ONLINE INFORMATION SEEKING BEHAVIOR: MODELS OF INFORMATION SOURCE SELECTION AND INFORMATION SEEKER SATISFACTION by LI LING HSU B.S., Oklahoma City University, 2004 M.B.A., Oklahoma City University, 2005 A thesis submitted to the Faculty of the Gr aduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Computer Science and Information Systems 2014
201 4 LI LING HSU ALL RIGHTS RESERVED
ii Th is thesis for the Doctor of Philosophy degree by Li ling Hsu has been approved for the Computer Science and Information Systems by Michael Mannino Chair Zhiping Walter Advisor Judy Scott Takeshi Nishikawa Ilkyeun Ra July 1 8 th 2014
iii Hsu, Li li ng (Ph.D., Computer Science and Information Systems ) Online Information Seeking Behavior: Models of Information Source Selection and Information Seeker Satisfaction Thesis directed by Associate Professor Zhiping Walter ABSTRACT Search engines currently do minate the information seeking market. To be in a better position to compete with search engines, every content website needs to understand 1) the underlying reasons behind source selection and information seekers satisfaction and 2) what can be optimized to gain a competitive advantage. To provide this understanding, this dissertation developed an information source selection model in the market of local information seeking to investigate how different types of information sources were perceived differen tly and how users utilizing different sources were different. An online survey was conducted to test the information source selection model. Data analysis results showed that direct experience, involvement, intensity of information seeking, habit strength of using search engines, perceived search skill, perceived ease of use, and perceived trustworthiness were significant in discriminating the three information source types. The data also suggested that even though search engines dominated the market of l ocal information seeking, local information websites still have a chance to compete with search engines. Implementable suggestions and future directions were provided for local information websites. To further explore what determines information seeker sa tisfaction, in the second
iv statements. A grounded theory approach was employed to derive an information seeker satisfaction model using the IS success model as the framewor k. The data analysis results suggested that quality of user generated content is one of the key factors of information seeker satisfaction and that information topic has a moderating effect on the relationship between the quality and its antecedents. The results also showed that no matter which type of information people sought, they always want to know what other people have experienced. User generated content enriches content websites and influences information seeker satisfaction. Implementable sugge stions were provided for product, health, and local information websites. The contribution of this dissertation is (1) to contribute to academic research by proposing an information source selection model and by proposing an information seeker satisfaction local information seeking behavior, by identifying factors that impact information seekers satisfaction, and by providing implementable suggestions to the content websites. T he form and content of this abstract are approved. I recommend its publication. Approved: Zhiping Walter
v ACKNOWLEDGMENTS Foremost, I would like to e xpress my sincere gratitude to my advisor, Professor Zhiping Walter, for her support, encouragement, mentorship, and immense knowledge. During my doctoral study, Professor Walter appreciated my strengths and encouraged me to improve my weaker area. She h as been supportive since the day I entered the Ph.D. program. I am very grateful for her insightful advice on my personal and academic li f e. She also inspired me to push my limits and unleash my maximum potential Professor Walter is and will be my be st role model for a researcher and a mentor. Besides my advisor I would like to acknowledge the rest of my doctoral commit t e e members : Professor Michael Mannino, Professor Judy Scott, Professor Takeshi Nishikawa, and Professor Ilkyeun Ra, for their time in reading previous drafts of my dissertation and providing valuable input that have shaped up this dissertation. Last, but not least, I would like to thank my family for their constant support through the ups and downs of my doctoral study. My hard working parents H uan chih Hsu and Mei ying Fang, have sacrified their lives for me and my sisters and helped me become the first person in my family to earn a doctoral degree. Their confidence in me helped me achieved my educational goal. I also wanted to expre ss my appreciation to my two sisters Li wen Hsu and Li ping Hsu, and three nieces Po hsuan Liu chang, Tsen hsin Liu chang, and Lu hao Liu chang They are the ones who cheer me up and make me laugh when my life gets a little rough W ithout their uncondi tional love and infinite care I wo uld not be here writing this acknowledgment.
vi TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ................................ ....... 1 Research Mo tivations ................................ ................................ .............................. 1 Trend of Source Preference ................................ ................................ .............. 1 Trend in Local Information Seeking ................................ ................................ 2 Effect of User generated Content on Information Sources ............................... 4 Moderating Effect of Information Topic ................................ .......................... 4 Research Questi ons ................................ ................................ ................................ 6 II. LITERATURE REVIEW ................................ ................................ ............................ 7 Everyday Life Information Seeking ................................ ................................ ........ 7 Pr oblem Statement -Local Information Seeking ................................ ............ 8 Information Source Selection ................................ ................................ ................. 9 Work related Information ................................ ................................ ............... 10 Everyday Life Information ................................ ................................ ............. 11 Problem Statement -Source Selection on the Internet ................................ .. 13 Information Seeker Satisfaction ................................ ................................ ............ 17 User Satisfaction ................................ ................................ ............................. 19 Problem Statement ................................ ................................ .......................... 23 III. RESEARCH MODEL ................................ ................................ ............................... 26 Comprehensive Model of Information Seeking ................................ .................... 26 Model and Hypotheses Development ................................ ................................ ... 28 Antecedents ................................ ................................ ................................ ..... 29 Information Carrier Factors ................................ ................................ ............ 38 Control Variables ................................ ................................ ............................ 41
vii IV. METHODOLOGIES ................................ ................................ ................................ 44 Quantitative ................................ ................................ ................................ ........... 44 Data Collection Procedure ................................ ................................ .............. 44 Measures ................................ ................................ ................................ ......... 45 Data Analysis ................................ ................................ ................................ .. 48 Qualitative ................................ ................................ ................................ ............. 62 Data Collection Procedure ................................ ................................ .............. 62 Data Analysis ................................ ................................ ................................ .. 64 V. FINDINGS AND DISCUSSION ................................ ................................ .............. 85 Quantitative ................................ ................................ ................................ ........... 85 Qualitative ................................ ................................ ................................ ............. 89 VI. CONCLUSION ................................ ................................ ................................ .......... 94 Recommendation ................................ ................................ ................................ .. 95 Local Information Websites ................................ ................................ ............ 95 Content Websites ................................ ................................ ............................ 98 Implications ................................ ................................ ................................ ......... 100 Academic Implications ................................ ................................ ................. 100 Managerial Implications ................................ ................................ ............... 102 Limitations ................................ ................................ ................................ .......... 105 REFERENCES ................................ ................................ ................................ ............... 110 APPENDI X ................................ ................................ ................................ ..................... 121 A. Forum Lists ................................ ................................ ................................ .... 121 B. Survey Items ................................ ................................ ................................ ... 124 C. Factor Loadings ................................ ................................ .............................. 127 D. List of Website Names ................................ ................................ ................... 128
viii E. Steps of Coding with Examples ................................ ................................ ..... 129
ix LIST OF TABLES T able 4 .1 Characteristics of Survey Sample. ................................ ................................ ........... 49 4 2 Type of Informati on and Source/Provider. ................................ .............................. 49 4 .3 Information Topic by Gender and Source Type ................................ ..................... 50 4 4 Convergent Validity ................................ ................................ ................................ 53 4 5 Discriminant Validity ................................ ................................ .............................. 54 4 6 Overall Model Fit ................................ ................................ ................................ .... 55 4 7 Pseudo R Square ................................ ................................ ................................ ..... 55 4 8 Hit ratio ................................ ................................ ................................ ................... 56 4 9 Likelihood Ratio Tests. ................................ ................................ ............................ 5 6 4 .1 0 Parameter Esti mates ................................ ................................ ................................ 57 4 .1 1 Extracted Concepts of Quality of User generated Content ................................ ...... 72 5 .1 Hypotheses Summary ................................ ................................ ............................. 85
x LIST OF FIGURES Figure 2 .1 Updated D&M IS Success (DeLone and McLean, 2003). ................................ ......... 21 2 2 The Proposed Integrated Research Model (Wixom and Todd, 2005) ....................... 22 3 1 Comprehensive Model of Information Seeking (Johnson et al., 1995). .................... 2 6 3 2 Information Source Selection Model. ................................ ................................ ........ 29 4 1 Information Seeker Satisfaction Model. ................................ ................................ .... 67
1 CHAPTER I INTRODUCTION With the popularity of the Web and growing number of online resour ces, the World Wide Web has become a major information source. The Web is considered a practical tool for information seeking, especially for young adults who grew up with Internet access (Li and Kirkup, 2007). The dramatic evolution in technology taking place has changed information seeking behavior and increased the competition among online information sources. The number of content websites is increasing, with traditional print media (e.g., YP.com Consumer Reports ) as well as new entrants (e.g., Yelp WebMD CNet ). Content webs ites, in this dissertation ref er to the websites that provide inform ative content to those who have information needs These content websites are striving to provide a compelling information seeking experience. Similarly, ma ny search engines (e.g., Google or Bing) have enhanced their search capabilities to provide personalized and cookies Some search engines also integrate maps into search results, allowing inf ormation seekers to find information more efficiently. Research Motivations Trend of Source Preference Search engines currently dominate the information seeking market (Solis, 2009), and a majority of people hold a positive perception toward search engines (Purcell, Brenner, & Rainie, 2012) Below are the findings: 91% of search engine users say they always or most of the time find the information they are seeking when they use search engines
2 73% of search engine users say that most or all the information they find as they use search engines is accurate and trustworthy 66% of search engine users say search engines are a fair and unbiased source of information 55% of search engine users say that, in their experience, the quality of search results is getting better over time, while just 4% say it has gotten worse 52% of search engine users say search engine results have gotten more relevant and useful over time, while just 7% report that results have gotten less relevant (Purcell et al., 2012, p. 3) What do t hese data mean to content websites? Can content websites compete with search engines? If the answer is yes, what should these content websites (e.g., Yelp WebMD CNet ) do to attract more information seekers? To answer this question, content websites ne ed to understand perception with different types of information sources and the underlying reasons behind information seeker satisfaction. By knowing 1) the user characteristics and perception that influence the behavior of source sel ection and 2) the features that satisfy the needs of information seekers, content websites can be better optimized and can play a better role to attract search engine users. Trend in Local Information Seeking One of the six trends identified by 15miles, L ocaleze, and comScore (comScore, 2012) was: Local search, as a category, continues to get bigger with 144 percent growth since 2007. Today, consumers are increasingly using new methods to research and select businesses in the local marketplace. This requir es markers to expand their traditional marketing plans to include these sources or risk a reduction in the volume of sales leads fr om the fragmenting marketplace. (comScore, 2012, p. 4) It is clear that local information seekers are moving to digital alter natives, such as online and mobile devices as their information sources. One consequence of this trend is that local advertising revenue is expected to grow from $11.1 billion to $21.8 billion from
3 year 2011 to 2016 (BIA/Kelsey, 2012). Many have seen the migration of local information seeking from traditional print media to the online sources, and have wanted to capture the market of local inform ation seeking on the Internet. Despite the fact that industrial researchers and practitioners have observed the growth of local information seeking on the Internet and emphasized how important the trend is to local businesses, little is known concerning online local information seeking in academic research. In contract, the existing industrial studies (e.g., Sterl ing, 2010; Wuchenich, 2011) provide only descriptive statistics and tend to lack theoretical support. Furthermore, industrial researchers have found that Google is the dominant tool for local information, followed by Yelp, YP.com, and Citysearch (Graham, 2011). These information sources are very different in terms of the way local information is delivered. Unlike content websites (e.g., Yelp, YP.com, Citysearch) that offer detailed local information, search engines do not provide local information direct ly. Instead, search engines are information retrieval systems, which provide information seekers with a list of local information websites. But the questions of 1) how local information sources are perceived by information seekers and 2) what the factors are that can help local still unknown. Accordingly, in the first study, the author of this paper developed a source selection model to understand: (1) How are different types of local information sources perceived? and (2) What are the user characteristics and perception that can help understand the behavior of source selection? More specifically, the purpose of the first study was to answer the following questions: (1) What are the factors that can help to
4 distinguish users who utilize different types of information sources? and (2) Who likes to use which type of information source? The source selection model was developed in the first study with the use of three major local information types: entertainment, professional services, and store sales. The information source selection model, which was developed in the first study, bridges the gap between academic research and industrial studies and contributes to the informa tion seeking literature. Effect of User generated Content on Information Sources Information providers are no longer limited to websites; instead, user generated content is available on many content websites. For example, consumer reviews are the most imp ortant asset to business directory and review websites; as such, these websites implement different strategies to encourage consumers to write more reviews (Wang, 2011). However, how does user generated content affect information seeker satisfaction towar d content websites? The effect of user generated content on information seeker satisfaction is still unknown. By use of an inductive approach, the author of this paper investigated what factors lead to information seeker satisfaction, in the anticipation that the analysis of the data will show whether user generated content has an effect on information seeker satisfaction. If the effect is found to be evident, this might be the leverage that can be used by content websites to compete with search engines. Moderating Effect of Information Topic The current information seeking literature has identified several information source characteristics that people care about when they look for information. However, the focus of majority of these studies (e.g., Crou ch & Louviere, 2004; Fidel & Green,
5 2004; Kwasitsu, 2003; Rains, 2007; Tustin, 2010) was on a single type of information, such as health, travel, or work related information. A standardized guideline of website design helps a content website look professi onal, but it does not provide unique elements that information seekers desire. The effect of factors like credibility and trustworthiness were found to be significant to information seeking behavior only in regard to a particular information topic. For ex information seeking, and they found that engineers tend to look for more information when they believe the source possesses expertise. However, expertise may not be an important factor when people look for entertainment information. This motivated the need to identify the distinctive factors among different types of information topics. Therefore, the second study extended beyond the scope of local information seeking from the first study, and added two more topics: pre purchase product and health information seeking in order to compare and determine whether information topic has a moderating effect on the relationship between the quality and its antecedents. In the comparison of the three t ypes of information topics, the findings showed which factors are important across all types of information sources. The results give content websites a standardized guideline on how to design a professional content website. Most importantly, the compari son shows distinctive factors that are essential in different types of information topics. These are the key factors that distinguish popular websites from other information providers. In the second study, the author collected recommendation statements th at one Internet user made to another. From these statements, the author explored how
6 information seekers perceived a content website and why they recommended particular websites to others. This research then aggregated the qualitative findings and employ ed the grounded theory approach to derive a research model using the IS (information systems) success model (DeLone & McLean, 2003; Wixom & Todd, 2005) as the framework. The proposed model contributes to the literature of information seeking by identifyin g factors that determine information seeker satisfaction. It is anticipated that, based on the findings, it will be possible to identify whether user generated content is one of the key elements of information seeker satisfaction and determine whether inf ormation topic has a moderating effect on the relationship between the quality and its antecedents. Research Questions Q1: What are the factors that can help to distinguish users who utilize different types of local information sources? Q2: Who likes to us e which type of local information source? Q3: Who should different types of local information websites target? Q4: Which factors will impact information seeker satisfaction? Q5: Will user generated content be a factor on information seeker satisfaction? Q6 : Will the IS success model be sufficient to cover all factors?
7 CHAPTER II LITERATURE REVIEW Everyday Life Information Seeking Local information seeking falls under the literature on information seeking. According to this literature, information seeking can be categorized into two domains: work related and non work related. The term, Everyday Life Information (ELI), was introduced by Savolainen (1995) to synthesize information topics that are not related to work. This term refers to information that is (Savolainen, 1995, p. 267) Savolainen (1995) proposed a conceptual model of everyday life information seeking (ELIS) that incorporates motivators, which trigger the action of information seeking, and the factors, which affect the information seeking behavior. In the ELIS resources (e.g., contact networks) could influence how they seek and consume info rmation. Savolainen (1995) certain type of channel source is determined by habitus and culturally determined system of thinking, perception, and evaluation that is inte (Savolainen, 1995, p. 261 262) In short, how people seek and use information is a rather complicated process that involves social and individual factors. Through interviews conducted with 18 people in Finland, Savolainen and Ka ri (2004a) explored how the Internet was conceptualized as an information source. The categorized into two types: metaphorical and experiential. Savolainen and Kari believ ed
8 (Savolainen & Kari, 2004a, p. 223) ), their conceptions were influenced by the image given by media. However, when respondents described the Internet base d on use experience, the focus of their responses was primarily on stating the pros and cons of using the Internet as an information source. Savolainen and Kari concluded that, when multiple channels (e.g., library, friends, Internet) are available throug h which to obtain information, usually, information seekers do not spend much time on evaluating the costs and benefits of utilizing each channel. Instead, the decision of which channel to use is made straightforwardly based on earlier use experience (Sav olainen & Kari, 2004a) Rieh (2004) studied online information seeking behavior in a home environment. In her comparison of seeking everyday life information at home with seeking work related information at work, she found that individuals seek informatio n more frequently home, unlike in the workplace or at school, the Internet is the primary channel used because there are no experts available. Also, her finding sug gested that people usually start with a known website; they turn to search engines only if they cannot think of any websites or are unsure of what they are doing. Interestingly, people show strong loyalty to their favorite search engines or portals even w hen search engines are not their first choices (Rieh, 2004) Problem Statement -Local Information Seeking Evidence in the ELIS literature (Rieh, 2004; Savolainen, 1999; Savolainen & Kari, 2004b) indicates that the Internet is one of the most popular info rmation channels. However, prior research (Rieh, 2003) found that it is difficult for people to find local
9 information on the Internet. Local information is closely associated with people's everyday life, and it is an important element of ELIS. Everyone needs some local information almost every day. For instance, one might want to find out what the top local restaurants are, which businesses provide a better cleaning service, or what upcoming local events or festivals are available in town. Data from co mScore, a marketing research company, showed that there were more than 447 million local information searches on search engines in August 2005 (Rainie, 2005) The number of local information searches will continue to increase as needs continue to grow. Companies, such as Google, Yahoo, AOL, and Microsoft have realized the market potential for local information seeking. They have tried to capture this market by offering localized search services (e.g., Yahoo Local). If content websites want to become the dominant players in their respective local markets, it is crucial for them to understand how people look for local information. Most published ELIS studies were conducted by studying how people look for: (1) how to type of infor mation or medical information (e.g., Agosto & Hughes Hassell, 2006; Branch, 2003; McKenzie, 2003; Rieh, 2003) (2) school work related information (Given, 2002) (3) orienting information (e.g., news and weather that people need to check every day to keep them up to date on the current events) (Savolainen, 1995) (4) recreational information (Ernest, Level, & Culbertson, 2005) and (5) travel information (Chang, 2009) literature, there has been little attention paid to local information seeking behavior. This research project paper is an attempt to fill that gap.
10 Information Source Selection Work related Information Much of the research on information seeking in work related topics has been de voted to the study of the determinants of information source selection when multiple channels of information sources, particularly including human experts, are available. This body of the literature applied two major models, which consistently identified information source accessibility and quality as major determinants. The two major models applied are cost/benefit framework and principle of least effort. According to the cost/benefit framework, information seekers evaluate the cost and benefit of the i nformation sources, select information sources based on that analysis, and perceive benefit as the most important factor. In comparison, in the principle of least effort, it is asserted that information seekers select information sources, which are availa ble and require the least effort. Accordingly, the former model advocated that information quality (or content of information) is the most important factor in the determination of source selection (Hertzum, 2002; Woudstra & van den Hooff, 2008; Woudstra, van den Hooff, & Schouten, 2012) whereas the latter model advocated accessibility as the most important factor (Anderson, Glassman, McAfeea, & Pinelli, 2001; Gerstberger & Allen, 1968; Jorosi, 2006) Hardy (1982) compared the cost/benefit framework with t he principle of least effort and proposed a third model: information seekers do evaluate the cost and benefit, but they put more weight on the cost than benefit. Accordingly, accessibility is more important than the information quality. A similar finding was found in the Xu, Tan, and (2006) seeker source information need framework. In this framework, task
11 importance was introduced as a moderator variable, and the researchers examined how it moderated the impact of source quality and accessibility on personal source preference. The result showed that, when a task becomes more important, information seekers prefer someone who is closer to them, and they do not pay much attention to information quality. In the context of academic related information seeking, Kim and Sin (2011) noticed that undergraduate students tend to use less accurate but more accessible information sources, such as Web sources. They conducted a survey to explore and understand such behavior. In the selection of information sourc es, five selection criteria were claimed by the students as important: 1) accuracy, 2) accessibility, 3) ease of use, 4) financial cost (free information sources are preferred), and 5) currency. However, the characteristics of the frequently used sources were identified as: 1) accessibility, 2) financial cost, 3) familiarity, 4) ease of use, and 5) comprehensiveness. It was implied in criteria. Everyday Life Informa tion Although there is much literature on source selection for work related information, there are few studies on source selection for everyday life information. However, similar to the findings in work related information seeking, researchers in the fiel d of ELIS found that the two key criteria for source selection are information quality and accessibility. For example, Savolainen (1995) found that availability, accessibility, and ease of use were three important determinants when teachers and industry w orkers looked for everyday life information. In his later studies of source preference in the
12 context of problem specific information (Savolainen, 2008) and orientating information (Savolainen, 2007) Savolainen identified content of information and acces sibility as the two major criteria why information sources are selected. The content of information, in these two studies, is similar to the information quality in the literature of work related information source selection. It refers to information seek information quality (e.g., the human is preferred as an information source because they are able to provide experience based information). (2000) related and everyday life information seeki ng had similar findings; speed of access, ease of use, and value of information were the most important criteria of source selection. Similar to Julien and (2002) used personal digital assistants (PDAs) to track undergraduate students related information and everyday life information. The result showed that search engines and portals were used most frequently. Ease of use and availability were the most frequent reasons stud ents mentioned for their choice of the first source. However, quality was mentioned as frequently as the aforementioned reasons for choice of the second source. A few studies examined information source selection behavior, but the researchers did not iden tify the determinants or did not relate determinants to source selection. For example, Chatzoglou and Vraimaki (2010) suggested that information seekers of different ages, education, employment status (e.g., self employed, unemployed, students), and emplo yment sectors (or industries) perceive selection criteria differently. For instance, information seekers with higher educational levels perceive
13 trustworthiness as a more important factor than ease of understanding, while information seekers with lower ed ucational levels perceive the opposite. However, how selection criteria differ in each information source was not examined. Problem Statement -Source Selection on the Internet Despite the fact that the Internet has become one of the main information sou rces, the majority of the existing studies on source selection focus on multiple channels (e.g., people, library, Internet, magazines). Determinants for the selection of different information channels vary, depending on the characteristics of the channel. For example, humans are favored as an information source because they are able to provide experience based information, while the Internet is selected because it is easy to access and it allows information seekers to compare among multiple information so urces (Kim & Sin, 2011; Savolainen, 2008) The focus of the first study was on the source selection on the Internet, and the findings helped to deepen the understanding of how different types of sources were selected. In the context of information seeking on the Web, the author of this paper observed that searching can occur at two levels: 1) searching on the entire Web using search engines before arriving at a specific content website (cross site search) and 2) searching within a specific content website (in site search). The processes through which people obtain information are different for these two types of information sources, because one is an information provider and the other is a pass through to information providers. With browsing, an informati on seeker has to know which websites provide the type of local information he desires. For example, when an individual wants to find out about upcoming local events, he has to know what website provides local events
14 information. Suppose he is aware of Ev entful .com he might visit the website by typing in the Web address or selecting it from bookmarks. To find the needed information, he would then browse pages within the website or visit other websites through links from Eventful .com By contrast, search ing on search engines does not require knowing specific website addresses. By entering a few keywords, one can retrieve a list of relevant websites. He might then visit each of the websites from the search result to find the needed information. When a se arch engine is used to find local information, the quality of the search skill. If one uses inexact keywords, numerous irrelevant pages may be retrieved. The individual might need to visit several irrelevant websites before the needed information is acquired. Being a pass through to information providers, a search engine is not able to content w ebsite is used for local information seeking, the quality of the information seeking result heavily depends on the quality of the website; that is, how much information is provided and how the information is presented and arranged. Being an information pr ovider, a content website is able to offer exhaustive information that people want. Being able to provide sufficient information so that people can find what they need from a single website is an asset, which provides the website owner with a competitive advantage. For news provider, for example, being local repositories can help their websites attract more visitors. However, there are several existing information providers on the Web. New entrants, such as the news providers, need to understand how exi sting local information
15 sources are used to determine whether they still have an opportunity to succeed in the local information seeking market. In the comparison of content websites and search engines, the author of this research study was able to discove r what the most popular type of information sources were and what percentage of the local information seekers used search engines. Should new entrants in the local information market (e.g., news providers) worry about competing against search engines? Th e comparison provided a useful insight about websites to determine their chance of success. Accordingly, it is important to compare how search engines and content websites a re perceived. To date, the focus of existing studies of online information seeking (e.g., Aula, 2005; Rains, 2007; Tustin, 2010; Zhang, Fang, & Sheng, 2006 7) has been either on information searching on search engines or on information seeking on websites, and researchers neglected how different modes of online information seeking using websites vs. using search engines affect information search behavior or what factors affect the mode of information seeking. Since content websites, such as news providers or Citysearch, face stiff competition from search engines, understanding online information seeking behavior is not enough. More importantly, every content website needs to understand when and why people choose search engines instead of going straight to websites. Further, news providers also face competition from specialized single content websites (as explained below). Therefore, it is imperative that news providers understand different ways people obtain local information online.
16 Information provided on one website can differ from similar information provided on another in terms of breadth and depth. For example, baseball related information provided by the Major League Baseball website is more detailed and comprehensive than the information provided by ESPN. This is logical because a Major League Baseball website is a website that is dedicated to baseball whereas ESPN.com covers a broader range of sports. After reviewing content websites that provide local information, the author categorized them in to two groups: centralized websites and specialized websites. In the context of this paper, centralized websites are those that provide a broad range of local information. They serve as a local information repository that enables information seekers to f ind a variety of local information (e.g., outdoor activities, local restaurants, sports information) from a single provider. Specialized websites, in the context of this dissertation are those that provide only a specific type of local information. A we bsite that provides only wedding venue information, for example, is considered a specialized website. Because of their targeted nature, specialized websites usually offer more complete and detailed information than centralized websites for a given type of local information. Therefore, the author compared three types of information sources: specialized websites, centralized websites, and search engines. In the comparison of the three types of information sources, the author was able to identify key factors that people care about when they look for local information. The comparisons also provide important insights to news providers, who plan to be local repositories. Since news providers have been serving essentially as a centralized source for all local i nformation through print or their websites, their transition into providers of centralized local information databases would
17 be logical (Gray, 2008) Specialized websites, for example, those devoted to local music scenes, will have better appeal if they a re backed by a local music organization than by a local news provider. However, both of them can be bypassed if people always start their search with a search engine. If this research study reveals that people prefer one stop centralized databases for al l their local information needs, then centralized websites, such as news provider or Local.com, are in a great position to capture the local information seeking market. Otherwise, centralized websites may be facing an uphill battle trying to lure people a way from search engines or established specialized websites. In conclusion, the first study investigated which Web information sources were used and how they were perceived in the context of local information seeking. The author identified factors that in of information source, and proposed a n information source selection model. The first local information seeki ng behavior, the author is filling a gap in ELIS literature. Second, the author investigated information seeking behavior on centralized websites, specialized websites, and search engines to identify key elements that make one more popular than the other. Comparing different modes of using the same information channel has never been done before in ELIS literature. Third, the results concerning information source Information Seeker Satisfaction There has been relatively little research that has examined what constitutes information seeker satisfaction and what affects it. Much of this research has found system quality to be an important factor. For example, Santosa, Wei, and Chan (2005)
18 conducted an experiment to study information seeker satisfaction on product related information. They found that situation motivation (e.g., download speed and visual appearance) positively influenced information seeker satisfaction. Fu and Salven dy (2002) compared the difference between browsing and searching on shopping websites Inherent usability refers to system usability evaluation (e.g., efficiency) when or a fter users have used websites. Apparent usability, on the other hand, refers to system usability evaluation (e.g., design and layout) before users actually use websites. The results from the experiment suggested that, for websites, where searching is th e main task, the focus should be more on inherent usability, and the website should be designed Similar to the findings from Santosa et al. (2005) and Fu and Salvendy (2002) Al Maskari and Sanderson (2010), who examined the factors affecting information seeker satisfaction in information retrieval, suggested that system effectiveness is one of the most important factors that influence information seeker satisfaction. In the cont ext of online health information websites, Bliemel and Hassanein (2006) proposed a research model to study the factors that contribute to information seeker satisfaction. In this model, it was suggested that satisfaction with system quality plays a larger role in the information quality (a medium effect) and trust (a small but significant effect). Besides system quality, studies in the literature of information seeker satisfa ction also found user related factors to be significant in influencing information seeker satisfaction. For example, Santosa et al. (2005) found that user involvement positively
19 influenced information seeker satisfaction. User involvement, in Santosa et refers to the level of importance and personally relevant information that seekers feel when they interact with information sources. Al suggested that information seekers have a higher level of satisfactio n when 1) less time is required to find the needed information (user effectiveness) and 2) less effort is required to find the needed information (user effort). characteristics, namely I nternet training, frequency of use, and expectation of success, were related to user satisfaction. The results from the interviews showed that information seekers perceived fairly high levels of satisfaction no matter how frequently they used the Internet as the information source or whether they received Internet training or not. However, Bruce did find that information seekers with different levels for expectation of ion of success increased, their level of satisfaction increased as well. User Satisfaction Information seeker satisfaction falls into the domain of user satisfaction. User satisfaction is one of the major streams to understand, evaluate, and measure infor mation systems (IS) success, and it has been thoroughly investigated in the discipline of information systems. DeLone and McLean (1992) reviewed the literature of IS success and developed a taxonomy of IS success. In this taxonomy, six categories of IS s uccess were identified: system quality, information quality, use, user satisfaction, individual descriptive model of IS success was drawn to explain the interdependent relations hip
20 among the six variables. In 2003, DeLone and McLean updated their IS success model and added service quality as the third quality. This model is reproduced in Figure 2. 1. In this updated IS success model, DeLone and McLean (2003) identified informat ion, system, and service characteristics, which influence information quality, system quality, and service quality. These three types of quality then affect user satisfaction and intention to use. The six dimensions of IS success are: System quality: mea sures the desired characteristics of an information system, such as reliability, accessibility, timeliness; Information quality: measures the desired characteristics of content provided by an information system. Completeness, relevance, and currency are e xamples of characteristics of information quality; Service quality: measures the quality of technical support one receives from a service provider; User satisfaction: system; Use: measures nature of use, navigation patters, number of visits, and number of transactions executed; and Net benefits: assesses the extent of benefit an information system brings to users. One example can be time an d cost savings on shopping.
21 Figure 2 1 Updated D&M IS Success (DeLone and McLean, 2003) B esides user satisfaction, the other stream of IS success research is technology acceptance. Wixom and Todd (2005) (1992) IS success (1989) technology acceptance model (TAM). The question of what determines user satisfaction and how satisfaction affects behavioral beliefs (usefulness and ease of use) and behavioral attitude can be explicitly explained in Wixom and Tod (2005) integrated research model. This model is reproduced in Figure 2 .2 In the first half of the integrated research model, Wixom and Todd (2005) split DeLone and and system satisfaction, which are influenced by information quality and system quality, respectively. Information satisfaction and system satisfaction then influence perception of usefulness and ease of use, which together determine attitude toward the u se of an informati on system and intention to use.
22 Figure 2 2 The Proposed Integrated Research Model ( Wixom and Todd 2005 ). D integrated research model have been widely adopted in num erous research areas. Depending on the domain of the research, different combinations of the three types of quality (information quality, system quality, or service quality) were applied to study user satisfaction, and they were found to be significant pr edictors of user satisfaction (e.g., Chen & Cheng, 2009; Cheung & Lee, 2008; Lin, 2010; Teo, Srivastava, & Jiang, 2008 9; Wang, 2008; Wang & Liao, 2008; Wu & Wang, 2006) In addition to information quality, system quality, and service quality, trust is ano (2006) study on health related information seeking demonstrated that, although the effect is not as large as information quality satisfaction or system quality satisfaction, trust w as found to be statistically significant on user satisfaction. Their findings suggested that health information seekers, who had a higher level of trust with content websites, had a higher level of user satisfaction. Besides the direct effect, trust, in the context of mobile
23 banking, was also found to be a significant moderator between information or system quality and user satisfaction (Chung & Kwon, 2009) Problem Statement Unlike traditional media, such as magazines or library systems, providers of inf ormation on the Internet no longer limits to the websites. With the proliferation of online discussion boards and social networking, a new trend of user generated content seems to be emerging. Many websites nowadays provide a space for users to share and obtain information from each other. Information seekers oftentimes are also information providers. For example, CNet, a technology information website, offers product specification, expert reviews, and consumer reviews. Information providers like CNet provide forums, discussion boards, and review tools for users to share their experience and contribute their knowledge. Information shared by users enriches websites, especially when the information provided by websites is not sufficient to satisfy inform For certain types of information, such as health related topics, user generated (2009) study on Internet users and nonusers health information usage and satisfaction suggested that information seekers were pleased with information and support from a teleconference once a month. And we share information that you cannot find any place (Taha et al., 2009, p. 670).
24 Applegate (1993) suggested t hat information seeker satisfaction consists of material satisfaction and emotional satisfaction. Material satisfaction was defined as a hand, referred to informatio (Applegate, 1993; Bruce, 1998) Hence, Bruce (1998) defined information seeker emotional responses to the (Bruce, 1998, p. 541) (1993) asserted that emotio nal satisfaction is imperative, because it is the key element in determining the intention of continued use in the future. From the literature review of information seeker satisfaction and user satisfaction, it seems apparent that the antecedents of satisf action (e.g., information quality, system quality, user involvement, user characteristics, or trust) are not sufficient to explain information seeker satisfaction on many of the content websites. As stated in the previous paragraph, much of the informatio n is provided by users. However, the effect of user Does user generated content improve the quality of content websites and impact user satisfaction? In order to unde rstand this question, it is necessary to identify the key factors that determine information seeker satisfaction. Therefore, in the second study, the author used a n in ductive approach to identify factors that lead to information seeker satisfaction throug h the collection of recommendation statements. From these findings, practitioners are able to understand why an information seeker recommends a particular
25 website to another information seeker. Also, these findings provide practitioners with a clear insi ght into the role of user generated content in terms of the promotion and enhancement of information seeker satisfaction.
26 CHAPTER III RESEARCH MODEL Comprehensive Model of Information Seeking In the first study, the author adopted the Comprehensive Model of Information Seeking (CMIS) from Johnson and Meischke (1993) and Johnson, Donohue, and Johnson (1995). This model was selected as the theoretical foundation for this research, because it can be used to reflect the causal structure of Web information sou rce selection. CMIS is reproduced in Figure 3 .1 Figure 3 1 Comprehensive Model of Information Seeking (Johnson et al., 1995) Initially, CMIS (Johnson & Meischke, 1993) was developed within the context of health related i nformation, more specifically, breast cancer related information, and it depicts the causal structure of the information seeking process for breast cancer related information. The model contains three elements: antecedent factors, information carrier fact ors, and information seeking actions (Johnson & Meischke, 1993; Johnson et al., 1995). Johnson and Meischke proposed that the action of information seeking is influenced by health related factors (antecedent factors) and information carrier factors.
27 Four antecedent factors motivate a person to seek cancer related information. The four cancer), salience (perceived significance of health information), and beliefs (a bout whether she can do something to prevent breast cancer). For example, a person is more likely to seek breast cancer related information if 1) someone around her has breast cancer (direct experience), 2) she thinks the breast cancer related information is important (salience), and 3) she believes she can do something to prevent breast cancer (beliefs). The second element of CMIS (Johnson & Meischke, 1993), information carrier factors, refers to the characteristics and utility of information carriers. S ince CMIS studies how women use magazines as an information source for breast cancer related information, the characteristic of the information carrier refers to message content attributes, including the editorial tone and communication potential of magazi nes. magazines, and communication potential is their perception about the style and comprehension of the information for which they look. The second information carr ier factor, utility, is determined by how important and relevant the information contained in the information source is to the information seeker. Utility is different from salience in that salience measures the significance of the cancer related informat ion; whereas, utility assesses the significance of the cancer related information provided by magazines (information source). In regard to the utility of the information carrier, Johnson and Meischke claimed that an information source is selected if the i nformation provided meets what the information seeker needs.
28 The third element, information seeking actions, consists of three dimensions: methods, scope, and depth. Method refers to the channel selected and utilized for information seeking. Examples of methods (or channels) include humans, the Web, and elements being read in a magazine. Depth, on the other hand, refers to the degree of readership (glancing over or readin g carefully) of a particular element (e.g., advertisements and articles) in the magazine. Model and Hypotheses Development context of health related information seeking with magazines as the information channel. Johnson et al. (1995) later applied CMIS to study the behavior of technical information seeking in organizations and confirmed the explanatory power of CMIS with regard to information search behavior. Along with prior studies demonstrated that CMIS is generalizable to other information channels and contexts. However, most of the constructs in CMIS were developed specifically for the domain of health related information seeking behavior. Mor eover, the channels, which were studied electronic sources; therefore, information carrier factors do not include information systems related constructs. Accordingly, a direct application of CMIS is not applicable to study online local information seeking behavior. Instead, the author used CMIS as a framework, and then she identified and redefined antecedents and information carrier factors to fit the context of online local information seeking. More specifically, the information source selection model consists of factors from three research domains.
29 First, the author integrated factors (e.g., involvement) from the marketing literature to help identify groups of consum ers who differ in the type of information sources used. Second, factors (e.g., direct experience) from the information seeking literature were used model (TAM), from the I S literature, was integrated into CMIS to asses s information carrier factors. The rationale of these integrations is discussed in the hypotheses section. Antecedents In the proposed model (see Figure 3.2 ), the author identified direct experience, involvem ent, intensity of information seeking, habit strength of using search engines, and perceived search skill as the an tecedents of information source selection. Figure 3 2 Information Source Selection Model
30 Direct Experience Prior experience with a webs ite is an essential factor, which influences continued IS use. Kim and Malhotra (2005) incorporated four theories, namely technology acceptance model, the theory of belief u pdating, the self perception theory, and habit, into a two wave panel model of IS use. This model was built on evidences in the literature that repeated use of an information technology re enforces positive evaluation and judgment of that information tech nology. Prior experience, in the context of e continued use (Liao & Lu, 2008). In the online shopping study, researchers (So, Wong, & Sculli, 2005) found that prior experience with online shopping indirectly influences shopping intention. (2004a) study on everyday life information seeking suggested that a source is selected based on previous use experience. Familiarit y was the most frequent reason cited when people were asked why they selected a specific information source (Fidel & Green, 2004) Continuous use could result in a long term and favorable attitude toward a website. Based on the literature of information seeking (Savolainen & Kari, 2004a; Fidel & Green, 2004) and information systems (Kim & Malhotra, 2005; Liao & Lu, 2008 ; So et al., 2005), when an individual spends a longer time on a specific website, he is more likely to resort to it when he encounters a problem; as such, every information seeker would have high direct experience with whatever information source he used. However, information in each type of information source is different from one another in terms of breadth and depth, and search engines provide more services than specialized websites and centralized
31 sources should be different. The rationale for this argument is presented in the following paragraphs. M ost search engines are embedded within portal websites (e.g., AOL or Yahoo); these portal websites also provide other services such as E mails, personal spaces, news, local weather information, etc. Some search engine sites (e.g., My Yahoo) even offer per sonalization features. The personalization features allow users to select tools and themes they prefer and customize it to a personalized homepage. Once the Web page is personalized and offers the tools they frequently use, users spend a longer time on t hat site for different purposes, such as checking E mail and watching videos. Websites, on the other hand, do not provide as many features as search engines. The scope of content being provided on websites is also limited to certain types of information. According to comScore and 24/7 Wall St.com (Weigley, 2013), among the top five websites where Americans spend most time online, four of them were search engine portal websites. Therefore, information seekers who use search engines as information sources would have higher direct experience with the information source than information seekers who use websites as information sources. Centralized websites contain a broad range of information topics. Information seekers are not necessarily interested in ever y topic being provided on centralized websites. When there is a need for local information, information seekers go to centralized websites and visit the page in which they are interested. As such, their direct experience with centralized websites is lowe r than those who use search engines and specialized websites. Hence, the author proposed,
32 H1: Information seekers with a higher level of direct experience with their preferred information source are more likely to be using specialized websites or search e ngines as opposed to centralized websites. Involvement Salience, in CMIS, was used to assess how significant the cancer related information was to information seekers (Johnson & Meischke, 1993) Salience is excluded in the source selection model for two reasons. First, salience is a trigger that motivates people to look for information. Since the author focused on studying source selection, motivators of information seeking should not be included in the model. Second, the nature of cancer related info rmation seeking is different from the nature of local information seeking. Cancer related information involves high risk. Ignoring the hand, local information is use high risk like cancer related information. Hence, for local information, the author believes that involvement is a better term to assess how significant the local information is to a person. The concept of involvement was introduced in the 1950s ( Arora, 1982) Since then, the term has been adopted in the marketing literature to study consumer behavior. Involvement has been proven to be a factor in consumer information processing (Arora, 1982; L aurent & Kapferer, 1985) Literature of product information seeking suggests that people with a higher level of involvement spend a longer time and a greater effort in searching for more information than others with lower involvement before they make a pu rchasing decision (Beatty & Smith, 1987; Clarke & Belk, 1978)
33 Applied to the context of local information seeking, involvement, in this study, is conceptualized as the perceived concern of a particular activity (e.g., making a decision on where to go duri ng leisure time). When individuals are highly involved with an activity, whether or not they would be satisfied with the decision being made becomes important. That is, they would be more active in information seeking and spend more time to acquire more detailed information about the activity. For example, if a person cares about how to save money on daily shopping, he would be more likely than others to devote more time on locating stores with better deals. Hence, broader and more detailed information might be needed to satisfy his information needs. Accordingly, the author hypothesized, H2: Information seekers with a higher level of involvement with an activity are more likely to be using specialized websites or search engines as opposed to centralize d websites. Intensity of Information Seeking Intensity of information seeking has been studied in both offline and online settings, and it is an important factor influencing information seeking behavior. In offline information seeking, Cole and Balasubr amanian (1993) found that younger adults seek more intensively than older adults, and higher intensity relates to higher consumer satisfaction. That is, information seeker satisfaction increases when they seek more intensively. Conversely, Kuruzovich, Vi swanathan, Agarwal, Gosain, and Weitzman (2008), who examined how information needs and online information seeking impact shopping processes in the context of automobile information, did not find a significant connection between offline information seeking intensity and online seeker satisfaction. Their research model proposed that, when
34 information seekers find more price information online, their intensity of visiting physical dealers (offline information seeking) increases. On the other hand, when info rmation seekers find more product information online, their intensity of visiting physical dealers decreases because, unlike price information, product information about a vehicle does not vary among dealers. Hence, information seekers would visit fewer d ealers (offline information seeking intensity) before making a purchase decision. As a result, offline information seeking intensity (frequency of visit dealers) would be negatively associated with online seeker satisfaction. The result of their study sh owed that the first two propositions were statistically supported, while the third proposition was not. In online settings, since the cost of information seeking is lower, people sought commerce had overcome the boom and bust cycle al., 2006 7, p. 75) than they did at the beginning of the e commerce era. For example, there were 62% more searches in music related information and almost double the searches in air travel information (Zhang et al., 2006 7). However, for a g iven information seeking task, information seeking intensity decreases over time, because marginal value of information seeking decreases when information seekers obtain more information (Erdem, Keane, nc, & Strebel, 2005). In the context of health info rmation, search intensity was found to be a significant mediator between frequency of use and communication effects. That is, frequency of using online health information sources was positively associated with intensity of information seeking, which in tu rn regarding the information they found online (Lueg, Moore, & Warkentin, 2003).
35 In this paper, intensity of information seeking is defined as the extent of information sources used and amount of information obtained by an information seeker before he makes a decision on a particular activity. Local information seeking is a repeated activity. When an information seeker uses a search engine as an information source, the re is a great chance that he does not look for local information as frequently as those who use a specialized website or a centralized website, so he does not know any local information website. By contrast, when people consult a great extent of informati on sources they are more likely to be frequent information seekers and have known a few websites. The author, therefore, proposed, H3: Information seekers who consult greater extent of information sources are more likely to be using content websites as opposed to search engines. Habit Strength of Using Search Engines sequences of acts that become automatic responses to specific situations, which may be anken, A arts, & van Knippenberg 1997, p.540). Once a behavior becomes a habit, the action is performed automatically, meaning that one does not think about the behavior before doing it (Limayem, Hirt, & Cheung, 2007). In other words, habit is not only a routine practice, but also an unconscious pattern of behavior. In the information systems (IS) literature, habit has been found to be a significant factor in predicting IS usage (Limayem & Hirt, 2003). The first search engine was created in 1990 by Alan Emtage, a student at McGill University in Montreal (Wall, n.d.). According to a report from the Pew Research
36 activity on the Web. Among daily Internet activities, 49% of the Web users searched information online on a typical day. Usage of search engines increased 69% from 2002 to May 2008 (Fallows, 2008). A search engine retrieves and lists websites according to search terms. People find the likely content websites by enteri ng key words. When information seekers have no idea where to look for particular information or when the URL of the content website search. In this study, the habit st rength of using search engines is defined as the extent to which an information seeker tends to use search engines automatically whenever he encounters a problem. When one develops the habit of doing search using search engines for any information, he may choose to use search engines to look for local information or employ search as a method to look for local information. Since centralized websites contain a variety of information, seeking information on centralized websites also requires in site searchin g. That is, centralized website users may have stronger habit strength of using search engines as well. Hence, the author proposed, H4: Information seekers with stronger habit strength of using search engines are more likely to be using search engines or centralize websites as opposed to specialized websites. Perceived Search Skill In the IS literature, computer self efficacy refers to a Higgins, 1995). Computer self efficacy was found to be a significant factor in influencing actual computer use. More specifically, people who perceive higher
37 computer self efficacy use computers more frequently, spend a longer time on computers, and experience less computer anxiety t han others (Compeau & Higgins, 1995). Computer self influenced by their perceived self efficacy. Perceived search skill, in this study, refers to the extent of confidence an informatio n seeker feels about utilizing search engines to find the needed information. Savolainen (2002) introduced a model of network competence in the context of information seeking. According to Savolainen, network competence refers to the required skills peop le need in order to find information on the Web. Among the four major criteria of network competence for information seeking, the skill of using search engines is one of the main requirements for information seeking. Based on the self efficacy theory and previous findings, Savolainen (2002) claimed that level of confidence is a strong predictor of the actual behavior. This means that, if an information seeker believes she has the skill of using search engines, she would be more likely to use search engin es when looking for information. In the Pew Internet & Research Project study (Fallow, 2005), it was found that 92% of the search engine users reported moderate to high confidence in their search skill, while only 8% users reported low confidence. With th e high percentage of confidence that search engine users have, it is reasonable to assume that people with greater perceived search skill would prefer to use search engines when they have information needs, whereas people with lower perceived search skill would be more likely to browse and seek needed information on a specialized website. As stated in earlier sections, the use of a centralized website also requires an in site search. Hence, the author proposed,
38 H5: Information seekers with greater perceiv ed search skill are more likely to be using search engines or centralize websites as opposed to specialized websites. Information Carrier Factors In the Model of Risk Information Seeking and Processing proposed by Griffin, Dunwoody, and Neuwirth (1999), th e researchers identified relevant channel beliefs as an essential factor, which affects information seeking and strategy selection. In Griffin et channel. Griffin et al. (1999) influences the way they look for information. Because how an information seeker perceives the information source influences his decision of whether to use the site or not, the concept of channel beliefs can be applied to information source selection. In CMIS (Johnson & Meischke, 1993) information carrier factors include characteristic (e.g., editorial tone) and utility. While characteristic refers to information perception toward the information, which is provided by the information sources. These two factors were developed and used to study non electronic information sources. Information sources, in this s tudy, focus on content websites and search engines. Accordingly, the technology acceptance model (Davis, 1989), from the discipline of information systems, is a more suitable theory to help understand and examine information carrier factors. The technolo gy acceptance model (TAM) is one of the most important and widely adopted models in the information systems literature. In TAM, intention to use a system.
39 In addition t o perceived usefulness and perceived ease of use from TAM, the author also included trustworthiness in her proposed model. Perceived usefulness, perceived ease of use, and trustworthiness are three beliefs that have been well studied in the field of infor website use (Davis, 1989) and purchase intention (Buttner & Goritz, 2008). Perceived Usefulness Perceived usefulness of a website has been proven to have statistically significant inf luence on the frequency of use (Lemire, Par, Sicotte, & Harvey, 2008). Information usefulness is conceptualized as the degree to which multiple choices of information sourc es in mind, one important criteria of selection is the usefulness (Yang, 1997), and source quality becomes more important when they have more alternative sources (Savolainena & Kari, 2004a). That is, when an information seeker uses a particular website as his preferred source, he must think the website is useful. Otherwise, he can go directly to a search engine and do a quick search. Hence, the perceived usefulness may be higher with centr alized or specialized websites. Perceived Ease of Use The least effort principle and source quality are two of the research findings in the literature of source choice decisions. The former contends that source accessibility is the main factor in source selection, while the latter contends that source quality is the k information source for their academic work, ease of use was ranked as the third important
40 factor. That is, students are more likely to use an information source if they believe they can find their needed information easily. Centralized websites provide a broader scope of lo cal information. When an information seeker looks for a particular piece of local information (e.g., local events during the holidays), often, it is necessary to review a variety of information to locate the information he wants. Hence, the perceived eas e of use with centralized websites may be lower than with search engines and with specialized websites. Perceived Trustworthiness Trustworthiness is an assessment of whether the information source offers truthful and unbiased information. The model of w ebsite stickiness, proposed by Li, Browne, and Wetherbe (2006), has proved that trustworthiness in online business to consumer relationships is positively associated with inform local information seeking usually does not involve as high risk as health related information seeking, some people might be concerned about monetary or time loss if inapprop riate decisions are made. If a local information seeker is highly interested in a particular topic, it is important for him to obtain accurate and reliable information to tion sources is an important factor in source selection. Search engines, unlike content websites, are information so urces, which produce a list of websites according to search keywords and algorithms. The information, which people obtain, is not directly provided by search engines. Furthermore, many Internet users are aware that, frequently, many modern search engines present paid advertising or
41 sponsored links on the search results pages. For such reasons, the perceived trustworthiness with search engin es may be lower than with centralized websites and with specialized websites. In sum, the author proposed the following, H6a: Information seekers who perceive a higher level of usefulness with their preferred information source are more likely to be using content websites as opposed to search engines. H6b: Information seekers who perceive a higher level of ease of use with their preferred information source are more likely to be using search engines or specialized websites as opposed to centralized website s. H6c: Information seekers who perceive a higher level of trustworthiness with their preferred information source are more likely to be using content websites as opposed to search engines. Control Variables A number of demographic variables, namely gender age, and education have r search engine usage patterns. Gender Teo and Lim (2000) found that, in comparison to women, men 1) browsed the Web more frequently, 2) spent a lo nger time on downloading files and software, and 3) were more likely to use the Web to look for information and to get free resources and product support. Although women have equal access to computers and the Web, researchers found that gender differences still exist in information technology usage (Li & Kirkup differences among Chinese and British students concluded that male students in both
42 countries were more experienced in using the We b and more confident in using search engines to locate needed information than female students. These conclusions supported prior findings. On the other hand, women had less confidence in their ability to find information although they performed as well as male students in finding information on the Web. In addition, Li and Kirkup also found that female students were more likely to feel lost when they looked for information on the Web. Age Regarding age, in offline information seeking research, researc hers have offered numerous explanations as to why people in different age groups have different information seeking patterns. Laroche, Cleveland, and Browne (2004) categorized these explanations into external and internal factors. Internal factors refer ability, memory, and learning abilities; external factors are related to the environmental changes. Laroche et al. (2004) explained that different generations experience different environmental stimuli. The main environmental stimulus for online information seekers is the popularity of Internet access. Younger generations, who grew up with convenient Web access, are expected to be more experienced on the Web. People from different generations experience different surroundings. The te rm, is the internet and a search engine, Google being the most popular one. This is in distinction to previous generations that grew up and were educated before th e widespread availability of the internet, and whose source of knowledge was through books and Internet & American Life Project on search engine users found that youn ger people were more likely to use search engines and used them more frequently than older people
43 (Fallows, 2005). It is thus reasonable to postulate that younger information seekers are more likely than older information seekers to develop stronger habit strength of using search engines. Education The third control variable, education, was also found to be an conducted by Times Mirror Center (1995), it was found that peo ple with higher education are more likely to be Web users than those with lower education. One possible reason is that people who have a college education have more exposure to the Web for classes and school work. In sum, gender, age, and education as con trol variables are included in the source selection model.
44 CHAPTER IV METHODOLOGIES Quantitative Data Collection Procedure Since the target group of the local information source selection study was people with online local information seeking experience, a Web based survey was used as the research methodology. The author used online forums to invite participation. For the purpose of this research study, online forums (see Appendix A) with topics, which were related to general local information, entertain ment information, services information, and shopping information, were targeted. The solicitation stated that the author is seeking responses only from people who are age 18 or above, live in the United States, and have experience in searching local infor mation on the Web. Since the author did not know how many people viewed the solicitation posts on the forum, she was not able to ascertain the response rate. The forum solicitations consisted of the research purpose, statement in regard to privacy and con fidentiality, and the hyperlink to the survey website. Potential participants were directed to the survey website by clicking on the hyperlink shown on the solicitations. The survey website was developed and hosted on the Ouray server operating under the University of Colorado Denver domain. The survey website outlined instructions on who should participate in the study, and survey lengths, etc. Respondents who clicked
45 During the survey, participants were asked to select one of the local information topics (local entertainment information, professional services information, or store sales information), which they had looked for recently. For the particular type of information being selected, participants were asked to select one type of local information source (search engines, centralized websites, or specialized websites) that they used most often. Dependi ng on their answers, the wording of follow up questions regarding the information source was different. Hence, the author had a total of six versions of survey questions: three types of information topics multiplied by two information seeking methods (sea rch engines versus content websites). If the information seeking method was a content website, then the respondent answered the same set of survey questions with respect to either a specialized website or a centralized website as per the most frequently u sed website identified by each respondent. See Appendix B for one of the survey versions. Measures Direct Experience familiarity with the preferred information source. For this construct, thre e items were adapted from Flanagin and Metzger (2000) The original items were used to measure Internet experience, and two of these items, namely, level of expertise (estimate whether one is an Internet expert) and level of access (estimate how easy it i s to access the author selected and modified the rest of the three items to fit the context of direct experience with information sources. Frequencies, level of experie nce, and familiarity
46 Involvement (1995) measurement of consumer purchase decision involvement were adapted to the context of local information. A fi ve point semantic differential scale was used to assess how much people care about the specific topic they wish to find. Involvement was examined by asking people how concerned they were when they made a decision about a certain topic. Intensity of Inform ation Seeking To measure intensity of information seeking, the author developed three items to assess the breadth (single versus multiple) of websites, which people visit to complete a certain local information seeking task. Intensity was measured with information as I can get Habit Strength of Using Search Engines Habit strength of using search engines is defined as the automatic behavior of going to a search engine website when one has information needs. The measurement of habit was adapted fro m Verplanken and (2003) Self Report Habit Index (SRHI). The index was developed based on three habit features, namely, history of repetition, automaticity, and expression identity. All 12 items were statistically proven to be reliable and valid, and they were loaded unto one dimension. Habit strength of using search engines is different from daily or weekly habits (e.g., drinking coffee in the morning), in that, it is not a spontaneous action or a routine. Instead, people tend to use search eng ines only when there is a need. Since the (Verplanken & Orbell, 2003, p. 1329) do not fit into this
47 research context. For the purpose of this study, 4 of the 12 items were adapted, and one Verplanken & Orbell 2003, p. 1329) were adapted becaus e the habit strength of using search engines would be stronger when people 1) use search engines frequently, 2) automatically resort to search engines for any information, 3) always start with a search engine, and 4) use search engines for a long time. Wo rding was modified to fit the context of habit strength of added to estimate wh ether a search engine is the first option people use. Perceived Search Skill Search engines allow users to apply search operators on query to narrow down or restrict their searches. With advanced search operators, more relevant websites and more specifi c information can be retrieved. Perceived search skill, in this study, was measured using 5 questions of online searching self efficacy beliefs adapted from Monoi, O'Hanlon, and Diaz (2005) Of the 12 questions, 5 were selected because these 5 questions measure whether one can identify keywords and synonyms, implement correct search strategy (e.g., when to browse search results or when to search further), and identify keywords from the search results and use it for further search. These 5 items were repo rted as a unidimensional scale in the original study. Perceived Usefulness Perceived usefulness is conceptualized as the perceived relevance of information regarding a specific seeking topic. To assess how useful the (2007 ) measurement of content usefulness was adapted.
48 Perceived usefulness was indicated on a 5 point Likert scale from (1) strongly disagree to (5) strongly agree. Perceived Ease of Use Perceived ease of use is estimated based on how easy a user feels about finding the needed information on a specific website or a search engine. Measurement from previous research on information technology usage (Agarwal & Karahanna, 2000; van der Heijden, 2003) was adapted to the context of information seeking. Four items with a 5 point Likert scale ranging from (1) strongly disagree through (5) strongly agree were employed. Perceived Trustworthiness Perceived trustworthiness is estimated based on four items: ability, credibility, benevolence, and overall trust. These fo ur measurements were developed by Battacherjee (2002) for the context of e commerce. Responses were Data Analysis Descriptive Statistic Table 4. 1 presents the demographic data for the su rvey respondents. The total number of responses was 99. During the survey, participants were asked to select an information topic that they recently looked for and to choose the website type for the identified information. For information topics, respon dents could choose from local entertainment, professional services, or store sales information. As for the website type, depending on their experience, respondents chose from search engines, centralized websites, or specialized websit es. Centralized webs ites refer to those that provide a broad range of local information. By contrast, specialized websites are those that provide only a specific type of local information. Of the 99 responses, almost 50%
49 looked for local entertainment information and nearly 60% utilized search engines (see Table 4. 2). Table 4. 1 Characteristics of Survey Sample Variables/Groups Frequency Percentage Gender Male 62 62.6 Female 37 37.4 Age 18 22 3 3.0 23 29 18 18.2 30 35 18 18.2 36 45 21 21.2 46 55 24 24.2 56 65 11 11.1 66 89 4 4.0 Education Less than high school 2 2.0 High school diploma or GED 20 20.2 Some college but no diploma 30 30.3 28 28.3 Some graduate level education 8 8.1 Master's degree or higher 11 11.1 Table 4. 2 Type of Information and Source/Provider Variables Frequency Percentage Information Topic Entertainment 48 48.5 Professional services 27 27.3 Store sales 24 24.2 Source/Provider Specialized websites 20 20.2 Centralized websites 21 21.2 Search engines 58 58.6 Even though search engines were the dominant tools for local professional services (70%) and store sales information (75%), search engine s were not as dominant for local entertainment information (44%). This result implied that specialized and centralized websites could emphasize entertainment information to attract more visitors. Also, there was a gender difference concerning information topics chosen: the number of
50 female respondents for each information topic was almost the same; however, male respondents apparently were more interested in entertainment information than in professional services or store sales information (see Table 4. 3) Table 4. 3 Information Topic by Gender and Source Type Information Topic: Entertainment Professional services Store sales Gender 1 Female 13 11 13 Male 35 16 11 Source/Provider 2 Specialized websites 13 3 4 Centralized we bsites 14 5 2 Search engines 21 19 18 1 Significance level for the Pearson chi square statistic is 0.074 2 Significance level for the Pearson chi square statistic is 0.053 There was no significant gender difference in overall perceived search skill (t = 0.123, df = 97, p = 0.902), suggesting that male and female respondents possessed the same level of perceived skill of utilizing search engines. This result was different from the earlier finding (Li and Kirkup, 2007) The difference may be caused by cultural (2007) survey focused on British and Chinese respondents. However, age (t = 2.882, df = 97, p = 0.005) and education (t = 2.331, df = 97, p = 0.022) differe nces existed in overall perceived search skill. Respondents between the ages of 18 and 45 possessed a higher higher possessed a higher perceived search skill than oth ers. Nonresponse Bias As mentioned earlier, the author was not able to calculate the response rate for the unknown number of people who saw the survey solicitations, so nonresponse bias tests were conducted. The author examined this with two methods.
51 F irst, she compared the response demographic with the demographic of online news readers. Online news readers were used to compare, because the news industry is one of the new entrants in the market of local information seeking. The American Press Institu te (Gray, 2008) suggested that news providers should go beyond the limits of news and serve as a local information and connection utility to connect consumers and local businesses. Several news providers, such as Chicago Tribune, Cox Ohio Publishing, and Seacoast Media Group, have done so and entered the local information seeking market. In addition, the Pew Research Center found that newspapers are one of the major sources for local information (e.g., events, business, restaurants) (Rosenstiel, Mitchell, Purcell, & Rainie, 2011). Accordingly, it is reasonable to use online news readers as the target population. A survey done by Nielsen (Olmstead, Mitchell, & Rosenstiel, 2011) reported that among the top 25 news web sites, 17 websites had more male visito rs than female 29%) or some college education (24 27%). This matched the demographic statistic (see Table 4. 1) in this study, suggesting that the sample represented the target local info rmation seekers. Second, the author divided responses into two groups, early and late respondents, and used late respondents as surrogates for non respondents (Armstrong & Overton, 1977) The two groups were compared in terms of their involvement, perce ived search skill, habit strength of using search engines, and intensity of information seeking. The results of t test showed that there was no significant difference in the means of aforementioned variables (p = 0.959, 0.251, 0.696, 0.457, respectively) between the two groups, indicating no nonresponse bias. The author also conducted tests to see if differences existed for age level, educational level, gender, type of information source,
52 and type of information topic. No significant differences were fou nd between the two groups on the demographic (p = 0.234, 0.849, 0.121, respectively), type of information source (p = 0.678), or type of information topic (p = 0.687). Both results suggested that nonresponse bias was not an issue in this study. Construct Reliability and Validity Confirmatory factor analysis was used to assess the validity of the eight latent constructs: habit strength of using search engines, involvement, intensity of information seeking, direct experience, perceived trustworthiness, per ceived usefulness, perceived ease of use, and perceived search skill. Confirmatory factor analysis (CFA) was performed in smartPLS 2.0 ( Ringle, Wende, & Will, 2005 ) A PLS based structural equation modeling (SEM) tool was selected rather than a covarianc e based SEM tool, such as AMOS, because of small sample size (N = 99). All latent variables were modeled with reflective indicators and the constructs, which were related, were connected to one another (see Appendix C). It has been suggested that the sta ndardized loading estimates should ideally be 0.7 or higher (Hair, Black, Babin, Anderson, & Tatham, 2006) One item of habit of using a search engine had a loading less than 0.7. It was removed, and the measurement model was re evaluated (see Table 4 .4 ) Furthermore, average variance extracted for each construct was higher than 0.5. To assess internal consistency of the latent constructs, recommended value of 0.7. Stan dardized loadings, average variance extracted, and internal consistency combined suggested good convergent validity.
53 Table 4.4 Convergent Validity Constructs Indicators Loadings AVE Alpha Composite Reliability Direct Experience EXP 1 .8 7 0.77 0.85 0.91 EXP 2 .86 EXP 3 .91 Involvement INV 1 .84 0.69 0.78 0.87 INV 2 .88 INV 3 .77 P. Search Skill SKILL 1 .72 0.73 0.91 0.93 SKILL 2 .78 SKILL 3 .93 SKILL 4 .93 SKILL 5 .90 Habit of Using SE HABIT 1 .70 0.70 0.85 0.90 HABIT 3 .89 HABIT 4 .87 HABIT 5 .87 Intensity of IS INTEN 1 .93 0.84 0.91 0.94 INTEN 2 .91 INTEN 3 .91 P. Ease of Use EASE 1 .83 0.72 0.87 0.91 EASE 2 .87 EASE 3 .83 EASE 4 .86 P. Usefulness USEFUL 1 .88 0.82 0.93 0.95 USEFUL 2 .91 USEFUL 3 .94 USEFUL 4 .91 P. Trustworthiness TRUST 1 .86 0.68 0.85 0.90 TRUST 2 .82 TRUST 3 .80 TRUST 4 .82 P.: Perceived; SE: Search Engines; IS : Information Seeking To verify discriminant validity, the author used (1981) criteria, which the square root of average variance extracted should be greater than correlations. As shown in Table 4. 5, the square root of average variance extracted was larger than the inter construct correlations, suggest ing good discrimina nt validity for all constructs.
54 Table 4.5 Discriminant Validity 1 2 3 4 5 6 7 8 1. Direct Experience 0.88 a 2. Involvement 0.25 b 0.83 3. P. Search Skill 0.38 0.31 0.86 4. Habit of Using SE 0.29 0.21 0.52 0.84 5. Intensity of IS 0.54 0.44 0.50 0.30 0.92 6. P. Ease of Use 0.68 0.38 0.56 0.42 0.61 0.85 7. P. Usefulness 0.61 0.32 0.38 0.27 0.62 0.75 0.91 8. P. Trustworthiness 0.45 0.28 0.44 0.44 0.47 0.64 0.59 0.83 a Bold numbers are the square root of average variance extracted. b Off diagonal elements are the inter construct correlations. P.: Perceived; SE: Search Engines ; IS : Information Seeking Once the reliability and validity were verified, the author used loadings to calculate a factor score f or each latent variable. The factor scores then were used in multinomial logistic regressions to assess the hypotheses. Multinomial Logistic Regression To identify independent factors that best discriminate among specialized websites, centralized websit es, and search engines, multinomial logistic regression (PASW Statistic 18.0) was applied. Multinomial logistic regression is a popular and widely used method to analyze three level categorical variables (e.g., Riggs, 2008; Cupini et al., 2002; Ariyachand ra & Watson, 2010) because of its flexibility in the assumptions. Multinomial logistic regression does not require the independent variables to be normally distributed, linearly related, or equal variance within each category (Tabachnick & Fidell, 1996). The more skewed the data, the better and more accurate the multinomial logistic regression is than the multivariate analysis of variance (MANOVA). The results of the tests of normality showed that none of the independent variables were normally distribut ed ( p logistic regression is flexible in the types of data contained in the independent variables (Tabachnick & Fidell, 1996). The independent variables, in the source selection model,
55 include nominal (e.g., gender), ord inal (e.g., education), and interval variables (e.g., direct experience). Accordingly, multinomial logistic regression was chosen to examine the three types of information source. In the first regression, the author used centralized websites as the refere nce category to compare centralized websites with specialized websites and search engines. In the second regression, specialized websites were used as the reference category to compare specialized websites with search engines. The result indicates that th e overall model fit (see Table 4. 6) was statistically significant (Chi Square = 68.30, df = 22, p 2 (the model explains 58.3% of total variation of the dependent variable) also suggested a good model fit (see Table 4. 7). Hence, the author concluded that the model was good enough to determine the behavior of source selection. Table 4.6 Overall Model Fit Model Model Fitting Criteria Likelihood Ratio Tests 2 Log Likelihood Chi Square df Sig. Intercept Only 191.123 Final 122.821 68.302 22 .000 Table 4.7 Pseudo R Square Cox and Snell .498 Nagelkerke .583 McFadden .357 In order to evaluate the predictive accuracy of the model, maximum chance criterion and proportional chance criterion tests were applied. Maximum chance criterion was 58.6% (largest group divide by total sample size 58/99), and the proportional chanc e criterion was 43% (20.2% 2 +21.2% 2 +58.6% 2 ). The hit ratio was 74.7% (see Table 4. 8) which was greater than the maximum chance criterion. The criterion was that the hit
56 ratio should be 25% greater than proportional chance criterion (43% *125% = 53.8%), an d the model accuracy also met this requirement. Table 4. 8 H it R atio Observed Predicted Specialized Websites Centralized Websites Search Engines Percent Correct Specialized Websites 9 5 6 45.0% Centralized Websites 2 12 7 57.1% Search Engines 2 3 53 9 1.4% Overall Percentage 13.1% 20.2% 66.7% 74.7% The likelihood ratio tests (see Table 4. 9) showed the contribution of each independent variable to the overall model. Direct experience, intensity of information seeking, habit strength of using search en gines, perceived search skill, perceived ease of use, and perceived trustworthiness are statistically significant ( p was marginally significant (p = 0.077). All control variables (age, education, gender) and perceived usefulness were not significant. Table 4. 9 Likelihood Ratio Tests Effect Model Fitting Criteria Likelihood Ratio Tests 2 Log Li kelihood of Reduced Model Chi Square df Sig. Intercept 122.821 a .000 0 Age 124.660 1.839 2 .399 Education 126.948 4.127 2 .127 Gender 124.472 1.651 2 .438 Direct Experience 131.733 8.912 2 .012 Involvement 127.938 5.117 2 .077 Intensity of IS 135. 741 12.920 2 .002 Habit of Using SE 133.355 10.534 2 .005 P. Search Skill 129.677 6.856 2 .032
57 Table 4.9 Likelihood Ratio Tests (Continued) P. Usefulness 122.842 .021 2 .990 P. Ease of Use 135.229 12.408 2 .002 P. Trustworthiness 134.954 12.133 2 .002 The chi square statistic is the difference in 2 log likelihoods between the final model and a reduced model. The reduced model is formed by omitting an effect from the final model. The null hypothesis is that all parameters of that effect are 0. a This reduced model is equivalent to the final model because omitting the effect does not increase the degrees of freedom. P.: Perceived; SE: Search Engines ; IS : Information Seeking Table 4. 10 shows the result of the two multinomial logistic regressi ons. As mentioned above, centralized websites were the reference category in the first regression, while specialized websites were the reference category in the second regression. Tables 4. 9 and 4. 10 together show the empirical evidence that supported th e proposed information source selection model. Table 4. 10 Parameter Estimates B Std. Error Wald df Sig. Exp(B) 95% Confidence Interval for Exp(B) Lower Bound Upper Bound Specialized Websites vs. Centralized Websites (1 st Regression) Intercept 1.699 3.595 .223 1 .637 Age .749 .993 .569 1 .451 .473 .068 3.309 Education .562 .818 .473 1 .492 .570 .115 2.832 [Gender = 1] 1.246 .998 1.558 1 .212 .288 .041 2.034 [Gender = 2] 0 a . 0 . . Direct Experience .648 .333 3.792 1 .052 1.911 .996 3.669 Involvement .219 .240 .837 1 .360 1.245 .778 1.993 Intensity of IS .080 .262 .094 1 .759 .923 .552 1.543 Habit of Using SE .477 .202 5.560 1 .018 .621 .417 .923 P. Search Skill .379 .193 3.846 1 .050 .685 .469 1.000 P. Usefulness .037 .2 58 .021 1 .886 1.038 .626 1.721 P. Ease of Use .695 .311 4.992 1 .025 2.004 1.089 3.686 P. Trustworthiness .333 .274 1.477 1 .224 .716 .418 1.227
58 Table 4.10 Parameter Estimates (Continued) Search Engines vs. Centralized Websites (1 st Regression) In tercept 8.126 3.692 4.844 1 .028 Age .299 .815 .134 1 .714 1.348 .273 6.657 Education 1.438 .772 3.468 1 .063 .237 .052 1.078 [Gender = 1] .628 .862 .530 1 .466 .534 .099 2.891 [Gender = 2] 0 a . 0 . . Direct Experience .730 .262 7.762 1 .0 05 2.074 1.242 3.466 Involvement .425 .202 4.434 1 .035 1.529 1.030 2.270 Intensity of IS .640 .248 6.645 1 .010 .527 .324 .858 Habit of Using SE .039 .194 .040 1 .842 1.039 .711 1.520 P. Search Skill .014 .161 .008 1 .930 1.014 .739 1.392 P. Usefuln ess .022 .229 .009 1 .923 1.022 .653 1.601 P. Ease of Use .983 .325 9.153 1 .002 2.672 1.413 5.050 P. Trustworthiness .716 .262 7.481 1 .006 .488 .292 .816 Search Engines vs. Specialized Websites (2 nd Regression) Intercept 9.825 3.648 7.255 1 .007 Age 1.047 .792 1.747 1 .186 2.850 .603 13.46 9 Education .876 .689 1.617 1 .204 .417 .108 1.607 [Gender = 1] .618 .787 .617 1 .432 1.855 .397 8.671 [Gender = 2] 0 a . 0 . . Direct Experience .082 .261 .098 1 .754 1.085 .650 1.811 Involvement .2 05 .198 1.076 1 .300 1.228 .833 1.810 Intensity of IS .560 .224 6.225 1 .013 .571 .368 .887 Habit of Using SE .516 .182 8.066 1 .005 1.675 1.173 2.391 P. Search Skill .393 .161 5.928 1 .015 1.481 1.080 2.032 P. Usefulness .015 .187 .006 1 .937 .985 683 1.422 P. Ease of Use .288 .314 .841 1 .359 1.333 .721 2.467 P. Trustworthiness .383 .179 4.589 1 .032 .682 .480 .968 a This parameter is set to zero because it is redundant. P.: Perceived; SE: Search Engines; IS: Information Seeking Direct experience was a significant factor (see Table 4. 10) in discriminating between specialized websites and centralized websites (p = 0.05) and in discriminating between search engines and centralized websites (p = 0.005); however, direct experience was not a significant factor in discriminating between search engines and specialized
59 websites (p = 0.754). Regression results showed that 1) information seekers who used specialized websites have higher direct experience than those who used centralized websites an d 2) information seekers who used search engines have higher direct experience than those who used centralized websites. Parameter estimates suggested one unit increase in direct experience, 1) the odds of using specialized websites as opposed to centrali zed websites increased by 91.1% and 2) the odds of using search engine as opposed to centralized websites increased by 107%. Hence, hypothesis 1 was supported. Involvement was statistically significant (p = 0.035) in discriminating between search engines that information seekers who used search engines were more involved in the activity than those who used centralized websites. With one unit increase in involvement, the odds of us ing search engines as opposed to centralized websites increased by 52.9%. Involvement, however, was not a significant factor in determining whether a specialized website or a centralized website was used (p = 0.36), or whether a search engine or a specia lized website was used (p = 0.30). Hereinafter, hypothesis 2 was partially supported. Intensity of information seeking was a significant factor in determining between search engines and centralized websites (p = 0.01) and in determining between search eng ines and specialized websites (p = 0.013); nevertheless, intensity of information seeking was not a significant factor in determining between specialized websites and centralized websites (p = 0.759). Regression results showed that information seekers who consulted more sources tended to use (centralized or specialized) websites as opposed to search engines. Parameter estimates suggested that one unit increase in
60 intensity, 1) the odds of using centralized websites as opposed to search engines increased b y 47.3% and 2) the odds of using specialized websites as opposed to search engines increased by 42.9%. Hence, hypothesis 3 was supported. The result also did not look fo r local information frequently; accordingly, they were not familiar with local information websites and used search engines as information sources. Habit strength of using search engines was a significant factor in discriminating between specialized websit es and centralized websites (p = 0.018) and in discriminating between search engines and specialized websites (p = 0.005), but there was no significant coefficients suggested th at information seekers with stronger habit strength of using = ) implied that with one unit increase in habit strength of using search engines, 1) the odds of using centralized websites as opposed to specialized websites increased by 37.9% and 2) the odds of using search engines as opposed to specialized websites incr eased by 67.5%. Hypothesis 4 was supported. Perceived search skill corresponded to habit strength of using search engines. Information seekers who perceived they had higher search skill tended to 1) use centralized websites as opposed to specialized webs 0.379) and 2) use search perceived search skill, 1) the odds of using centralized websites as opposed to specialized websites increased by 31.5% (p = 0.05) and 2) the odds of using search engines as
61 opposed to specialized websites increased by 48% (p = 0.015). So, hypothesis 5 was supported. Hypotheses 4 and 5 together demonstrated that using centralized websites, similar to using search engines, required in site search; therefore, when information seekers obtain more experience in using search engines and develop confidence in their search skill, they are more likely to use centralized websites as opposed to specialized websites where users mainly use browsing as an information seeking method. The likelihood ratio tests (see Table 4. 9) indicated that both perceived ease of use and perceived trustworthiness were significant, but perceived usefulness was not significant. So hypothesis 6a was not supported; perceive d usefulness was not significant factor in determining which information source was used (p = 0.886, 0.923, 0.937, respectively). Perceived ease of use was a significant factor in discriminating between specialized websites and centralized websites (p = 0. 025) and in discriminating between search engines and centralized websites (p = 0.002), but there was no significant coefficients suggested that 1) information seekers who used specialized websites 0.695) and 2) information seekers who used search engines perceived a higher level of implied that with one unit increase in perceived ease of use, 1) the odds of using specialized websites as opposed to centralized websites increased by 100.4% and 2) the odds of using search engines as opposed to centralized websit es increased by 167.2%.
62 argument that centralized websites contained a broader scope of local information and required greater effort to locate a particular piece of local information. Perceived trustworthiness was not a significant factor in discriminating between specialized websites and centralized websites (p = 0.224), but it was significant in discriminating between search engines and centralized websites (p = 0.0 06) and in coefficients suggested that 1) information seekers who used centralized websites = 0.716) and 2) information seekers who used specialized websites perceived a higher level implied that with one unit increase in perceived trustworthiness, 1) the odds of using centralized websites as opposed to search engines increased by 51.2% and 2) the odds of using specialized websites as opposed to search engines increased by 31.8%. Hypothesis proposition that information seekers perceived a higher level of trustworthiness with content websites r ather than with search engines. Qualitative Data Collection Procedure Three types of websites were chosen to study information seeker satisfaction: prod uct information websites, health information websites, and local information websites. To investigate what website features people care about when looking for pre purchase product information, health information, and local information, an inductive
63 approa ch, based on qualitative analysis on actual recommendation statements, was applied. The author used recommendation as a surrogate of satisfaction. The rationale of using recommendation was that people seldom express how satisfied they are about a particu lar website. Instead, people are more likely to make a recommendation if they are satisfied with a website. Moreover, researchers (e.g., Lymperopoulos & Chaniotakis, 2008; Bunker, Rajendran, & Corbin, 2013), in different disciplines, have found that cons umer satisfaction is positively related to recommendation. Therefore, it is reasonable to use recommendation statements to examine satisfaction. Google was used to collect recommendation statements, which one person made to others. The author combined ke ywords, including recommend provide and check with website names, in search queries. For example, she inputted as a e the use of double quotes tells Google to search for exact words, operator works as a placeholder that can be any unknown words (Google, n.d.). For pre purchase product information, the author used website names that were s (2008) articles. With regard to health information, the ranking list provided by eBizMBA (2011) was used. This list provided the 15 most Traffic Rank, and U.S. Traf 2011). When the author evaluated the recommendation statements regarding these 15 health information websites, she found other health information websites, such as Go Ask Alice, PatientsLikeMe, and MedHe lp. The author marked down the names of these
64 websites, and included them in search terms later. The method of snowball sampling was applied to local information websites. Since there was no list of popular local information websites, the author started with the websites that she ha d known from the local information source selection study. More website names were collected when the author evaluated the recommendation statements. Appendix D shows the full list of website names used for each type of info rmation topic. For each search query, such as the author went over each page until she reached the end of the search results. Statements that stated why a website was recommended or not recommended were collected. Typically, recommend ation statements specified why certain websites were useful, while the non recommendation statements helped understand what were missing from these websites. The author continued to collect recommendation statements until the saturation point was reached and no new concept was emerged. A total of 201 recommendation statements were collected: 79 for product information websites, 89 for health information websites, and 33 for local information websites. Data Analysis In qualitative analysis, the processes o f data analysis and data collection were performed interactively. Each recommendation statement was analyzed after it was collected. The recommendation statements were analyzed using the grounded theory approach suggested by Strauss and Corbin (1998). M ore specifically, the author used open coding, axial coding, and selective coding techniques to: 1) identify and label concepts, 2) group these concepts into a more abstract level of categorization, 3) relate categories to their subcategories, and 4) integ rate the categories. The purpose of open
65 coding was to examine the statements closely and to compare the differences and were categorized under the same concept. During the open coding, the author wrote memos to: 1) define the concepts, 2) interpret the statements, 3) memo what she thought, and 4) summarize the findings. While open coding was the process of fracturing data into different concepts, axial coding was to reassemble data back together (Strauss & Corbin, 1998). In axial coding, preliminary concepts developed in open coding were compared and related to one another. Axial coding helps researchers to think deeper and to understand the relationship among concepts. Once the open coding and axial coding were conducted, selective coding was performed. Similar to axial coding, selective coding involved integration of categories. In selective coding, a more abstract c ore category was identified, and the remaining categories were linked to the core category to form the initial theory. The aforementioned three types of coding were not sequential analytics processes; instead, they proceed simultaneously (Strauss & Corbin 1998). The following example illustrates the three types of coding. In the open coding, certified health professional s As open coding proceeded to axial coding, similar labels, such as information written by professionals or website run by doctors were compared. The author found that the exp ertise was identified as one of the core concepts. In the selective coding, expertise and other core concepts (e.g., reliability) were grouped under a more abstract category website
66 quality. These core categories (e.g., information quality or website q uality) represent the concepts developed from the open coding and axial coding. In the final phase of coding, core categories were developed to form the theoretical framework of information seeker satisfaction. The detailed steps of analytic process are also demonstrated in Appendix E and in the section of quality of user generated content. Figure 4.1 shows the final information seeker satisfaction model.
67 Figure 4 1 Information Seeker Satisfaction Model
68 Information Seeker Satisfact ion Information seeker satisfaction, in this content websites. The results of the data analysis showed that information seeker satisfaction consists of satisfaction of user ge nerated content, information satisfaction, and website (Strauss & Corbin, 1998, p. 117) The recommendation statements in satisf action of user generated content share the property of addressing features of user generated content as to why a particular website was recommended; the recommendation statements in the information satisfaction category shared the property of addressing in formation features as to why a particular website was recommended; the recommendation statements in the website satisfaction category shared the property of addressing website features as to why a website was recommended. For example, one recommendation r egarding website satisfaction stated: Metromix, because the website has a search tool, whi ch allows information seekers to narrow down information based upon different needs. Hence, this person expresses his/her satisfaction with Metromix because of the search tool provided by Metromix. One recommendation statement can be in more than one cate gory, depending on today site that I have for you is webmd.com. I am sure many of you have heard about WebMD who are always looking for the symptom checker because it's a grea t little tool to try figure out if you are sick, if you have a cold, if you have a flue or whatever as we
69 all know the symptom checker is right here on the front page. WebMD also has a great collection of articles about your health and things about dieting and exercising, and a expressed his/her satisfaction with the system checker (website satisfaction) and with the amount of articles (information satisfaction) on WebMD. Satisfaction of User generated Content Satisfaction of user generated content on content websites was expressed as one of the key factors of information seeker satisfaction. User generated content, in this dissertation refers to the content provided by generated content MedHelp. There is bound to be a forum for abusive realationships or something similar. They h ave forums for everything here. You might meet like minded people who can offer MedHelp has a supportive forum, where users share their experience and support one a nother. Accordingly, an individual would be satisfied and make a recommendation if he believes the quality of user generated content on a content website is high. Information Satisfaction Information satisfaction was identified as one of the key factors perception on the information quality provided by content websites affected their resources to keep you
70 two examples sug gested that, when users bel ieve the quality of information provided by a content website is high, they feel satisfied and recommend that website to others. Website Satisfaction Website satisfaction was identified as one of the key factors of information the website quality affected their satisfaction. For example, one recommendation stated: interested in trying to sell you one way or the other, so their information is often biased to increase sales. Sometimes sales sites have the .org extension, so even .org isn't information on websites with .org or .edu extensions, because these websites are more reliable than those with .com or .org extensions, which are perceived to be many ways to filter what individual recommended CNet, because CNet provides a filter system, which helps information seekers quick ly narrow down specific products based on their preferences. These two examples suggested that, when users believe the website quality is high, they are satisfied and make a recommendation to other information seekers. In the last section, the author had discussed how quality of user generated content, information quality, and website quality relate to satisfaction. In the next section, the author discusses what determine these three types of quality and which factors are important on different types of c ontent website.
71 Quality of User generated Content Quality of user generated content, in this research study, is defined as the quality of the content provided by website users. Six factors, namely direct experience, dependable, neutral, quantity of revi ews, responsiveness, and community support share the same property of quality of user generated content. Table 4. 11 demonstrates how these concepts were emerged. firstha nd exposure to a specific information topic. Direct experience was mentioned as an important factor in all three types of information topics. On product information websites and local information websites, consumers share their direct experience with a p would recommend going to cnet.com and reading some reviews from people who share loc al information, such as education or housing, in forums. People trust the information provided by the local people, because these people have real experience with data.com, specifically t health information from other patients. For example, one recommendation stated: medhelp forums online, particularly the patient one or ovarian cancer (covers other women's issues). There are all sorts of posting on what tests are good for what generated content is va luable to information seekers, because it is written by people who have
72 firsthand experience with a specific product, local businesses or communities, or health problems. Table 4.11 Extracted Concepts of Quality of User generated Content Info Topic Recom mendation Statements Open Coding Axial Coding Category P* make an impact on my purchases Owners of electric juicers may have an issue that would not normally be mentioned in the product experience with produ cts Unbiased product reviews Dependable product reviews Direct experience Neutral Dependable Quality of User generated Content L* rely on people like you and me to provide feedback about our experience wi th local business L out metromix for your restaurant needs. When I'm going to a new neighborhood or a new place, I always check it out there Dependable business reviews Unbiased business reviews H* D.com It's how severe the condition may experience P place where thousands of people submit and rate reviews about Number of produc t reviews Quantity of reviews P answer questions and point out the best and worst features of Responsive answer from users Responsiveness H have a section where you can ask ques tions or read information from others who Responsive answer from users
73 Table 4.11 Extracted Concepts of Quality of User generated Content (Continued) Info. Topic Recommendation Statements Open Coding Axial Coding Category H have found a really good support board online that really webmd and their community section. I like the arthritis/back pain board lots of really good people in the same situation as Emotional support ex perience with health problems Community support Quality of User generated Content H I also rec that you check the website "PatientsLikeMe".There is a section for Bipolar illness and you'll find just tons of info/feedback(even graphs) from patients and their Informational support *P: Pre purchase product information; L: Local information; H: Health information In this paper, dependable refers to the extent to which an information seeker believes she or he can rely on the user gener ated content to make a purchase decision while neutral refers to the degree in which the user generated content provides independent opinions On product information websites and local information websites, consumer reviews are feedbacks provided by cons umers, who have firsthand experience with a product, business, or community. When the author looked closer, she found that consumer reviews were valuable for information seekers, because these reviews are unbiased opinions shared by consumers. In additio n, people also believed that they can depend on the consumer reviews to help them make a purchase decision, because these reviewers are the general public, who are just like them. These reviewers provide diversified opinions based on the truth, and they a re honest and unbiased in writing what they experienced. For example, one recommendation, in the topic of pre purchase check Amazon, Epinions and Price Grabber. Customer reviews really make an impact on
74 my purchases. Owners of electric juicers may have an issue that would not normally be mentioned that they always check consumer reviews before they go to any new restaurant, bar, or event, implying that consumer reviews are perceived to be unbiased check out metromix for your restaurant needs. When I'm going to a new neighborhood or Another concept related to consumer reviews is the quantity of reviews. The quantity of reviews was found to be an important factor in the pre purchase product information seeking. P eople prefer product information websites that contain an extensive amount of consumer reviews. For example, an individual recommended CNet Another individual recommended information website is recommended when it contains a great amount of positive or negative consumer reviews. The quantity of r eviews helps information seekers determine the quality of a product. As one of the recommendation stated Responsiveness was found to be an important factor when people loo k ed for pre purchase product information and health information. In this research paper, responsiveness refers to the extent to which a question is answered by other members, who are enthusiastic about a topic, in a timely manner. Some product informatio n websites allow information seekers to ask additional questions in the comment area or in
75 the bulletin boards. People like product information websites where they can ask questions and receive responses. For example, an individual recommended Viewpoints communities (e.g., chat rooms or bulletin boards) for users to exchange opinions, to share experience, to ask questions, and to provide answers to one another. These virtual communities become important when the information provided by websites is not cann ot find the most recent treatment on a website, he can post questions in the virtual communities and get responsive answers from other users. One recommendation you can In addition to responsiveness, virtual communities, on health information websites, also provide support to information seekers. Health information seekers or their loved ones normally suffer from certain health conditions. Knowing that there is someone who suffers from the same condition and understands your symptoms is important to hea lth information seekers. In virtual communities, information seekers discuss symptoms with other pa tients and receive emotional support from other community members. In addition, information seekers also receive informational support from members who share alternative medications and treatments. Accordingly, community support is one key factor of the quality of user generated content. Community support, in this paper, refers to the perceived emotional and informational
76 need of treatment or want to avoid medical problem s check out Patients Like Me, problems, including what course of treatment they followed, what medications they used, ual recommended Patients LikeMe, because PatientsLikeMe has the community of hudge support and great info. of Information Quality Information quality, in this research study, is defined as the quality of the content provided and presented (e.g ., articles written by professionals) by websites. Six factors, namely completeness, currency, organization, accuracy, demonstration, and understandability share the same property of information quality. Information completeness and currency were mention ed as important factors across all three types of information topics. Information completeness, in this study, refers to the extent to which the information provided by content websites is sufficient in meeting a variety of information needs. When people look for pre purchase product information, health information, or local information, they want to know the full coverage of the topic. For example, in the topic of pre purchase product information, an individual recommended CNet, because list of diets a very complete and thorough list and evaluation of essentially any diet that information, thorough content websites, which cover different cities or cover everything
77 for a part icular city, are useful to information seekers. For example, one recommendation re (link in Resources). Big city or small town, they'll supply the names and contact Information currency, in this paper, refers to the ext ent to which the information provided by websites is perceived to be up to date. Local information seekers need websites that provide the latest local information. For example, an individual eep up with what's purchase product information, information seekers also need the latest product information. For example, guide.com, and reviews.cnet.com provide a wealth of accurate, up to date information about cheap high fast. New drugs, diseases, or research trials are reported on a d aily basis. Health information seekers need the latest information, especially when they think their doctors are not aware of the latest options for treatment of various diseases. One such example ecent ones too. Some doctors still run off of the old levels and they were changed in 2003 and believe me some places the doctors haven't changed what they are judging them by in decades. I know mine ype of information people sought, they liked content websites that provide thorough and up to date information.
78 Information organization, information accuracy, and demonstration were identified as important factors by pre purchase product information seeke rs and health information seekers. Information organization, in this research study, is characterized as the structure and arrangement of the information provided by content websites. Product information websites and health information websites contain a variety of information; as such, people like content websites that provide well organized information. For example, purchase product information, an individual Information accuracy, in this paper, refers to the correctness of the information provided by content websites. Information a ccuracy was identified as an important factor in the topics of pre purchase product information and health information. Pre purchase product information seeking involves the risk of monetary loss, so information seekers need accurate information to help t hem purchase the right product. For example, one Dictionaries, n.d.). This individual recommended CNet, because he or she believed that the information provided by CNet is accurate. Information accuracy was also important to health information seekers, because inaccurate informa tion may result in improper care only like content websites that provide accurate information. One such example is:
79 cover on the internet, particularly something as important as medical info. Check out WEBMD.com, type in STD (sexually health information websites that provide well organ ized and accurate information were perceived to have high information quality. Dictionaries Onlin multimedia files (e.g., videos or images), which are used to support written words. On product information websites, written words, which are static in nature, do not convey much prod uct information to consumers. Consumers like websites that use videos or images to demonstrate products. These videos or images provide visual demonstrations from multiple angles, and help to bring the brick and mortar shopping experience to online infor mation seekers. For example, an individual recommended CNet, because health information, people also need images to help understand skin problems, identify pills by th eir appearance (e.g., shape or imprint), or learn to perform stretches and have a picture of what your bite/ rash looks like. Thats where I usually look up medical information websites and health information websites that use multimedia files to help
80 people better unders tand an information topic were perceived to have high information quality. Understandability, in this paper, refers to the extent to which the information provided and presented by content websites is understandable by people who do not have a medical back ground. Understandability was identified as an important factor by health information seekers. Health information contains medical terms, which are hard to understand for the general public. A health information website would be recommended if the infor mation is written in a simpler language. One recommendation, for example, terms of symptoms, causes, and also explains diagnosis and treatment... no medical background n Website Quality p erception on the features, functionalities, and operation of content websites. Flexibility, expertise, reliability, neutral, and integration share the same property of website quality. Among these five factors, flexibility is the factor identified in all three types of information topics. Flexibility, in this paper, refers to the ability of a system that responds and provides customized search results in meeting a variety of information needs. In the topic of health information, many of the recommendati on statements mentioned the Symptom Checker, which is an interactive system that lists a number of possible causes, diagnosis, or treatments based on the medical symptoms entered. The Symptom Checker is flexible and adaptable enough to accommodate one or more
81 then you can scroll on to any part of the body that you think is wrong, and they will have a list of what could be the problem. You can pick more then one part of the body if information seekers, as it allows users to easily enter symptoms and receive a list of possible causes, diagnosis, or treatments based on the personalized symptoms. Similar to the Symptom Checker, product information websites and local information websites also provide tools that return search results based on user preferences. On product information websites, the search or filter tools allow information seekers to narrow do wn products according to multiple criteria. For example, one person recommended CNet, instance, you can filter laptops based on price, screen size, manufacturer, and a variety of community information across multiple cities. Local information seekers need the search or filter tools to narrow down their search results based on personalized c riteria, such as neighborhood, budget, or interest. For example, an individual recommended Metromix, conte nt websites that are flexible in meeting different information needs. Systems that help narrow down information based on different needs are helpful for information seekers, because they shorten the time of information seeking. Expertise, reliability, and neutral are important to pre purchase product information seekers and health information seekers. On product information websites,
82 editor or expert reviews were perceived to be professional and authoritative, because these editors or experts are subject matter experts. For example, an individual recommended websites that are perceived to be pro fessional, and suggested not to rely on search engines or websites that do not have certified medical doctors. For example, an you should not rely on yahooligans!! unless they information websites that provide information from experts or profe ssionals were perceived to have expertise. In the topic of pre purchase product information and health information, reliable was frequently mentioned in the recommendation statements. To product information websites, reliability means that a website is co nsistently good in providing quality reviews, on which people can rely and trust when making a purchase decision. For checking out the reviews to see how the varying pieces of equipment are rated. I trust the that a website is consistently good in providing health information, on which people can rely and trust. A health website was perce ived to be reliable when 1) the domain name is .edu or .gov, 2) there is no advertisement, or 3) it is referred by professionals (doctors or nurses). For example, an individual recommended websites that use .gov or .edu
83 WebMD, because it was referred by a nurse. Similar to reliability, people who look for pre purchase product information and health information also like websites that are neutral and unbiased. Pre purcha se product information seekers need unbiased product information to help them select the right Information seekers also prefer unbiased health information websites. As one recommendation stated e of the advertisements, and they perceive a website to be biased if there are too many advertisements. As one recommendation stated interested in trying to sell you one way or the other, so their information is often biased to perceived quality of that website is low and information seekers are less likely to use it as an information source. The remaining factor, integration, was identified in the topic of pre purchase product information. Integration, on product information websites, refers to the service that integrates reviews from multiple websites. Information seekers like websites that collect and analyze reviews from multiple websites. One recommendation used the term busy shoppers, who do not have time to browse consumer reviews from multiple websites. For example, an individual recommended Consum erSearch, because ConsumerSearch
84 ConsumerSearch for a comprehensive look at mattress reviews compiled from a variety purchase product information, they service, information seekers are able to save time on v isiting multiple websites. As one surfing multiple sites and more time free to spend playing our never ending backlog of
85 CHAPTER V FINDINGS AND DISCUSSION Quanti tative In Table 5. 1, a summary of the results of the hypotheses testing is presented. Multinomial logistic regression was used to examine the research model. Some factors were found to be more significant than others in Web information source selection, and different combinations of factors influenced the likelihood of using a particular type of information source rather than others. Direct experience, involvement, intensity of information seeking, habit strength of using search engines, perceived search skill, perceived ease of use, and perceived trustworthiness were statistically significant in distinguishing users who utilized three types of information sources. Table 5.1 Hypotheses Summary H1: I nformation seeker s with a higher level of direct experie nce with their preferred information source are more likely to be using specialized website s or search engine s as opposed to centralized website s Supported H2: I nformation seeker s with a higher level of involvement with an activity are more likely to be using specialized website s or search engine s as opposed to centralized website s Partially supported H3: I nformation seeker s who consult great er extent of information sources are more likely to be using content website s as opposed to search engine s Suppo rted H4: I nformation seeker s with stronger habit strength of using search engines are more likely to be using search engine s or centralize website s as opposed to specialized website s Supported H5: I nformation seeker s with greater perceived search skill are more likely to be using search engine s or centralize website s as opposed to specialized website s Supported H6a: I nformation seeker s who perceive a higher level of usefulness with their preferred information source are more likely to be using content website s as opposed to search engine s Not Supported H6b: I nformation seeker s who perceive a higher level of ease of use with their preferred information source are more likely to be using search engine s or specialized website s as opposed to centralized w ebsite s Supported H6c: I nformation seeker s who perceive a higher level of trustworthiness with their preferred information source are more likely to be using content website s as opposed to search engine s Supported
86 Hypothesis 6a was not supported. Cont usefulness was not a significant factor in information source selection, meaning that people who used specialized websites, centralized websites, or search engines have no differences in their perceived usefulnes s with the information sources. Hypothesis 2 was partially supported. The author proposed that information seekers with a higher level of involvement prefer broader and more detailed information, thus search engines or specialized websites are preferred; as such, i nformation seekers with a higher level of involvement with an activity will be more likely to be using specialized websites or search engines as opposed to centralized websites. However, the findings from the data suggested that information see kers, who were more involved with the activity, did not necessarily use specialized websites. Based on this finding, the search engine users were more involved than centralized websites users. Direct experience, in the information source selection model, familiarity with their chosen information source. It was found to be a significant factor in comparing 1) between specialized websites and centralized websites and 2) between search engines and centralized websites. Therefore, hypothesi s 1, in which it was proposed that i nformation seekers with a higher level of direct experience with their preferred information source are more likely to be using specialized websites or search engines as opposed to centralized websites was supported. T he data supported the relatively lower experience than those who use search engines and specialized websites. The rationale for this argument was that, since centralized website s provide a broad range of local information, people may need only partial information from the websites. If an
87 information seeker visits a website only for particular information, the experience he has on that website would be low. Intensity of informat ion seeking, in this study, referred to the extent of information sources used and the amount of information obtained by an information seeker before he made a decision on a particular activity. Intensity of information seeking was found to be a significa nt factor in comparing 1) between search engines and centralized websites and 2) between search engines and specialized websites. When an information seeker wants to seek more intensively, the odds of him choosing a centralized website or a specialized we bsite over a search engine are higher. In short, people prefer content websites over search engines when they seek more intensively. Therefore, hypothesis 3, in which it was proposed that i nformation seekers who consult greater extent of information sour ces are more likely to be using content websit es as opposed to search engines was supported. habit strength of using search engines and with higher perceived search skill were more likely to use search engines or centralized websites than specialized websites. The rationale for this argument was that search engines and centralized websites require either cross or in site searches. On the other hand, the topics being provided by specialized websites are focused on a particular domain; as such, locating needed information on specialized websites does not require any search skill. Individuals who do not develop the strong habit of using search engines or who are not confident in their search skill can use browsing as the information seeking method on specialized websites. Therefore, hypothesis 4, in which it was proposed that i nformation seekers with stronger habit
88 strength of using search engines are more likely to be using sea rch engines or centralize websites as opposed to specialized websites and hypothesis 5, in which it was proposed that i nformation seekers with greater perceived search skill are more likely to be using search engines or centralize websites as opposed to s pecialized websites were supported. Perceived ease of use was found to be a significant factor in comparing 1) between specialized websites and centralized websites and 2) between search engines and argument that centralized websites contain a broad range of information, which decreases the ease of locating needed information. Difficulty of locating needed information on centralized websites also explained why centralized website users had higher se arch skill. Therefore, hypothesis 6b, in which it was proposed that i nformation seekers who perceive a higher level of ease of use with their preferred information source are more likely to be using search engines or specialized websites as opposed to cen tralized websites was supported. Perceived trustworthiness was found to be a significant factor in comparing 1) between search engines and centralized websites and 2) between search engines and specialized websites. Centralized and specialized website us ers perceived a higher level of trustworthiness than search engine users. Even though a majority of the survey respondents used search engines as a local information source, they did not perceive a high level of trustworthiness with the search engines. T rustworthiness can be a competitive advantage for local information websites. Therefore, hypothesis 6c, in which it was proposed that i nformation seekers who perceive a higher level of trustworthiness with their preferred information source are more likel y to be using content websit es as opposed to search engines was supported.
89 Qualitative The major finding of the information seeker satisfaction study was the important role of user generated content to information seekers. Quality of user generated conte nt refers to the quality of the content provided by website users. Six factors, namely direct experience, dependable, neutral, quantity of reviews, responsiveness, and community in the information topic. Direct experience was identified as an important factor in all three types of information topic, suggesting that no matter which type of information people s eek information seekers always like to get information from other users. That is, a content website, which does not contain user ble to information seekers, because these people have firsthand experience with products, local businesses or communities, or health problems. In this paper, neutral refers to the degree in which the user generated content provides independent opinions, wh ile dependable refers to the extent to which an information seeker believes she or he can rely on the user generated content to make a purchase decision. On product information websites and local information websites, information seekers liked consumer re views, because these reviewers were written by the general public, who provided their opinions based on what they experienced. People believed that consumer reviews are free from bias or commercial influences; as such, information seekers are able to rely on consumer reviews to help them make purchase decisions. Since consumer reviews are valuable, a product information website with a
90 large number of reviews is perceived to be a better information provider. As one reviews, the more people that have reviewed a product, the more reliable the overall rating is. If there is only one review and it is a scathing one, that doesn't necessarily mean that the product is a bad one. It may simply be that someone received a dud and they happen to be the only person who has taken the time and effort to go online and review the product. On the other hand, if a product has hundreds or even thousands of reviews, you can be pretty confident that the overall rating listed (based on th Many product information websites and health information websites allow users to ask questions and obtain responses from one another. Information seekers like to have their questions responded in a timely manner ; as such, how responsive other users are was an important factor of the quality of user generated content. In this research paper, responsiveness refers to the extent to which a question is answered by other members, who are enthusiastic about a topic, i n a timely manner. The sixth factor, community support, was identified as an important factor in health information seeking. The data suggested that people liked virtual communities on health information websites, because these virtual communities provid e emotional and informational support. In sum, direct experience, dependable, neutral, quantity of reviews, responsiveness, and community support were factors of quality of user generated content. Information quality and system quality have been thoroughl y studied in the discipline of information systems, and these two types of quality were found to be two important antecedents to user satisfaction. Many of the factors identified in this study were consistent with the previous studies (e.g., DeLone & McLe an, 1992; DeLone &
91 McLean, 2003; Lin, 2010; Wixom & Todd, 2005). Information completeness and information currency emerged in all three types of information topic, suggesting that information seekers always want thorough and up to date information regardl ess of the information topics. Both pre purchase product information seeking and heath information seeking involve a higher risk than local information seeking. People who look for pre purchase product information want to select the right product and redu ce the risk of monetary loss. crucial to pre purchase product information seekers and health information seekers. Moreover, pre purchase product information seekers and health information seekers also like websites that are well organized, so they can easily locate their needed information. Multimedia files are commonly embedded in th e majority of the content websites to support written words. Multimedia files are important to pre purchase product information seekers and health information seekers. On product information websites, multimedia files (images or videos) help consumer vis ualize and examine products up close. On health information websites, images are useful to information seekers in identifying skin problems, in identifying pills by their appearance, or in learning to perform stretches and exercises, because these types o f health information are normally hard to describe or understand with written words. With images, information seekers are able to quickly and easily understand the health information. Since medical terminologies are hard to understand, conveying and trans lating scientific terminologies into an understandable language is critical on health information
92 websites. Health information websites that use medical jargon, complex wordings, or abbreviations may impede health information seekers from using their webs ites. Health information seekers like websites that are written for people who do not have a medical background. This finding is consistent with the previous study (Zahedi & Song, 2008), which suggested that understandability is one key factor of informa tion quality on health information websites. A few implementable suggestions regarding understandability are presented in the recommendation section. functionalities, and operatio n of content websites. The data analysis results showed that no matter which type of information topic people looked for, information seekers liked websites that are flexible in meeting a variety of information needs. Flexibility, in this paper, refers t o the ability of a system that responds and provides customized search results in meeting a variety of information needs. Consistent with the previous studies (Lin, 2010; DeLone & McLean, 2003), information seekers liked systems that can be adapted to mee t different information needs. Accordingly, content websites that are able to help information seekers narrow down a particular piece of information based on different needs were preferred. Expertise, reliability, and neutral are three important factors t o pre purchase product information seekers and health information seekers. This is reasonable, because making the wrong decision based on inaccurate and biased product or health information may lead to monetary or health loss. Product and health informat ion websites that are perceived to have the expertise, to be reliable, and to be neutral from the commercial influences were recommended. In the topic of pre purchase product information, people
93 also liked websites that integrate reviews from multiple web sites, as this integration allows users save time on visiting multiple websites. This finding supported Al Maskari level of satisfaction when 1) less time is required to find the needed information (user effectiveness) and 2) less effort is required to find the needed information (user effort). Finally, the information seeker satisfaction model shows that information topics moderate the effect between antecedents and the t hree types of quality. That is, for different types of information topic, different quality factors are expected. A future research direction is to test whether information topic changes the strength of the effect between quality and quality antecedents.
94 CHAPTER VI CONCLUSION Currently, search engines dominate the information seeking market. To be in a better position to compete with search engines, every content website needs to understand the information seeking behavior. To provide this understandi ng, this dissertation developed an information source selection model and an information seeker satisfaction model. The goal of this dissertation is to help content websites understand 1) how different types of information sources are perceived differentl y and how users utilizing different type of sources are different from one another and 2) the underlying reasons behind information seeker satisfaction and what can be optimized to gain a competitive advantage. In the first study, the author developed the information source selection model based on Johnson and Meischke (1993) 1995) Comprehensive Model of Information Seeking. An online survey was administered to investigate the research questions. Data analysis of 99 responses reveal ed that the most popular tool for local information is search engines. Direct experience, involvement, intensity of information seeking, habit strength of using search engines, perceived search skill, perceived ease of use, and perceived trustworthiness w ere identified to be significant in discriminating three types of information sources. The author then used the grounded theory methodology to develop an information seeker satisfaction model using the IS success model (DeLone & McLean, 2003; Wixom & Todd, 2005) as the framework. A total of 201 recommendation statements were collected: 79 for product information websites, 89 for health information websites, and 33 for local information websites. The information seeker satisfaction model suggested that
95 1) the quality of user generated content is a valuable asset for content websites and 2) different quality factors are expected when people look for different types of information topics. Recommendation Local Information Websites The market of local informati on seeking is one of the most competitive battlefields on the Internet. Currently, search engines are the dominate tools in the market of local information seeking (Graham, 2011). Local information websites face critical challenges, and they are striving to provide a compelling local information seeking experience. In the comparison of user characteristics and user perception among three different types of information sources, namely centralized websites, specialized websites, and search engines, the inf ormation source selection model offered unique insights into local information seeking behavior. The data analysis results indicated that almost 50% of the survey respondents looked for local entertainment information and that the majority of them were ma le. In addition, people who looked for local professional services and store sales information mainly used search engines (70% and 75%, respectively) as information sources; however, search engines were not as dominant in local entertainment information s eeking (44%). Based upon aforementioned findings, new entrants in the local information market, such as news providers, have a better chance to compete with search engines in the market of local entertainment information, and male is the target audience o f this particular market. centralized websites seems to be lower than that of others who used specialized websites
96 or search engines. This is reasonable since information se ekers may visit centralized websites for a particular information topic; as such, their experience with centralized websites was low. Since the information seeker satisfaction model suggests that user generated content is an important factor in informatio n seeker satisfaction, centralized websites, such as Citysearch, may consider providing a social network platform for local communities to connect and engage with one another. Engagements will keep information seekers coming back and will help increase th eir direct experience with centralized websites. In the survey, respondents who looked for a greater extent of information sources liked to use content websites instead of search engines as information sources, suggesting that local information websites ha ve the edge over search engines. Based on this finding, both centralized and specialized websites should provide deeper and more comprehensive local information and attract information seekers who use search engines as information sources. However, ident ifying information seekers who need a greater extent of local information is a challenge. The centralized websites users and search engines users perceived higher search skill and stronger habit strength of using search engines than the specialized website users. This is a positive sign for new entrants (e.g., news providers) who want to develop a centralized local information website. With the growing market of search engines stronger and their search skill will be better; as such, new centralized websites have a competitive edge over specialized websites. New centralized websites should target younger higher) local
97 information seekers, as these two groups perceived higher search skill than other demographic groups did. Ease of use was the key underlying issue for existing centralized websites. Compared with specialized websites and search engines, peo ple found that it was more difficult to find the information they needed on existing centralized websites. The author suspected that ease of finding needed information on centralized websites was related positively to number of information topics. For ex isting centralized websites, the author recommended that they focus on information organization and structure, and then target to attract local information seekers who are not confident in their search skill and who do not have the habit of using search en gines. Even though a majority (58.6%) of the survey respondents used search engines as a local information source, they did not perceive a high level of trustworthiness with search engines. Apparently, information seekers perceived a higher level of trust worthiness with content websites than with search engines in providing local information, implying that trustworthiness is one competitive advantage that local information websites can leverage in the market of local information seeking. One of the new en trants of the local information seeking market is the news providers. They have the advantage over other types of websites. According to a survey conducted by comScore, newspaper websites are the most trustworthy local information sources ( Sigmund, 2010) The advantage of high perceived trustworthiness will help news providers capture the local information seeking market on the Internet.
98 Content Websites The information seeker satisfaction model showed that user direct experience, information completenes s, information currency, and website flexibility are the four factors that constantly emerged across three types of information topics: pre purchase product information, health information, and local information. Therefore, every content website needs to include user generated content, because information seekers like to gain additional insights from other users, who share their experience. With regard to information quality, content websites must ensure that the information they provide is thorough and u p to date. Moreover, content websites need to be flexible in customizing information to meet a variety of needs. More specifically, content websites must provide a system that allows information seekers to filter out and narrow down a particular piece of information quickly. For example, for information seekers who look for wireless routers with different price ranges or manufacturers, product information websites need to provide a system that drill s down into subsets of products based on their informati on needs. The information seeker satisfaction model also demonstrated that information quality and system quality are not the only two factors influencing information seeker satisfaction; quality of user generated content is also an important factor in inf ormation seeker satisfaction. User generated content is important to pre purchase product information seekers and local information seekers, as these two types of information seekers believed that the information shared by other users is free from bias or commercial influences; as such, they can rely on the user generated content to make purchase decisions. Accordingly, content websites need to play a neutral role when it
99 comes to the maintenance of user generated content. That is, content websites need to needs. Once information seekers perceive the quality of user generated content, they will be more willing to come back, to engage, to share their experience and knowledge, and to enrich the content websites. Three of the quality of user generated content factors, namely quantity of reviews, responsiveness, and community support, are directly related to user participation and engagement. In an effort to address th e quality of user generated content, content websites need to encourage participation and engagement. To promote participation, content websites need to provide a user friendly virtual environment, which allows users to easily post and share their experie nce or knowledge. In addition, the virtual environment needs to be adaptable to different devices (e.g., smartphones or tablets). To foster user engagement and involvement, content websites need to encourage social siness review websites, namely Yelp, Citysearch, and Yahoo Local, found that social image is positively correlated with reviewer productivity. Yelp surpassed Citysearch and Yahoo Local in reviews and traffic, because Yelp encourages users to establish a s ocial image and reputation, while Citysearch and offline social events. Each Yelp member has a public profile page that records her activities, including reviews written, number of useful, funny, and cool review votes received, Yelp friends made, and compliment letters displayed. Yelp also recognizes
100 Citysearch and Yahoo Local do not allow social interaction, nor do they attempt to encourage posting us on social image increases not only the Implications Academic Implications The first theoretical implication of this dissertation was the formation of the information source selection mod el. Channel (e.g., Internet, humans, TV, library) selection has been well studied in the information seeking literature (e.g., Kim and Sin, of the information source selec tion literature, no model has been developed for studying different modes of information seeking for the same channel the Web. The contribution of the local information source selection study was to fill the gap in the information seeking literature by synthesizing previous factors suggested by researchers, by developing a model of source selection, and by examining what factors were associated with local information source selection on the Internet. Second, the local information source selection study p rovided the first step in addressing the gap in the literature of Everyday Life Information Seeking by local information needs are growing on the Web. Many practition ers have noticed the potential market for local information seeking and have started to provide local search services. The data analysis showed that search engines were the most popular tool for local information. This does not bode well for new entrants who hope to capture the
101 local information seeking market by providing centralized websites or specialized websites. Comparing with centralized website users, search engine users were able to find the needed information more easily and they were more fami liar with the information sources. Obviously, information seekers still found it difficult to identify their needed information on existing centralized websites. Implementable recommendations were addressed in earlier sections of this paper. Third, the i nformation seeker satisfaction study extended the IS success model (DeLone & McLean, 2003; Wixom & Todd, 2005) to the context of information seeking, and it introduced a third quality the quality of user generated content, to the IS success model. The e xisting studies (e.g., Bliemel & Hassanein, 2006 ; Chen & Cheng, 2009; Cheung & Lee, 2008; Chung & Kwon, 2009; Wang, 2008; Wang & Liao, 2008; Wu & Wang, 2006) that adapted the IS success model have generally not considered how the quality of user generated content impacts user satisfaction. The results of the data analysis suggested that, in addition to information quality and website quality, an additional element quality of user generated content, is also an essential factor of information seeker satisf action. With the proliferation of user generated content on websites, quality of user generated content needs to be included to examine user satisfaction and overall quality of a website. This finding indicated that the high quality of user generated con tent is a valuable asset to content websites, and content websites should use it as an advantage to compete with search engines in the market of information seeking. Fourth, the finding of the information seeker satisfaction study confirmed (19 93) proposition that emotional satisfaction is one of the elements in
102 feelings about the result of information seeking (Applegate, 1993; Bruce, 1998). The data sugge sted that emotional support from the virtual communities on health information websites is an essential factor in information seeker satisfaction. Health information seekers need emotional support from other users who experience similar health problems an d understand how they feel. Knowing that there is someone to talk to is important to health information seekers. In addition, responsiveness is also related to the emotional satisfaction. Product information websites allow information seekers to ask add itional questions in the comment area or in the bulletin boards; health information websites provide virtual communities (e.g., chat rooms or bulletin boards) for users to exchange opinions, to ask questions, and to provide answers to one another. When in formation seekers get responses from other users in a timely manner, they feel satisfied and pleased with the content websites. Managerial Implications Search engines have dominated the market of local information seeking. How and what should local inform ation websites do to compete with search engines? The preference for local information. Based on the results of the data analysis, the author recommended the following implementable solutions to centralized websites: 1) Emphasize on local entertainment information 2) Provide a social network platform for local communities to engage with one another 3) Provide more detailed thorough and up to date local information
103 4) Focus on in formation organization and structure and develop a well structured local repository to attract information seekers who have lower habit strength of using search engines and who perceive lower search skills 5) pete with search engines For specialized websites, they should: 1) Attract male audiences and information seekers who have stronger habit strength of using search engines and who perceive higher search skill 2) Target information seekers who need greater amount of local information 3) Target centralized website users who have a hard time to find needed local information 4) The qualitative study analyzed the recommendation statements to identify essential factors that lead to information seeker satisfaction. The insights from the qualitative study can be employed to guide the website design and maintenance. User direct experience, information completeness, information currency, and website flexi bility are the four fundamental requirements to every content website. To be a better information provider, product information and health information websites need to 1) encourage user engagement and participation to ensure questions are responded in a ti mely manner, 2) focus on the information organization and structure the material logically, 3) ensure information accuracy and provide up to date information, 4) use multimedia files (e.g., images, videos) to help information seekers better understand a pa rticular information topic, 5) provide information written by professionals or experts, 6) ensure reliability (e.g., consistently good in providing quality reviews), and 7) make
104 sure the information is free from bias or commercial influences. In addition, product information websites and local information websites need to play a neutral role when it comes to the maintenance of user generated content. They need to focus on satisfying eds. Health information contains medical terms, which are hard to understand, health information websites need to use simple languages so the general public can easily understand and comprehend the medical information. For example, health information web sites may 1) find alternative words for medical jargon or 2) avoid using abbreviations. As for product information websites, one unique feature that influenced information seeker satisfaction was review integration. Accordingly, product information websi tes may consider integrating reviews from multiple websites and attract information seekers from their competitors. Furthermore, the information seeker satisfaction study also made a managerial implication 1) by helping content websites understand the impo rtance of user generated content and 2) by assessing the role of user generated content in promoting and enhancing information seeker satisfaction. The results indicated that user generated content enriches content websites and increases information seeke r satisfaction. User generated content is valuable to content experience, which is perceived to be neutral and dependable. Thus, on product information websites, the quantity of reviews is positively related to the quality of user generated content. Moreover, user generated content enriches content websites when the On health information websites, user generated content se rves as informational and
105 emotional support to information seekers. In conclusion, user generated content is a valuable asset to every content website, and content websites should take user generated content as an advantage to compete with search engines, which normally do not provide much user generated content. How to leverage this asset and ensure the quality of user generated content are critical to every content website. Limitations The major limitation of the local information source selection study is that the author only examined the information sources that information seekers recalled when they responded the survey. In real life situations, multiple Web information sources may be used during the information seeking process. For example, one may start with visiting information needs are not satisfied. In addition, some people may switch from one type of information source to another to verify that the information t hey found is accurate. Since navigation from one site to another can be done by a few simple clicks, visiting multiple websites to obtain the needed information should be a common practice during the information seeking process. The first research study did not address this issue. Hence, one future research direction is to explore how different types of Web sources are used at the same time for local information. A second limitation of the local information source selection study is its small sample size No definite conclusions can be made until a larger sample size is obtained. With the small sample size (N = 99), it is possible that the author missed some significant factors, particularly in the group of specialized (n = 20) and centralized website ( n = 21)
106 users. Accordingly, natural directions for future research include testing the selection model with a larger data set by gathering more data. focus on how people lo ok for online information via computers. According to the 2013 U.S. Mobile Path to Purchase Study conducted by Telmetrics and xAd (Loechner, 2013), 45% of the consumers searching for local products and services used mobile devices first. How do mobile de vices influence the local information seeking market? First, mobile devices are more portable than computers; as such, information seekers can do a quick search whenever they have information needs (e.g., on daily commute or during lunch breaks). Given t his convenience, there is a great chance that, on mobile devices, people seek more frequently but for shorter periods of time than on computers. Hence, on mobile devices, the need for a responsive website to ease the information seeking process is more ev ident than on computers. If information seekers cannot quickly find what they need, they may not come back. Second, in addition to browsers, people can also use apps to look for local information on mobile devices. The 2013 Local Mobile Search Study (Wu chenich, 2013) showed that 77 million smartphone owners relied on mobile apps to find local information. This impacts the local information seeking market, as such apps are competitors to content websites on mobile devices. Accordingly, one possible futu re research direction is to compare how information sources are used on mobile devices and on computers. A fourth limitation of the local information source selection study is that it only included three types (entertainment, professional services, and sto re sales) of local information, and the data showed that search engines were the most popular information
107 sources for all three types of local information. However, for certain types of local information, specialized websites or centralized websites may b e more popular than search engines. For example, for local education information, specialized websites (e.g., GreatSchool.org) may be more popular than search engines, because people need professional sources. Hence, it is possible that search engines ar e not the most popular tools for other types of local information. However, based on the results of the data analysis, the source selection model is generalizable to other types of information. In this model, only involvement and intensity of information seeking are related to information topics. In terms of involvement and intensity of information see king, no significant difference w as found among the three types of information topics. Other factors, such as direct experience, habit of using search eng ines, perceived search skills, or perceived ease of use, are not influenced by information topics. In addition, the information source selection survey only asked respondents to select one type of information topic they recently sought. An information see ker may use different types of information sources for different information topics. For example, if one person looks for local entertainment information frequently, he may use the content website he normally visits; however, for unfamiliar topics, he may use search engines. Accordingly, one future research direction is to examine whether frequency of information seeking is a significant factor in source selection. With regard to the information seeker satisfaction study, the findings are constrained by s everal limitations. First, even though the online recommendation world experience, they were limited by the lack of further clarification. That is, the author cannot explore
108 issues in depth, nor can she clarify recommendation statements through further questions, as the people who posted the recommendation statements were unidentifiable. For example, in the topic of pre purchase product information, many people recommended websites whe re they can ask questions and obtain responses from other users, implying that information seekers like to get responsive answers. However, the author was not able to ask any further questions to explore whether social interaction is important to pre purc hase product information seekers. Consequently, it is possible that the author missed some important factors. Second, the information seeker satisfaction model was developed from the qualitative data. The findings in the information seeker satisfaction m odel were not statistically verified, and may not be generalizable in another setting. Consequently, a needed expansion of this study is the verification of the information seeker satisfaction model. To verify the effect of user generated content, resear chers may use one of the information topics or a combination of the two information topics as a starting point for a broader examination of other information topics. Third, the information seeker satisfaction study was conducted when branding and reputatio n management companies were not commonly hired by businesses. Branding and reputation management companies provide services (e.g., writing phony reviews) that help businesses improve their client rating and reviews. Phony reviews not only decrease the qu ality of user generated content, but also impact the credibility of content websites. According to a recent study released by the Maritz Research (Ensing, 2013), one in four of the survey respondents believed that online review websites are unfair and the reviews are biased or fake. Thus, how consumers react to the growing
109 number of phony reviews and how phony reviews affect information seeker satisfaction will be an important research direction in the future.
110 R EFERENCES Agarwal, R., & Karahanna, E. ( 2000). Time flies when you're having fun: Cognitive absorption and beliefs about information technology usage. MIS Quarterly 24 (4), 665 694. Agosto, D.E., & Hughes Hassell, S. (2006). Toward a model of the everyday life information needs of urban teenager s, part 1: Theoretical model. Journal of the American Society for Information Science and Technology 57 (10), 1394 1403. Al Maskari, A., & Sanderson, M. A. (2010). Review of factors influencing user satisfaction in information retrieval. Journal of the Ame rican Society for Information Science and Technology 61 (5), 859 868. Anderson, C.J., Glassman, M., McAfeea, R.B., & Pinelli, T. (2001). An investigation of factors affecting how engineers and scientists seek information. Journal of Engineering and Technol ogy Management 18 (2), 131 155. Applegate, R. (1993). Models of user satisfaction: Understanding false positives. RQ 32 (4), 525 539. Ariyachandra, T., & Watson, H. (2010). Key organizational factors in data warehouse architecture selection. Decision Suppo rt Systems 49 200 212. Armstrong, J.S., & Overton, T.S. (1977). Estimating nonresponse bias in mail surveys. Journal of Marketing Research 14(3), 396 402. Arora, R. (1982). Validation of an S O R model for situation, enduring and response components of involvement. Journal of Marketing Research 19 (4), 505 516. Universal Access in the Information Society 4 (1) (2005), 67 81. Battacherjee, A. (2002). Individual trust in online firms: Scale development and initial test. Journal of Management Informalion Systems 19 (1), 211 241. Beatty, S.E., & Smith, S.M. (1987). External search effort: An investigation across several product categories. Journal of Consumer Research 14 (1), 83 9 5. media ad view reports reveal where local ad dollars are flowing. Retrieved from http://www.biakelsey.com/Company/Press Releases/120517 Local Ad Spending Moving More to Mobile and Online.asp
111 Bliemel, M., & Hassanein, K. (2006). Consumer satisfaction with online health information retrieval: A model and empirical study. e Service Journal 5 (2), 53 83. Branch, J.L. (2003). Nontraditional undergraduates at home, work, and sc hool: An examination of information seeking behaviors and the impact of information literacy instruction. Research Strategies 19 (1), 3 15. Bruce, H. (1998). User satisfaction with information seeking on the Internet. Journal of the American Society for In formation Science 49 (6), 541 556. Bunker, M.P., Rajendran, K.N., & Corbin, S.B. (2013). The Antecedents of Satisfaction for Facebook "Likers". Marketing Management Journal 23 (2), 21 34. Buttner, O. B., & Goritz, A. S. (2008). Perceived trustworthiness o f online shops. Journal of Consumer Behaviour 7(1), 35 50. Chang, S. J.L. (2009). Information research in leisure: Implications from an empirical study of backpackers. Library Trends 57 (4), 711 728. Chatzoglou, P.D., & Vraimaki, E. (2010). Investigating Internet usage as innovation adoption: A quantitative study. Journal of Information, Communication and Ethics in Society 8 (4), 338 363. Chen, C. W.D., & Cheng, C. Y.J. (2009). Understanding consumer intention in online shopping: A respecification and vali dation of the DeLone and McLean Model. Behaviour & Information Technology 28 (4), 335 345. Cheung, C.M.K., & Lee, M.K.O. (2008). The structure of Web based information systems satisfaction: Testing of competing models. Journal of the American Society for I nformation Science and Technology 59 (10), 1617 1630. Chung, N., & Kwon, S.J. (2009). Effect of trust level on mobile banking satisfaction: A multi group analysis of information system success instruments. Behaviour & Information Technology 28 (6), 549 562 Clarke, K., & Belk, R.W. (1978). The effects of product involvement and task definition on anticipated consumer effort. Advances in Consumer Research 6 (1), 313 318. Cole, C.A., & Balasubramanian, S.K. (1993). Age differences in consumers' search for inf ormation: Public policy implications. Journal of Consumer Research 20(1), 157 169. Compeau, D. R., & Higgins, C. A. (1995). Computer self efficacy: Development of a measure and initial test. MIS Quarterly 19 (2), 189 211.
112 comScore. (2012). 5th annual 15mi les/localeze local search usage study. Retrieved from http://www.localsearchstudy.com/Local%20Search%20Study%202012.pdf Cooney, M. (2008, November 18). 20+ Product Review Sites. Retrieved from http://smartlifeblog.com/20 plus product review sites/ Crouch, G.I., & Louviere, J.J. (2004). The determinants of convention site selection: A logistic choice model from experimental data. Journal of Travel Research 43 (2), 118 130. Cupini, L.M, Pasqualetti, P., Diomedi, M., Vernieri, F., Silvestrini, M., Rizzato, B., Ferrante, F., & Bernardi, G. (2002). Carotid artery intima media thickness and lacunar versus nonlacunar infarcts. Stroke 33 689 694 Davis, F.D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quar terly 13 (3), 319 340. DeLone, W.H., & McLean, E.R. (1992). Information systems success: The quest for the dependent variable. Information Systems Research 3 (1), 60 95. DeLone, W.H., & McLean, E.R. (2003). The DeLone and McLean model of information system s success: A ten year update. Journal of Management Information Systems 19 (4), 9 30. Demonstration. (n.d.). In Cambridge Dictionaries Online Retrieved from: http://dictionary.cambridge.org/dictionary/american english/demonstration_1?q=demonstration eBiz MBA. (2011). Top 15 Most Popular Health Websites. Retrieved from http://www.ebizmba.com/articles/health websites Encyclopedia. (n.d.). Google generation. Retrived from http://www.azencyclopedia.info/g/1019243_Google_generation/ Ensing, D. (2013, July). Customer rating and reviews site: An upcoming crisis of confidence? Retrieved from http://www.maritzresearch.com/~/medi a/Files/MaritzResearch/Whitepapers/Custo mer Rating and Reviews Site_rev.pdf Erdem, T., Keane, M.P., nc, T.S., & Strebel, J. (2005). Learning about computers: An analysis of information search and technology choice. Quantitative Marketing and Economics 3 207 246. Ernest, D.J., Level, A.V., & Culbertson, M. (2005). Information seeking behavior for recreational activities and its implications for libraries. Reference Services Review 33 (1), 88 103.
113 Fallows, D. (2005, January 23) Search engine users. Retriv ed from http://www.pewinternet.org/~/media//Files/Reports/2005/PIP_Searchengine_users .pdf.pdf Fallows, D. (2008, August 6). Search engine use. Retrived from http ://www.pewinternet.org/~/media//Files/Reports/2008/PIP_Search_Aug08.pdf. pdf information sources. Information Processing and Management 40 (3), 563 581. Flanagin, A.J. & Metzger, M.J. (2000). Perceptions of Internet information credibility. Journalism & Mass Communication Quarterly 77 (3), 515 540. Fornell, C., & Larcker, D.F. (1981). Evaluating structural equation models with unobservable variables and measurement err or. Journal of Marketing Research 18 (1), 39 50. Fu, L., & Salvendy, G. (2002). The contribution of apparent and inherent usability to a user's satisfaction in a searching and browsing task on the Web. Ergonomics 45 (6), 415 424. Gerstberger, P.G., & Allen T.J. (1968). Criteria used by research and development engineers in the selection of an information source. Journal of Applied Psychology 52 (4), 272 279. Given, L.M. (2002). The academic and the everyday: Investigating the overlap in mature undergraduat es' information seeking behaviors. Library & Information Science Research 24 (1), 17 29. Google. (n.d.). Search operators. Retrieved from https://support.google.com/websearch/answer/13 6861?hl=en Graham, J. (2011, February 15). Yellow Pages battles Google in local search. Retrived from http://usatoday30.usatoday.com/tech/news/2011 02 15 yellowpages 16_ST_N.htm http://www.newspapernext.org/Making_the_Leap.pdf Griffin, R., Dunwoody, S., & Neuwirth, K. (1999). Proposed model of the relationship of risk information seeking and processing to the development of preventive behaviors. Environmental Research 80 (2), 230 254. Hair, J.F., Black, W.C., Babin, J.B., Anderson, R.E., & Tatham, R.L. (2006). Multivariate data analysis: Sixth edition New Jerse y: Pearson Education.
114 Hardy, A.P. (1982). The selection of channels when seeking information: Cost/benefit vs. least effort. Information Processing & Management 18 (6), 289 293. Hertzum, M. (2002). The importance of trust in software engineers' assessment and choice of information sources. Information and Organization 12 (1), 1 18. Johnson, J.D., Donohue, W.A., Atkin, C.K., & Johnson, S. (1995). A comprehensive model of information seeking: Tests focusing on a technical organization. Science Communication 16 (3), 274 303. Johnson, J.D., & Meischke, H. (1993). A comprehensive model of cancer related information seeking applied to magazines. Human Communication Research 19 (3), 343 367. Jorosi, B.N. (2006). The information needs and information seeking behavio urs of SME managers in Botswana. Libri 56 (2), 97 107. Julien, H., & Michels, D. (2000). Source selection among information seekers: Ideals and realities. Canadian Journal of Information and Library Science 25 (1), 1 18. Kim, K. S., & Sin, S. C. J. (2011). Selecting quality sources: Bridging the gap between the perception and use of information sources. Journal of Information Science 37 (2).178 188. Kim, S. S., & Malhotra, N. K. (2005). A longitudinal model of continued IS use: An integrative view of four m echanisms underlying postadoption phenomena. Management Science 51 (5), 741 755. Kuruzovich, J., Viswanathan, S., Agarwal, R., Gosain, S., & Weitzman, S. (2008). Marketspace or marketplace? Online information search and channel outcomes in auto retailing. Information Systems Research 19 (2), 182 201. Kwasitsu, L. (2003). Information seeking behavior of design, process, and manufacturing engineers. Library & Information Science Research 25 (4), 459 476. Laroche, M., Cleveland, M., & Browne, E. (2004). Explor ing age related differences in information acquisition for a gift purchase. Journal of Economic Psychology 25 (1), 61 95. Laurent, G., & Kapferer, J. N. (1985). Measuring consumer involvement profiles. Journal of Marketing Research 22 (1), 41 53. Lemire, M ., Par, G., Sicotte, C., & Harvey, C. (2008). Determinants of Internet use as a preferred source of information on personal health. International Journal of Medical Informatics 77 (11), 723 734.
115 Liao, H., & Lu, H. (2008). The role of experience and innova tion characteristics in the adoption and continued use of e learning Websites. Computers & Education 51 (4), 1405 1416. Li, D., Browne, G.J., & Wetherbe, J.C. (2006). Why do Internet users stick with a specific Web site? A relationship perspective. Interna tional Journal of Electronic Commerce 10(4), 105 141. Li, N., & Kirkup, G. (2007). Gender and cultural differences in Internet use: A study of China and the UK. Computers & Education 48 (2), 301 317. Limayem, M., & Hirt, S.G. (2003). Force of habit and in formation systems usage: Theory and initial validation. Journal of the Association for Information Systems 4 (1), 65 95. Limayem, M., Hirt, S.G., & Cheung, C.M.K. (2007) How habit limits the predictive power of intention: The case of information systems co ntinuance. MIS Quarterly 31 (4), 705 738. Lin, H. F. (2010). An investigation into the effects of IS quality and top management support on ERP system usage. Total Quality Management 21 (3), 335 349. Loechner, J. (2013, June 3). Mobile shoppers want local r elevancy. Retrieved from http://www.mediapost.com/publications/article/201595/mobile shoppers want local relevancy.html#axzz2XEh70Lmc Lueg, J.E., Moore, R., & Warkentin, M. (2003). Patient health information search: An exploratory model of Web based search behavior. Journal of End User Computing 15 (4), 49 61. Lymperopoulos, C., & Chaniotakis, I.E. (2008). Price satisfaction and personnel effi ciency as antecedents of overall satisfaction from consumer credit products and positive word of mouth. Journal of Financial Services Marketing 13 (1), 63 71. McKenzie, P.J. (2003). A model of information practices in accounts of everyday life information seeking. Journal of Documentation 59 (1), 19 40. Mittal, B. A. (1995). Comparative analysis of four scales of consu mer involvement. Psychology & Marketing 12 (7), 663 682. Monoi, S., O'Hanlon, N., & Diaz, K.R. (2005). Online searching skills: Development of an inventory to assess self efficacy. Journal of Academic Librarianship 31 (2), 98 105.
116 Morrison, E.W., & Vancouv er, J.B. (2000). Within person analysis of information seeking: The effects of perceived costs and benefits. Journal of Management 26 (1), 119 137. Olmstead, K., Mitchell, A., & Rosenstiel, T. (2011, May 9). Navigating News Online. Retrieved from http://www.journalism.org/analysis_report/demographics Spot on. (n.d.). In Oxford Dictionaries Retrieved from: http://www.oxforddictionaries.com/us/definition/american_english/spot on?q=spot+on PP, R. (2009, March 2). Top 10 Product Review Websites. Retrieved from http://t echpp.com/2009/03/02/top 10 product review websites/ Purcell, K., Brenner, J., & Rainie, L. (2012, March 9). Search engine use 2012. Retrived from http://pewinternet.org/~/media/Files/Reports/2012/PIP_Search_Engine_Use_2012 .pdf Rainie, L. (2005, November 20). Big Jump in Search Engine Use. Retrived from http://www.pewinternet.org/Reports/2005/Big jump in search engine use/Data Memo.aspx Rains, S.A. (2007). Perceptions of traditional information sources and use of the World Wide Web to seek health information: Findings from the health information national trends survey. Journal of Health Communication 12 (7), 667 680. Rieh, S.Y. (2003). Investigating Web sear ching behavior in home environments. Proceedings of the American Society for Information Science and Technology 40 (1), 255 264. Rieh, S.Y. (2004). On the Web at home: Information seeking and Web searching in the home environment. Journal of the American S ociety for Information Science and Technology 55 (8), 743 753. Riggs, D. (2008). Using multinomial logistic regression analysis to develop a model of International Journal of Emerging T echnologies and Society 6(2), 106 123. Ringle, C .M., Wende, S., & Will, A. (2005 ). SmartPLS 2.0 (M3) Beta, Hamburg. Retrieved from http://www.smartpls.de. Rosenstiel, T., Mitchell, A., Purcell, K., & Rainie, L. (2011, September 26). How people learn about their local community. Retrieved from http://www.journalism.org/2011/09/26/local news/
117 Santosa, P.I., Wei, K.K., & Chan, H.C. (2005). User involvement and user satisfaction with information seeking activity. European Journal of Information Systems 14 (4), 361 370. Savolainen, R. (1995). Everyday life information seeking: Approaching information seeking in the context of "way of life". Library & Information Science Research 17 (3), 259 294. Savolainen, R. (1999). The role of the Internet in information seek ing. Putting the networked services in context. Information Processing & Management 35 (6), 765 782. Savolainen, R. (2002). Network competence and information seeking on the Internet: From definitions towards a social cognitive model. Journal of Documentat ion 58 (2), 211 226. Savolainen, R. (2007). Information source horizons and source preferences of environmental activists: A social phenomenological approach. Journal of the American Society for Information Science and Technology 58 (12), 1709 1719. Savola inen, R. (2008). Source preferences in the context of seeking problem specific information. Information Processing and Management 44 (1), 274 293. Savolainen, R., & Kari, J. (2004a). Conceptions of the Internet in everyday life information seeking. Journal of Information Science 30 (3), 219 226. Savolainen, R., & Kari, J. (2004b). Placing the Internet in information source horizons: A study of information seeking by Internet users in the context of self development. Library & Information Science Research 2 6 (4), 415 433. Shim, W. (2002). Probing the process of information source selection using palm pilots: A pilot study. Proceedings of the American Society for Information Science and Technology 39 (1), 220 227. Sigmund, J. (2010, February 24). Newspaper Web sites continue to be the most valued local news and information sites online. Retrieved from http://cms.naa.org/News and Media/Press Center/Archives/2010/Newspaper websites continue to be the most valued local news and information sites online.aspx Solis, B. (2009, October 28). Social media accounts for 18% of information search market. Retrieved from http://socialmediatoday.com/SMC/137537 So,W. C., Wong, T. N., & Sculli, D. (2005). Factors affecting intentions to purchase via the Internet. Industrial Manag ement & Data Systems 105 (9), 1225 1244.
118 Sterling, G. (2010, December 9). Survey: Local reviews gaining in importance. Retrieved from http://searchengineland.com/survey local reviews gaining in importance 58391 Strauss, A., & Corbin, J.M. (1998). Basics of qualitative research: Techniques and procedures for developing grounded theory (2nd ed.) Thousand Oaks, CA: SAGE Publications. Tabachnick, B.G., & Fidell, L.S. (1996). Using multivariate statistics New York: Harper Collins College Publishers. Taha, J., Sharit, J., & Czaja, S. (2009). Use of and satisfaction with sources of health information among older Internet users and nonusers. The Gerontologist 49 (5), 663 673. Teo, T., & Lim, V. (2000). Gender differences in Internet usage and task preferences. Beh aviour and Information Technology 19 (4), 283 295. Teo, T.S.H., Srivastava, S.C., & Jiang, L. (2008 9). Trust and electronic government success: An empirical study. Journal of Management Information Systems 25 (3), 99 131. Times Mirror Center. (1995, Octob er 16). Technology in the American household: Americans going online...Explosive growth, uncertain destinations. Retrived from http://people press.org/files/legacy pdf/136.pdf Tustin, N. (2010). The role of patient satisfaction in online health information seeking. Journal of Health Communication 15 (1), 3 17. van der Heijden, H. (2003). Factors influencing the usage of Websites: The case of a generic portal in the Netherlands. Information and Management 40 (6), 541 549. Verplanken, B., Aarts, H., & van Kni ppenberg, A. (1997). Habit, information acquisition, and the process of making travel mode choices. European Journal of Social Psychology 27 (5), 539 560. Verplanken, B., & Orbell, S. (2003). Reflections on past behavior: A self report index of habit stren gth. Journal of Applied Social Psychology 33 (6), 1313 1330. Wall, A. (n.d). History of search engines: From 1945 to Google 2007. Retrived from http://www.searchenginehistory.com Wang, Y. S. (2008). Assessing E commerce systems success: A respecification a nd validation of the DeLone and McLean Model of IS success. Information Systems Journal 18 (5), 529 557.
119 Wang, Y. S., & Liao, Y. W. (2008). Assessing eGovernment systems success: A validation of the DeLone and McLean model of information systems success. G overnment Information Quarterly 25 (4), 717 733. Wang, Z. (2010). Anonymity, social image, and the competition for volunteers: A case study of the online market for reviews. The B.E. Journal of Economic Analysis & Policy 10 (1), 1 33. Weigley, S. (2013, Ma rch 9). 10 Web Sites Where Surfers Spend the Most Time. Retrived from http://www.usatoday.com/story/money/business/2013/03/09/10 web sites most visited/1970835/ Wixom, B.H., & Todd, P.A. (2005). A theoretical integration of user satisfaction and technology acceptance. Information Systems Research 16 (1), 85 102. Woudstra, L., & van den Hooff, B. (2008). Inside the source selection process: Selection criteria for human information sources. Information Processing and Management 44 (3), 1267 1278. Woudstra, L. van den Hooff, B., & Schouten, A.P. (2012). Dimensions of quality and accessibility: Selection of human information sources from a social capital perspective. Information Processing and Management 48 (4), 618 630. Wu, J. H., & Wang, Y. M. (2006). Measuri ng KMS success: A respecification of the Information & Management 43 (6), 728 739. Wu, Y. T., & Tsai, C. C. (2007). Developing an information commitment survey for assessing students' Web information searching strategies and eval uative standards for Web materials. Educational Technology & Society 10 (2), 120 132. Wuchenich, N. (2011, June 28). 5 Considerations for connecting with local searchers. Retrived from http://searchengineland.com/5 considerations for connecting with local searchers 83201 Wuchenich, N. (2013, April 1). Mobile growth highlights importance of comprehensive local strategy. Retrived from http://searchengineland .com/mobile growth highlights importance of comprehensive local strategy 153617?utm_medium=email&utm_campaign=Post%20Blast%20%28sai%29%3 A%20BII%20MOBILE%20INSIGHTS%3A%20Windows%20Phone%20Gains% 20Ground%20In%20Some%20Markets&utm_source=Triggermail&utm_term= M obile%20Insights Xu, Y., Tan, C.Y., & Yang, L. (2006). Who will you ask? An empirical study of interpersonal task information seeking. Journal of the American Society for Information Science and Technology 57 (12), 1666 1677.
120 Yang, S. C. (1997). Informati on seeking as problem solving using a aualitative approach to uncover the novice learner's information seeking processes in a perseus hypertext system. Library & Information Science Research 19 (1), 71 92. Yoo, E. Y., & Kim, K. S. (2008). Undergraduates' s ource selection behavior: A correspondence analysis. Proceedings of the American Society for Information Science and Technology 45 (1), 1 3. Zahedi, F. M., & Song, J. (2008). Dynamics of Trust Revision: Using Health Infomediaries. Journal of Management Inf ormation Systems 24 (4), 225 248. Zhang, J.J., Fang, X., & Sheng, O.R.L. (2006 7). Online consumer search depth: Theories and new findings. Journal of Management Information Systems 23 (3), 71 95.
121 APPENDI X Appendix A. Forum Lists http://www.gocorgi.com/forums/ http://aspecialthing.com/forum/ http://bbs.clubplanet.com/ http://bow.fishingcountry.com/bfcforums/index.php http://citywidetalk.com/ http://community.myhomeideas.com/ http://connectedcommunities.us/ http://forum.aboutnewjersey.com/ http://forum.freeadvice.com/ http:/ /forum.notebookreview.com/ http://forumnyc.com/ http://forums.alpinezone.com/ http://forum.gon.com/ http://forums.dealofday.com/ http://forums.mtbr.com/ http://forums.outdoorreview.com/ http://forums.wat erskimag.com/ http://forums.somd.com/ http://knoxmoms.com/data/assets/boards/knoxmoms/index.php http://metaldetectingforum.com/ http://nysgo.com/forums/newyork/ http://orlandoforums.com/forum/ http ://forums.roadbikereview.com/ http://straightrazorplace.com/forums/ http://talk.sheknows.com/ http://thisbluemarbl e.com/ http://whiteblaze.net/forum/index.php?s=8ebb740f96e90ea8cfd0b629eab05981 http://smdailyjournal.com/f orum/index.php http://talk.baltimoresun.com/ http://wirednewyork.com/forum/ http://www.discussny.com /new york forum.php http://www.fishingmessageboards.com/ http://www.flaglerchat.com/forum/ http:// www.flagleronline.com/forum/ http://www.hookupsportfishing.com/forum/ http://www.tampaforums.com/forums/index.php http://www.thecouponcupboard.com/ http://www.unicyclist.com/forums/
122 Appendix A. Continued http://www.vickysdeals.com/forums/ http://www.womens health.com/boards/ http://www.buckeyeplanet.com/forum/ http://www .city data.com/forum/#u s forums http://www.mototips.com/forums.php http://www.spearfishingplanet.com/ http: //www.styleforum.net/ http://www.mothering.com/discussions/ http://www.planetsidexside.com/forum.php http://www.pwcforum.com/forums.php http://www.resellerratings.com/forum/ http://www.resellerratings.com/forum/index.php http://www.salmoncrazy.com/vforums/ http://forum.appliancepartspros.com/ http://forums.slickdeals.ne t/index.php http://loudounmoms.com/forum/ http://www.corvetteactioncenter.com/forums/ http://www.cycl ingforums.com/ http://www.frisco online.com/forums/index.php http://www.getbentsportfishing.com/forum/ http://www.worldlawdirect.com/forum/ http://www.grom.tv/forums/ http://www.hawaiitalks.net/ http://www.hotcouponworld.com/forums/index.php http://www.njguido.com/forum/index.php?act=idx http://www.southbaychat.com/ http://community.cookinglight.com/ http://www.discussionforums.us/forum/ http://www.freesamplesite. com/ydf/index.php http://www.njflyfishing.com/vBulletin/#main forums fly fishing state http://www.tennisw.co m/forums/index.php http://www.mycoupons.com/boards/ http://www.orangepower.com/ http://www.talkdelaware.c om/forums.php http://www.talkpa.net/forums.php http://www.talkvirginia.net/forum.php http://www.survivaltopics.com/forums/ http://www.yosemitearea.com http://www.coloradomoms.com http://www.utahonthefly.com
123 Appendix A. Continued http://www.windychat.com http://northwestfishing.info
124 Appendix B. Survey Items Direct Experience EXP 1 : How often do you go to the Website 1 ? Seldom __:__:__:__:__ Very often EXP 2 : How experienced are yo u at using the Website ? Slightly experienced __:__:__:__:__ Very experienced EXP 3 : How familiar are you with the Website ? Not at all familiar __:__:__:__:__ Extremely familiar Involvement INV 1 : When shopping at retail stores 2 what's most like you? I do n ot care at all about __:__:__:__:__ I care a great deal about INV 2 : When shopping at retail stores how important is it to you that you get the best retail price ? Not a t all important __:__:__:__:__ Extremely important INV 3 : How concerned are you about saving money when shopping at retail stores ? Not at all concerned __:__:__:__:__ Very much concerned Intensity of Information Seeking INTEN 1 : I usually visit multiple Web sites to find out which stores are having sales on items I want to buy Not at all like me __:__:__:__:__ Exactly like me INTEN 2 : I usually use various Websites to review sales information before choosing a retail store Not at all like me __:__:__:__:__ E xactly like me INTEN 3 : I usually look for as much information as I can get from various Websites about stores sales before going shopping Not at all like me __:__:__:__:__ Exactly like me Habit Strength of Using Search Engines HABIT 1 : How frequently do y ou use search engines (such as Google, Yahoo, AOL, etc.) when you need to find information on the Web? __:__:__:__:__ Always HABIT 2 : How long have you been using search engines to look for information? __:__:__:__:_ More than 10 years HABIT 3 : When I need to find information on the Web, I usually start with a search engine. Not at all like me __:__:__:__:__ Exactly like me HABIT 4 : When I need to find information on the Web, I automatically resort to a search engine. Not at all like me __:__:__:__:__ Exactly like me
125 Appendix B. Continued Perceived Search Skill SKILL 1 : I can use the most appropriate keywords or phrases when looking for information I need. Not at all like me __:__:__:__:__ Exactly like me SKILL 2 : I can use alternative terms (e.g., synonyms) when looking for information I need. Not at all like me __:__:__:__:__ Exactly like me SKILL 3 : I usually know when to browse the search results and when to search further. Not at all like me __:__:__:__:__ Exactly lik e me SKILL 4 : When a relevant term shows up in my search results, I can recognize it and use it to do further searches if desired. Not at all like me __:__:__:__:__ Exactly like me SKILL 5 : I can narrow or broaden my search to get the appropriate amount of i nformation. Not at all like me __:__:__:__:__ Exactly like me Perceived Usefulness USEFUL 1 : With respect to store sales information that Website provides related links to other Websites Strongly disagree __:__:__:__:__ Strongly agree USEFUL 2 : That Websi te helps me find further information about store sales Strongly disagree __:__:__:__:__ Strongly agree USEFUL 3 : With respect to store sales information the content provided by that Website fits my information searching goal. Strongly disagree __:__:__:__ :__ Strongly agree USEFUL 4 : With respect to store sales information that Website provides the information I need. Strongly disagree __:__:__:__:__ Strongly agree Perceived Ease of Use EASE 1 : I am good at using that Website to find information about store sales Strongly disagree __:__:__:__:__ Strongly agree EASE 2 : I can quickly find the information I need on that Website Strongly disagree __:__:__:__:__ Strongly agree EASE 3 : It is easy to navigate that Website Strongly disagree __:__:__:__:__ Strongly agree EASE 4 : It is easy for me to find information about store sales on that Website Strongly disagree __:__:__:__:__ Strongly agree
126 Appendix B. Continued Perceived Trustworthiness TRUST 1 : That Website has the ability to offer complete information abo ut store sales Strongly disagree __:__:__:__:__ Strongly agree TRUST 2 : That Website offers objective information about local stores and shops Strongly disagree __:__:__:__:__ Strongly agree TRUST 3 : That Website keeps users' best interests in mind in off ering information about store sales Strongly disagree __:__:__:__:__ Strongly agree TRUST 4 : Overall, that Website is trustworthy. Strongly disagree __:__:__:__:__ Strongly agree _____________________________________________________________________________ *: Item removed due to low loading 1 Italicized words: Actual wording changes depending on type of information source chosen by the respondent. Not italic in the original survey. 2 Underlines: When a different information topic was chosen by the responde nt, wording changed accordingly. Not underlined in the original survey.
127 Appendix C. Factor Loadings
128 Appendix D. List of Website Names Name of Product Information Websites Name of Health Information Websites Name of Local Information Websites Buzzilli ons Drugs.com AmericanTowns Cnet EasyDiagnosis AskCity ConsumerSearch EverydayHealth Bing local ConsumerReports Go Ask Alice City data Crowdstorm HealthGrades CityGuide Digital Photography Review Healthline AOL City's Best Epinions Mayo Clinic CitySe arch Kelley Blue Book MedHelp Local.com MouthShut MedicineNet MerchantCircle Reviews.Omgili PatientsLikeMe Metromix PCMag RealAge Windows Live Local PEbuzz RightHealth Yahoo Local RateItAll RxList Yelp Reevoo The U.S. National Institutes of Health Retrevo WebMD Review Centre WrongDiagnosis ReviewGist Yahoo Health Sazze SmartRatings TestFreaks TopTenReviews Tribesmart TrustedReviews WIRED ZDNet
129 Appendix E. Steps of Coding with Examples Recommendation Statement Open Cod ing Axial Coding Selective Coding like Metromix rely on people like you and me to provide feedback about our experience with products Information Seeker Satisfaction: Satisfaction of user generated content Information s atisfaction Website satisfaction Quality of User generated Content: Direct experience Dependable Neutral Quantity of reviews Responsiveness Community support reviews really make an impact on my purchases. Owners of electric juicers may ha ve an issue that would not normally be mentioned in the product experience with products Unbiased product reviews Dependable product reviews epinions.com, a place where thousands of people submit and rate reviews about t housands of Number of product reviews Thorough information reviewers answer questions and point out the best and worst features Responsive answer from users Unbiased reviews experience with products
130 A ppendix E Continued Patients Like Me, where you can see how others faced the medical dealing with handled their problems, including what course of treatment they followed, what medications they used, and how they avoided complic ations Informational support direct experience with health problems really good support board online that really has helped a out webmd and their community section. I like the arthritis/back pain board lots of really good people in the same situation as Emotional support direct experience with health problems
131 Appendix E Continued list of diets a very complete and thorough list and evaluation of es sentially any diet that you may have heard of, or will hear of, in the Complete and thorough information Information Quality: Completeness Currency Organization Accuracy Demonstration Understandability happenings a nd bar deals, be sure to check out Metromix, Centerstage, or The Local Tourist. Up to date information a list of 76 possible symptoms organized in an alphabetical list and a common symptoms Organized information provider reviews. Reputable sites like freelist.com, theispguide.com, and reviews.cnet.com provide a wealth of accurate, up to date information about cheap high speed Internet Accurate information Up to date information s ometimes they have pictures and other descriptions that you may find very Multimedia files to support written words good and explains diseases in terms of symptoms, causes, and also explains diagnosis and treatment... no medical background needed. Information is easy to understand
132 Appendix E Continued and offers you many ways to filter what type for. For instance, you can filter laptops based on price, screen size, manuf acturer, and a System is flexible to meet different information needs Website Quality: Flexibility Expertise Reliability Neutral Integration wikipedia (somewhat reliable) WebMD.com (very reliable, it's run by docto rs) and find any mental illness boards you Website run by professionals legitimate source of health information on the web and do some research before you see your doctor again. I recommend WebMD or Website is re liable and legitimate sites, or .edu sites, though, because commercial sites are always interested in trying to sell you one way or the other, so their information is often biased to increase Commercial websites are biased i n providing information ConsumerSearch for a comprehensive look at mattress reviews compiled from a variety Integrate reviews from other websites