase I
Data were extracted from the student unit database of the liberal-arts
college using SPSS ODBC connections directly to the data. Matching of data
across tables was accomplished using the student information-system database
unique identifier. AMOS structural-equation modeling software was utilized to
calculate the statistics for the path analysis. Because the data were previously
existing data, the issues of instructions to subjects and informed consent did not
exist during this phase of the research. Data Analysis Procedures Phase I
For Phase I, because replication of the CCHE model was important,
demographic data were used only to report descriptive data and were not included
while building the predictive model. Theoretically, the path analysis was
constructed so that courses predicted the high-school performance and college
entrance-examination indices, and the indices, in turn, would predict first
semester college GPA. Data were extracted from the student unit database of the
less selective, liberal-arts college using SPSS ODBC connections directly to the
data. Fall 2005 enrollment data were used for students who had graduated high
school in the spring of 2005, who were attending college for the first time in the
fall of 2005, and who were baccalaureate students. Data for compilation of the

CCHE index were also matched into the dataset. Data points included students
ACT and SAT test scores, high-school grade-point averages, and high-school
class ranks. College entrance-examination indices and high-school performance
indices were calculated as dictated by the CCHE Admission Policy Technical
Appendix (2003).
High-school transcripts were coded by subject using the codebook in
Appendix A. Once all coursework for the class of 2005 was coded, the data were
entered into the database by college admission staff. Upon completion of data
entry, data were extracted and summed by discipline to determine HEAR subject
counts. Once the data were compiled, AMOS structural-equation modeling
software was utilized in calculating the statistics for the path analysis in an
attempt to replicate CCHE analysis.
Building the model in this manner served to address the first two research
1. To what extent do the components in the CCHE admission policy
predict first-semester college GPA for students at one less selective liberal-arts
2. Which components are the best predictors?
The path coefficients denoted, for the liberal-arts college in the study, the
extent of prediction using solely the CCHE model by explaining a proportion of
the variance and identifying the variance explained. Also, for the liberal-arts

college in the study, the path coefficients identified which components were the
best predictors.
Assumptions involved in the use of path analysis were also tested. As path
analysis techniques are extremely sensitive to model specification, various direct
and indirect causal paths were compared after the initial model had been fit
(Kline, 1998; Tabachnick & Fidell, 2006). Additionally, the same sample is
required for all of the regressions in the path analysis. As with regression, missing
data do present a problem with structural-equation modeling, including path
analysis (Tabachnick & Fidell, 2006). For purposes of this analysis, 19 cases with
missing data were removed, not imputed (Muthen, Kaplan, & Hollis, 1987).
Violation of certain assumptions can lead to a misrepresentation of effect
sizes (Cohen, 1988), leaving the interpretation of results problematic. Path
analysis assumes normality, so normal probability plots were examined which
also served to assess the normality of residuals (Chambers, Cleveland, & Tukey,
1983; Sims-Knight, 2004). Also, as with linear regression, the relationships
between variables must be linear. Curvilinear relationships can misrepresent the
strength of independent variables (Osbome & Waters, 2002). These relationships
were assessed through the use of an ANOVA test of linearity (Tabachnick &
Fidell, 2006). Multicollinearity was also assessed using tolerance statistics. For
this study, examination of this assumption was important due to the fact that
multicollinearity causes individual variable effects to be underestimated (Cohen,

1988) . Independence of variables and the independence of prediction errors were
assessed using the Durbin-Watson d-coefficient (Durbin & Watson, 1950). Next,
scatterplots of regression predictions versus residuals were examined in order to
determine whether the errors of prediction and the independent variables were
uncorrelated (Sims-Knight, 2004). Additionally, for this technique, the sample
was examined to determine whether it was large enough to assess significance,
with the number of cases needing to be 10 times greater than the number of
parameters in the model (Kline, 1998; Tabachnick & Fidell, 2006).
By examining model fit, one can examine whether residuals are
uncorrelated with any of the variables other than the one caused by the residuals
(Tabachnick & Fidell, 2006). Model fit was assessed in a number of ways. By
comparing the relationship between the original and the reproduced correlation
matrices, one can test the model for significance (Specht, 1975; Schumacker &
Lomax, 1996). Fit was initially tested by calculating a chi-square statistic. Models
where chi-square is significant indicate that the model does not fit (Schumacker &
Lomax, 1996). As larger sample sizes can influence the chi-square statistic, a
relative chi-square statistic was used to adjust for the size of the population
(Wheaton, Muthen, Alwin, & Summers, 1977).
Additionally, baseline comparison measures of fit were examined (Bollen,
1989) . Comparative Fit Index (CFI) was analyzed to address the issues associated
with sample size. The CFI is another calculation where interpretation assesses the

level of fit and specifies (a) less than 0.85 indicates an unacceptable fit, (b) 0.85
through 0.89 is a mediocre fit, (c) 0.90 through 0.94 is an acceptable fit, (d) 0.95
through 0.99 indicates a close fit, and (e) 1.00 is an exact fit (Hu & Bentler,
1995). Supporting the baseline comparison, other similar fit measuresthe
Tucker-Lewis index (TLI) and Normed Fit Index (NFI) were also examined
(Schumaker & Lomax, 1996).
3.1.2 Phase II Qualitative Analysis
Following the completion of the quantitative phase, a phenomenological
study was conducted in order to develop a better understanding of any adverse
impact the policy may have on African American male college students. To
identify students that may have been adversely affected if the policy had already
been in effect, results of the path analysis were used to identify college students
whom the model predicted would be placed on probationary status in their first
semester (GPA below 1.70), who had not completed the requisite high-school
coursework of the HEAR, yet who had surpassed the prediction and were enrolled
in their second spring semester as sophomores in good standing. Subjects and Sampling Phase II
The sample for Phase II was selected purposively (Patton, 1990). Subjects
were chosen based on their status as students who were predicted by the CCHE
model to fail yet had been successful through their fourth college semester.
Archived records were examined to attempt to discover similarities between the

cases and further narrow the sample. Records researched included (a) student
demographic information, (b) high-school demographic information for the high
school of record, (c) courses students had taken while at the college, (d)
instructors students had had at the college, (e) advisors of record at the college, (f)
tutoring services received from the college, (g) financial aid need and amounts
received, and (h) demographics of housing where students lived.
In order to narrow the sample further, students who identified themselves
as African American were selected. As stated in the Colorado Higher Education
Performance Contract Act (2004), the States interest in increasing minority
enrollment in colleges was mandated to be part of each Colorado public
institutions performance contract. Examining the demographics of the sample,
the study focused solely on male students. Because one-third of this demographic
can expect to be jailed, imprisoned, paroled, or on probation (Weathersbee, 2006),
understanding the reasons these students were able to remain in college is critical.
Once identified, students were contacted via letter to inform them of their
selection and to ask them for their consent to participate. Students were informed
of the purpose of this study, asked to sign consent forms, and paid monetary
compensation in the amount of $20 for each interview session in which they
67 Setting and Materials Phase II
The second phase of the study was conducted at the same less selective
liberal-arts institution in Colorado. Data were obtained from the purposive sample
of students via interviews. All research was conducted during the last three weeks
of the students second spring semester at the institution being studied. Interviews
were conducted in the early afternoon in the library because it is a place where
both students and staff regularly visit (Seidman, 1998). Selection of Interviewees
Using the predictive equations developed in the path model, students were
chosen based on the results of Predictive Equation 3, where first-semester college
GPA = b3i high-school index + b32College entrance-examination performance
index + e3. The total population of African American men in the sample was 18.
Those student for whom their GPA was predicted to be 1.70 or below, the actual
first-semester GPA was 2.00 or higher, and the student continued to be enrolled at
the college were examined as part of the pool. Nine men remained in the pool.
This pool was further restricted by examining transcripts. Students who did not
meet the curricular requirements of the HEAR were chosen to continue in the
selection pool with none being eliminated. The final restriction was that the
students were still enrolled and in good standing in their second spring semester,
reducing the African American male population to five students in the interview

Once identified, these students were contacted via letter to invite their
participation in the study. Two of the students did not respond to the letter. Three
students agreed to assist, and each took a pseudonym. Interviews with Kyle,
Rahim, and Jamal began at the end of spring semester, 2007. Each had been
enrolled at the college for two years. Kyle and Rahim completed the interview
process, but Jamal chose not to return for his third year, so completed only the
initial interview. Data Collection Procedures Phase II
Students were interviewed multiple times in a tiered process (Seidman,
1998; Merriam, 1998; Yin, 2003). The interviews were driven by the participants
responses as well as the interviewers understanding of those responses (Seidman,
1998). The questions focused on factors identified as important in the success of
the student and included questions on student-support systems, student integration
in the classroom and courses, student social integration, race and gender role
models, characteristics of the students high school, and the students perception
of motivating factors (Tinto, 1993; Braxton et al., 2004; Bean & Eaton, 2000;
Baird, 2000).
The phenomenological structure of the interviews began by following
Seidmans (1998) three-interview series, with each interview lasting between 40
and 60 minutes. Initially, students were asked a series of open-ended questions
designed to make the students comfortable and to begin to determine the factors

to which the students contributed their success. Students were specifically asked
how they decided to attend the institution, with the focus on reconstructing past
events that had led them to their current situations (Seidman, 1998). Questions
asked of all students in the initial interview included:
1. Tell me about yourself.
2. Tell me about the process of how you decided to attend this college.
3. How do you feel about your time at this college?
4. Describe how your high-school experience helped or did not help your
ability to succeed at this college.
5. Tell me about your family while you were growing up.
During each of the first interviews, branching questions were also asked that were
not necessarily the same across participants in order to solicit responses.
Examples of the exploratory prompts used included:
1. Can you tell me more about that?
2. How did you feel about that?
3. What was that experience like?
The participants were then thanked for their responses and scheduled for the
following week. Once compiled, the data were evaluated to determine where
further information was necessary. Subjects to be explored further were identified,
whether they were expressed in this first session or not.

The second interview, also following Seidmans (1998) model, was
conducted one week after the first session and focused on expanding the details of
each participants first interview. The goal of this stage of the interviews was to
reconstruct the students experience. The questions posed in these sessions were
more tailored to the individual participants and were structured to build on
responses from the first interview: for example, Last time, you said [quote from
first interview]. Can you reconstruct that experience for me, walk me through it?
Additionally, where information appeared to be absent, such as no mention of a
father, questions were directed to fill in the gaps: Last time, we didnt talk about
the impact your father had on your youth experiences. Can you tell me about that
now? At the conclusion of the interview, the subjects were again thanked for
their time, and the next interview was scheduled. Afterward, responses were
evaluated and initial themes were developed through the use of chunking.
Sentences, phrases, and words were chunked into categories and evaluated for
initial themes for the interviewees to evaluate for accuracy. These themes were
the guide for the third interview.
Finally, the third interview was conducted one week after the second
interview to elicit the participants understanding of the meaning of their
experiences (Seidman, 1998). Two overarching themes were considered. Factors
contributing to school success and educational barriers were both consistent
for the three students. Within these themes, a series of sub-themes were also

explored to determine whether or not the interviewers interpretation of the
themes could describe the students understanding of their experiences. Each
participant was asked to explore further the interpretations of his experience,
based on what he had already outlined. The researcher had put the responses into
a timeline of experiences, based on the first two interviews. The timeline was laid
out for each interviewee, and each was asked to fill in gaps and to state which
entries were significant and which were not important. The secondary school,
postsecondary, and family experiences were most rigorously targeted because
these were the areas in which the students barriers and success were most
identified in the first two interviews. Students were also asked to rate their overall
experiences and identify the greatest successes as well as barriers. An overview of
the HEAR was also given to the two remaining students during this interview, and
they were asked for their opinions regarding the standards, based on their
After the third interview was completed, the data were analyzed to
determine whether saturation had been attained. The interviewer used the
following questions to determine whether saturation was achieved. Were any new
ideas given during the third interview that needed further clarification? Were any
new lines of questioning apparent? What had the students not said when
transcripts were compared to the conceptual framework? The tapes from the three
interviews were transcribed, and notes taken during the interviews were reviewed

to determine whether new lines of questioning were deemed appropriate. After re-
reading the notes and transcriptions multiple times, the interviewer found no new
lines of questioning were apparent. Data Analysis Procedures Phase II
The analysis of the phenomenological inquiry was designed to answer the
following research question and sub-questions:
1. Why do some students defy the prediction model by performing at a
much higher level than is predicted?
2. Do these students have family support structures that assist them?
3. Does the college provide support structures that have proven effective
for these students?
4. Do these students have role models in their school who are
demographically similar to them?
5. Have these students become integrated socially and/or academically at
the college?
6. How do the students perceive their high-school experience?
The results of the study were analyzed in an attempt to answer these
In order to triangulate methodologically the analyses, multiple analytical
methods were employed (Leech & Owuegbuzie, 2007). First, the data from the
interviews were coded using a grounded-theory approach as advocated by Strauss

and Corbin (1998) in a continuation of the initial coding for the third interview.
The tapes were first compared to the transcription, and the notes and the
transcribed data were evaluated. Materials were then read multiple times in order
to insure accuracy of transcription.
After completing the data collection and transcription, the researcher
convened a research team to assist with data analysis. The team was comprised of
the researcher, one advanced graduate student, and one additional professional
researcher whose background is in assessment. All members of the research team
were experienced conducting qualitative research. Transcripts from the interviews
were analyzed using a sequential coding process (Glaser & Strauss, 1967; Strauss
& Corbin, 1998).
To begin this stage of the coding, each person coded the data
independently, asked to chunk responses into categories that would be considered
for themes within the overarching structure. Once completed, the two overarching
themes used to construct the third interviewfactors contributing to school
success and educational barriers were introduced to the team for
consideration. Coders were also asked to determine whether the overarching
themes seemed logical or whether other overarching themes seemed more
appropriate, given the transcripts. Once finished, the team reconvened to review
the authors coding structure. Distinct categories resulted from the analysis, and

emergent themes pertaining to the phenomena were modified from the authors
original conceptions. Constant-Comparison Analysis and Emergent Themes
The team worked together until ideas could be developed into distinct
categories and molded into emergent themes regarding the phenomena. The
results of this coding were then used to perform a constant-comparison analysis.
Initially, data were coded using an inductive approach, to name phenomena seen
in the data. Phrases were chunked into small, meaningful pieces and assigned a
label or code (Leech & Owuegbuzie, 2007; Creswell, 2003). In developing the
coding scheme, the same steps were followed for each code (Weber, 1990). First,
the recording unit was determined, whether the unit was a word, a phrase, or a
sentence. From these units, general categories were developed, and the data were
subsequently coded. The team had also been asked to determine whether any
success factors or barriers appeared to be missing. Once completely coded, the
researcher consulted with two other coders until agreement was reached. Codes
were modified as necessary, and the coding scheme was restructured three times.
Once the descriptive codes were generated, a smaller number of emergent themes
were generated from the codes (Creswell, 2003). The themes were analyzed
across and between the cases in order to shape the information into a
phenomenological description, which served to address how and why these men

continued to succeed in college in spite of the States admission model (Creswell,
Once this process was complete, eight coding schemes were defined from
the interviews: (a) positive pre-collegiate experiences, (b) negative pre-college
experiences, (c) youth environment, (d) math deficiencies, (e) college
connections, (f) positive college experiences, (g) negative college experiences,
and (h) resiliency/agency. Within several of the schemes, categories needed
further delineation. For positive pre-collegiate academic experiences, the data
were coded to represent experiences in sports and experiences with teacher
interaction. Additionally, negative pre-collegiate experiences were divided into
lack of guidance and experiences with racism. Positive college experiences fell
into two different categories: classes and financial aid. Next, negative college
experiences were divided into six groupings that included (a) academics, (b)
faculty, (c) racism, (d) stress, and (e) student services. The students youth
environments were also separated into the parental aspect and the neighborhood
environment. Math deficiencies, college connections, and agency and resilience
were not split into any further categories.
Resultant themes consisted of college connections or meaningful
relationships in college, lack of guidance in high school, issues with mathematics
at all levels of schooling, resiliency or agency that allowed these students to
persist in college, and single-parent homes with little ability to support academics.

The team came to consensus on these themes (Miles & Huberman, 1994)
although the overarching themes were slightly modified as a result, becoming
barriers in education as opposed to educational barriers and success in
higher education instead of factors contributing to school success.
Member checking was done with the interviewees once the final coding
was established. Jamal did not choose to participate in this phase of the research,
but Kyle and Rahim both took part in this phase. Using the transcripts, the
categories and the emergent themes, the researcher went through the results with
each of the two remaining participants to determine whether the researchers
interpretations of the students voices were consistent with their stories. Prompts
were given by the researcher. Examples of prompts in this phase included:
1. Is this what you meant when you said___________?
2. During the interview, it came across to me that_________was a major
barrier in your educational career. How accurate is that statement?
3. When you spoke about_____________, it appeared to be a significant
contributor to your current success in college. How accurate is that statement?
Each member who chose to participate in the member checking stated his
satisfaction with the researchers interpretation of his interviews. Affirming the
interpretations given to him, Kyles response when finished with this process was
powerful: Maybe someone would actually listen to me so it can stop. Each man
expressed a sense of relief that his story was being told.
77 Content Analysis
In addition to the constant-comparison analysis, the information obtained
from the coded responses was methodologically triangulated using the codes to
perform a content analysis. Performing a content analysis strengthened the
analysis by allowing the investigator to determine those codes that were most
used in the responses (Leech & Owuegbuzie, 2007). Codes that were most used
correlated with the categories and emergent themes derived in the constant-
comparison analysis. Codes were loaded into SPSS to generate counts and
proportions in order to determine which codes were most frequently used. Once
generated, the counts were compared to the emergent themes for comparison.
3.2 Summary of Methodology
The mixed methodology used in this study was critical in developing a
comprehensive understanding of the impacts of the CCHE HEAR implemented
with the high-school graduating class of 2008 (Carnahan et al., 2007). The
quantitative phase of the study was necessary in order to define the policy in the
same terms as it was developed and to ensure that, at the very least, the model did
predict what it purported. The same type of regression analysis used for the
implementation of the policy was employed.
In order to understand the impact of this policy fully, quantitative analysis
falls short. By defining students who challenged the quantitative model and
overcame it, the impact of the implementation of the HEAR could then be fully

assessed. Determining what made these students successful despite the predicted
outcome of failure could only be accomplished by using a methodology that
explored the factors affecting these students more deeply. The sequential,
quantitative-to-qualitative methods used in this study were a good fit for the
research questions.

4. Results Phase I
Phase I of the research was designed specifically to answer the following
research questions:
1. To what extent do the components in the CCHE admission policy
(HEAR) predict first-semester college GPA for students at one less selective
liberal-arts institution?
2. Which components are the best predictors?
Through reconstruction of a path analysis, an understanding of the significance of
these factors in predicting success was assessed; and the model developed by
CCHE staff validated by data from one less selective liberal arts college.
4.1 Results
For these two research questions, the statistical findings of the model did
indicate that the premise of the HEAR (see Figure 4.1) significantly predicted
first-semester college GPA; although not all of the individual components of the
HEAR appeared to add to the model significantly in the manner in which the
model was built. In determining the overall significance, descriptives, fit, effect
sizes, probability values, assumptions and statistical power were all examined.

Figure 4.1. Path Diagram of Colorado Admission Policy
Descriptives including sample size, minimum values, maximum values,
means and standard deviations for each of the variables contained in the model
are shown in Table 4.1. As this was a path model, degrees of freedom for the
model were calculated by subtracting the number of distinct parameters estimated
from the number of distinct sample moments (Arbuckle & Wothke, 1999). This
calculation yielded 4 degrees of freedom subtracting 31 estimated parameters
from 35 distinct sample moments.

Table 4.1
Descriptives of Model Variables
Variable N Min. Max. Mean SD
English Hours 764 1.0 5.0 3.85 0.853
Mathematics Hours 764 0.0 8.0 2.94 1.252
Social Science Hours 764 0.5 8.5 3.70 1.027
Natural Science Hours 764 0.0 7.0 3.07 1.011
HS Performance Index 764 22.0 64.0 47.38 9.200
Exam Performance Index 764 30.0 71.0 49.21 7.294
First Semester GPA 764 0.0 4.0 2.19 1.136
Examining the overall fit of the model, the construction of the HEAR
model as developed by CCHE was found to have good fit across a number of
different fit indices. Without a solid consensus on which fit index is the most
reliable, Bollen (1990) recommends reporting on a number of different indices.
The first fit index examined was the chi-square goodness of fit (CMIN). As chi-
square can be sensitive to sample size, Wheaton, Muthen, Alwin & Summers
(1977) suggested examination of a relative chi-square (X^df) where a ratio of five
or less indicates an acceptable level of fit. Under this fit index, the HEAR model
did appear to have acceptable fit X^df = 3.414 (Table 4.2). Other fit indices are
based on comparisons to baseline models and are also shown in Table 4.2. The

purpose of these indices is to compare the model to a worst-case scenario, or a
null model (Arbuckle & Wothke, 1999; Schumaker & Lomax, 1996). For these
indices, the results follow: (a) The Normed Fit Index (NFI) = .988; (b)
Comparative Fit Index (CFI) = .992; and (c) Tucker-Lewis Index (TLI) = .955;
thus again suggesting the model had acceptable fit. For these indices, values of .9
or above indicate good model fit (Schumaker & Lomax, 1996). For each fit index
examined, the CCHE model, as constructed, appeared to have acceptable fit.
Table 4.2
Indices of Fit
Fit Type Fit Measure Level of Fitness
x2/df 3.414
Normed Fit Index (NFI) 0.988
Comparative Fit Index (CFI) 0.992
Tucker-Lewis Index (TLI) 0.955
The next step in the analysis of the data was to determine the significance
of the model, thus addressing the first question in this research. Regression
equations within the model were analyzed for statistical significance. Once the
significance for the overall model was determined, the individual components
were examined to focus on the second research question.

4.1.1 Significance of the Model
Across all examined measures of fit, the model of the HEAR appeared to
be a reasonable fit when applied to this population, and the prediction of first-
semester GPA was significant. The various equations in the model were then
analyzed to determine strength of the individual components of the CCHE
admission policy. This model was specified by the following path equations:
Equation One. High School Performance Index = bnmath courses +
b^English courses + b^ science courses + bi4 social science courses + bi5
college entrance examination performance index + ei
Equation Two. College Entrance Examination Performance Index =
b2imath courses + b22English courses + b23 science courses + b24 social
science courses + e2
Equation Three. First Semester GPA = b3ihigh school index +
b32College entrance examination performance index + e3
For each of these equations, b represented the regression coefficients and the
subscripts indicated the equation number and variable number (for example, b2i
represented the coefficient in equation two for variable one, which was the
number of math courses taken). Error was also represented in each of the three
equations by the letter e. Significance in each equation was measured by
examining R2 to determine the level of variance explained by the variables in each

equation. Initial regression analyses (sum of squares, degrees of freedom, mean
squares, F) are shown in Table 4.3.
Table 4.3
Initial Regression Analyses
Variable S.S. df M.S. F Sig (P)
Equation One 26,781.0 5 5,356.2 107.4 .000**
Residual 37,793.8 758 49.9
Equation Two 8,886.5 4 2,221.6 53.2 .000**
Residual 31,706.6 759 41.8
Equation Three 287.1 2 143.6 156.7 .000**
Residual 697.4 761 0.9
*p < .05; **p < .001 Equation One
For the first equation, R2 = .415. Thus, the number of courses taken within
the various subjects, along with student college entrance examination
performance explained 41.5% of the variance in predicting the students high
school performance index. Considering the regression weights for each of the
components (Table 4.4), each was positively statistically significant (p < .05) in
the prediction of the high school performance index. Discussion of the individual

components will be further expanded in the discussion of effect sizes. This
equation did significantly add to the model.
Table 4.4
Regression Weights for Equation One (High school performance index = bumath
courses + b^English courses + b^ science courses + bn social science courses +
b^ college entrance examination performance index + ei)
Variable Estimate S.E. C.R. 0 P
Math Hours 2.690' 0.245 10.990 0.366 .000**
English Hours 1.470 0.321 4.575 0.136 .000**
Science Hours 1.940 0.287 6.770 0.213 .000**
Social Science Hours 0.541 0.257 2.104 0.060 .035*
College Entrance Exam Index 0.176 0.040 4.463 0.140 .000**
Intercept 17.170 2.034
*p < .05; **p < .001 Equation Two
The second equation explained the least variance out of all of the
equations in the model. For this equation, R =.219. Thus, the number of courses
taken within the various subjects explained only 21.9% of the students college
entrance examination performance index. Investigating the regression weights for
each of the components in the college entrance examination performance index
(Table 4.5), each was positively statistically significant (p < .05) except for the

number of social science courses (p = .177). Discussion of the individual
components will be further expanded in the discussion of effect sizes. Although
smaller than the other equations, this equation also added significantly to the
Table 4.5
Regression Weights for Equation Two (College Entrance Examination index =
b2imath courses + b22English courses + b23 science courses + b24 social science
courses + e2)
Variable Estimate C.R. S.E. |3 P
Math Hours 2.275 10.906 0.209 0.391 .000**
English Hours 1.470 2.586 0.293 0.089 .010*
Science Hours 1.940 1.966 0.262 0.045 .049*
Social Science Hours 0.541 1.349 0.235 0.071 .177
Intercept 36.849 1.300
*p < .05; **p < .001 Equation Three
The third equation predicted the first-semester college GPA. In this
equation, R2 = .292. The combination of student high school performance index
along with the college entrance examination performance therefore explained
29.2% of the variance in this equation. Examining the regression weights for each
of the components in the prediction of the first-semester GPA (Table 4.6), each

was positively statistically significant (p < .05). Discussion of the individual
components will be further expanded in the discussion of effect sizes.
Table 4.6
Regression Weights for Equation Three (First Semester College GPA = b3ihigh
school index + t^college entrance examination performance index +
Variable Estimate C.R. S.E. 0 P
High School Performance Index 0.057 13.790 0.004 0.458 .000**
College Entrance Examination Index 0.024 4.716 0.005 0.157 .000**
Intercept -1.691 0.254
*p < .05; **p < .001
4.2 Total Effect Size of Each Equation
In addition to the effect sizes of the specific moments, the effect size of
each regression equation (/) was calculated (Cohen, 1988; Cohen, 1992). For
multiple regression equations, the effect sizes were calculated based on the
following equation:
1 -R2
Once run, the effect sizes were compared to Cohens (1992) suggested levels of
size for/. The effect size for this calculation is considered large where/ >= .35,
medium where/ >= .15 and/ < .35 and small where r < .15. Results for equation