FACEBOOK AND IMPRESSIONS OF NEW ROOMMATES IN THE TRANSITION TO COLLEGE: THE IMPACT OF DISCREPANCIES BETWEEN ONLINE AND OFFLINE ROOMMATE IMPRESSIONS ON THE DEVELOPMENT OF ROOMMATE RELATIONSHIPS AMONG FIRST YEAR STUDENTS By Ying-ju Lai A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOPOPHY Communication Arts and Sciences-Media and Information Studies 2011 ABSTRACT FACEBOOK AND IMPRESSIONS OF NEW ROOMMATES IN THE TRANSITION TO COLLEGE: THE IMPACT OF DISCREPANCIES BETWEEN ONLINE AND OFFLINE ROOMMATE IMPRESSIONS ON THE DEVELOPMENT OF ROOMMATE RELATIONSHIPS AMONG FIRST YEAR STUDENTS By Ying-ju Lai This dissertation is an attempt to explore first year college students’ Facebook use in association with their relationship development with their previously unacquainted roommates. Survey data indicated that it is very common for freshmen to look up their roommates on Facebook after receiving their roommate assignment from school. Being able to get an idea of who the roommate is helps reduce a student’s uncertainty about living with a complete stranger. Both the uncertainty reduction theory (URT) and the predicted outcome value (POV) theory provide a solid theoretical framework to predict students’ information-seeking behaviors on Facebook. Social information processing (SIP) theory and hyperpersonal model are proved to be helpful in explaining students’ impression formation process on Facebook. This research predicts that information-seeking behaviors as well as the impressions formed based on information available on Facebook will reduce students’ uncertainty about the roommates. Moreover, the study aims to take a further step by applying expectancy violations theory to investigate whether the discrepant impressions formed between Facebook and offline experience have an impact on students’ level of uncertainty upon move-in with the roommates, and more importantly, the impact on the development of roommate relationship closeness. A combination of a three-wave survey method and in-depth interviews with 19 students was used for this study. Survey data were collected at three different time points: (1) two weeks before college freshmen moved in with their roommates; (2) one week after they moved into the dormitory; and (3) seven weeks after living together with the roommates. The interviews were conducted after the three-wave survey was completed. Statistical analyses using multiple linear regressions, multiple analysis of variance, and mixed-design ANOVA were applied for the hypotheses testing. The findings were mostly consistent with the hypotheses: (A) before moving in with the roommates, incoming first year students’ uncertainty level was affected by how often they interacted with the roommates on Facebook, how many channels they used to communicate with the roommates, and their impressions of roommates’ appearance and task attractiveness; (B) freshmen who formed positive initial Facebook impressions engaged in more information-seeking behaviors and had greater certainty than those who formed negative initial impressions of the roommates; (C) upon moving in with the roommates, students’ uncertainty level was influenced by their offline impressions of the roommates’ social and appearance attractiveness; (D) discrepancies between initial Facebook and offline impressions produced significant group differences in students’ level of uncertainty and relational outcomes with the roommates; and (E) students’ uncertainty and the impressions of roommates change over time and among groups. Copyright by YING-JU LAI 2011 TABLE OF CONTENTS LIST OF TABLES ……………………………………………………………………………….iv LIST OF FIGURES ………………………………………………………………..…………....vii CHAPTER 1 INTRODUCTION ………………………………………………………………………………..1 CHAPTER 2 THEORETICAL FRAMEWORK ………………………………………………………………..8 Uncertainty Reduction ………………………………………...……………………….....8 Uncertainty Reduction Theory ………………………………………………........9 Predicted Outcome Value Theory ……………………………………………….11 Impression Formation Online ………………………………………………….………..13 Impressions of the Roommates’ Social, Physical, and Task Attractiveness……………..15 Impression, Uncertainty, and Anticipated Future Interaction in the Initial Interactions...16 Impressions and Uncertainty in Offline Interactions ……………………………………18 Impression Discrepancies …………………………………………………………….....19 Relationship Development with the Roommates …………………………………….….22 Expectancy Violations Theory ……………………………………………………….….23 Changes in Uncertainty and Impressions over Time ………………………………........25 CHAPTER 3 METHODS ………………………………………………………………………………...……27 The First-Wave Study: Two Weeks before Move-In …………………………………...27 Procedure ………………………………………………………………………..27 Participants …………………………………………………………………...….28 Roommate Assignment ………………………………………………………….31 Roommate-Related Facebook Experience ……………………………………....32 Information for Impression Formation on Facebook …………………………....35 The Second-Wave Study: A Week after Living Together …………………………....…37 Procedure ………………………………………………………………………..37 Participants ……………………………………………………………………...37 Information Accuracy for Impression Formation on Facebook ………………...42 The Third-Wave Study: Seven Weeks after Living Together …………………………..43 Procedure ………………………………………………………………………..43 Participants ………………………………………………………….………..…44 Measurement ……………………………………………………….……….………......46 Uncertainty ……………………………………………………….……………..46 Impressions of the Roommates’ Social, Appearance, and Task Attractiveness...47 Relationship Closeness Inventory ……………………………………………....48 i Research Design …………………………………………………………………………51 Interviews ………………………………………………………………………………..52 CHAPTER 4 RESULTS …………………………………………………………………………………….....54 Hypotheses Testing for the First-Wave Study ……………………………………....…..54 Hypothesis 1 ……………………………………………………………...……..56 Hypothesis 2 ………………………………………………………………...…..57 Hypotheses Testing for the Second-Wave Study ……………………………………….59 Hypothesis 3 …………………………………………………………………….60 Hypothesis 4 ………………………………………………………………...…..61 Hypothesis 5 ………………………………………………………………….....63 Hypotheses Testing for the Third-Wave Study …………………………………………64 Hypothesis 6 …………………………………………………………………….64 Hypothesis 7: Changes in Certainty …………………………………………………….65 Hypothesis 8: Changes in Impressions ………………………………………………….68 Impressions of the Social Attractiveness ………………………………………..68 Impressions of the Appearance Attractiveness ……………………………….…71 Impressions of the Task Attractiveness …………………………………………74 CHAPTER 5 INTERVIEW SUMMRY ……………………………………………………………………….80 Before Moving in with the Roommates …………………………………………………80 Roommates’ Facebook Profiles …………………………………………………81 Communication Channels for Roommate Interaction before Move-in ……...….85 Upon Moving in: Impression Discrepancies …………………………………………….86 Developing a Close Relationship with the Roommates ………………………………....89 Relationship Closeness = 1~2 …………………………………………………...89 Relationship Closeness = 4~5 …………………………………………………...90 Relationship Closeness = 6~7 …………………………………………………...90 Relationship Closeness = 8~10 ………………………………………………….91 Conclusion ………………………………………………………………………………91 CHAPTER 6 DISCUSSION AND CONCLUSION …………………………………………………………..93 Hypothesis 1 ……………………………………………………………………………..93 Hypothesis 2 ……………………………………………………………………………..95 Findings from the Interviews: Initial Impression Formation on Facebook ……….…….96 Hypothesis 3 ……………………………………………………………………………..98 Hypothesis 4 ……………………………………………………………………………..98 ii Findings from the Interviews: Impression Discrepancies ……………………………..100 Hypothesis 5 & 6 ………………………………………………………………………100 Findings from the Interviews: Roommate Relationship Development ………………..102 Hypothesis 7 …………………………………………………………………………...104 Hypothesis 8 …………………………………………………………………………...105 Limitations ……………………………………………………………………………..107 Future Research ………………………………………………………………………..109 Conclusion ……………………………………………………………………………..111 APPENDICES ………………………………………………………………………………...114 Appendix A: First-Wave Survey ………………………………………………………115 Appendix B: Second-Wave Survey…………………………………………………….123 Appendix C: Third-Wave Survey ……………………………………………………...129 Appendix D: Interview Questions ………………………………………………..……136 Appendix E: Scale Items, Descriptives, and Intercorrelations ………………………...138 REFERENCES ………………………………………………………………………………...151 iii LIST OF TABLES Table 1 Four Different Possibilities for Impression Discrepancies ………………...……………………21 Table 2 Age, Gender, Ethnicity, Family Income, and Facebook Use for the Sample of the First-Wave Study …………………………………………………………………………………………….29 Table 3 Descriptive Statistics for Students’ Roommate Assignment ……………………………………31 Table 4 Summary of Roommate-Related Facebook (FB) Experience for the Sample of the First-Wave Study …………………………………………………………………………….………………33 Table 5 Information for Impression Formation on Facebook ……………………………………………36 Table 6 Summary of Facebook (FB) Experience for the Sample of the Second-Wave Study ….……….38 Table 7 The Perception of the Accuracy of Roommates’ Self-Presentation on Facebook ………………42 Table 8 Summary of Facebook (FB) Experience for the Sample of the Third-Wave Study …………….44 Table 9 Summary of Scale Reliability, Means, and Standard Deviations for CL7, Social, Appearance, and Task Attractiveness for the Three-Wave Study …………………………………...………..48 Table 10 Descriptive Statistics for the Raw and the Converted Total Frequency, Diversity, Strength, and RCI Scores ………………………………………………………………………………………50 Table 11 Impression Discrepancies in Four Different Conditions ……………….……………………….52 Table 12 Intercorrelations for Information-Seeking Frequency, Interaction Frequency, Number of iv Channels, Facebook Impressions, and Certainty about the Roommates in the First-Wave Study ……………………………………………………………………………………….……55 Table 13 Regression Coefficients for Students’ Certainty about their Roommates in the First-Wave Study …………………………………………………………………………………………….57 Table 14 Summary of Means and Standard Deviations for Variables by the Types of the Initial Facebook Impressions and Cronbach’s Alphas for Scales in the First-Wave Study …………………..…..58 Table 15 Intercorrelations for Information-Seeking Frequency, Interaction Frequency, Number of Channels, Offline Impressions, and Certainty about the Roommates in the Second-Wave Study ………………………………………………………………………………………….…59 Table 16 Regression Coefficients for Students’ Certainty about their Roommates in the Second-Wave Study …………………………………………………………………………………………… 61 Table 17 Summary of Means and Standard Deviations for Variables by the Types of the Offline Impressions and Cronbach’s Alphas for Scales in the Second-Wave Study…………….………62 Table 18 Summary of Means and Standard Deviations for Scores on Students’ Certainty at the Second Stage by Different Impression Discrepancies ………………………………….………………..63 Table 19 Summary of Means and Standard Deviations for Scores on Students’ Relationship Closeness Index (RCI) by Different Impression Discrepancies ……………………………...…………….65 Table 20 Summary of Descriptive Statistics for Students’ Scores of Certainty………………..………….67 Table 21 Summary of Descriptive Statistics for Impressions of Social Attractiveness …………………...71 Table 22 Summary of Descriptive Statistics for Impressions of Appearance Attractiveness ………….....74 Table 23 Summary of Descriptive Statistics for Impressions of Task Attractiveness …………………….77 v Table 24 Hypotheses Testing and Results ………………………………………………………...………78 Table 25 CL7 Attributional Confidence Items ………………………………………………….……….138 Table 26 Summary of Means, Standard Deviations, and Intercorrelations of Items for CL7 Attributional Confidence Scale …………………………………………………………………………...….139 Table 27 Impressions for Social Attractiveness Items …………………………………………………...140 Table 28 Summary of Means, Standard Deviations, and Intercorrelations of Items for Social Attractiveness …………………………………………………………………………………..141 Table 29 Impressions for Appearance Attractiveness Items ………………………………………….….142 Table 30 Summary of Means, Standard Deviations, and Intercorrelations of Items for Appearance Attractiveness ………………………………………………………………………..…………143 Table 31 Impressions for Task Attractiveness Items ……………………………………………...……..144 Table 32 Summary of Means, Standard Deviations, and Intercorrelations of Items for Task Attractiveness …………………………………………………………………………………..145 Table 33 RCI Frequency Items ……………………………………………….………………………….146 Table 34 RCI Diversity Items ……………………………………………………………………………147 Table 35 RCI Strength Items ………………………………………………………………………….....148 Table 36 RCI Conversion Table ……………………………………………………………….………...149 vi LIST OF FIGURES Figure 1 Changes in Certainty at the Three Stages ………………………………………………..….......66 Figure 2 Changes in Students’ Certainty by Groups ………………………………….……………..……66 Figure 3 Changes in Students’ Certainty at the Three Stages by Groups ………………………….……..67 Figure 4 Changes in Impressions of the Roommates’ Social Attractiveness at the Three Stages .……….70 Figure 5 Changes in Students’ Impressions of the Roommates’ Social Attractiveness by Groups ……....70 Figure 6 Changes in Students’ Impressions of the Roommates’ Social Attractiveness at the Three Stages by Groups ………………………………………………………………………………………..70 Figure 7 Changes in Students’ Impressions of the Roommates’ Appearance Attractiveness at the Three Stages ……………………………………………………………………………………………73 Figure 8 Changes in Students’ Impressions of the Roommates’ Appearance Attractiveness by Groups…73 Figure 9 Changes in Students’ Impressions of the Roommates’ Appearance Attractiveness at the Three Stages by Groups ………………………………….…………………………………………….73 Figure 10 Changes in Impressions of the Roommates’ Task Attractiveness at the Three Stages……….…76 Figure 11 Changes in Students’ Impressions of the Roommates’ Task Attractiveness by Groups ……..…76 Figure 12 Changes in Students’ Impressions of the Roommates’ Task Attractiveness at the Three Stages by Groups ………………………………......……………………………………………………….76 vii CHAPTER 1 INTRODUCTION Seven years after Facebook’s first debut on the Harvard campus in 2004, it now has more than 500 million active users around the world, and about 30% of them are inside the United States (Facebook, 2011). A recent Pew report indicated that 72% of American young adults (1829 years old) are using social network sites (SNSs), and among them, 71% have a profile on Facebook (Lenhart, Purcell, Smith, & Zickuhr, 2010). As the Facebook phenomenon has spread, many questions and concerns have drawn attention from both the general public and researchers regarding its influences on the social, psychological, and behavioral aspects of its users, such as privacy issue, users’ self-presentation and impression management in the profiles, and their psychological well-beings. For example, research showed that young adults are at risk for stalking occurrences (Cass, 2011; Fisher, Cullen, & Turner, 2002) due to the accessibility of personal information on SNSs; therefore, young adults are advised to take steps to protect themselves from being harmed. There have been many studies specifically focusing on its implications for college students nowadays (e.g., Ellison, Steinfield, & Lampe, 2007; Lampe, Ellison, & Steinfield, 2008; Lou, 2009; Pempek, Yermolayeva, & Calvert, 2009; Steinfield, Ellison, & Lampe, 2008). The studies found that about 85% to 95% of college students have joined Facebook. The time they spent on Facebook each day has increased significantly from 29.48 minutes in 2006 to more than 63 minutes in 2007 on average (Steinfield, et al., 2008). While Facebook has become one of the routine activities for college students, many parents, educators, and researchers are curious about its effects on students’ intellectual, personal, and interpersonal developments. For example, Ellison and her colleagues (Ellison, et al., 2007; Steinfield, et al., 2008) investigated the 1 relationships between students’ Facebook use, social capital, psychological well-being, and selfesteem; and Pempek, Yermolayeva, and Calvert (2009) conducted a study on college students’ social networking experiences on Facebook. These studies indicated a positive relationship between Facebook use and students’ social interactions—it helps maintain their former connections and facilitate a sense of belongings in the college environment—which is critical for students who are at their emerging adulthood stage, where friendships at this stage are influential in students’ development of identity, well-being, and family relationships in the future (Steinfield, et al., 2008). This study aims to explore first year college students’ Facebook use in association with their relationship development with their previously unacquainted roommates. In 2007, several articles reported that there were an increasing number of incoming college freshmen and their parents requested housing officials for a new roommate assignment in that summer after they looked up their future roommates’ Facebook profiles (e.g., Collura, 2007; Eberhardt, 2007; Walsh-Sarnecki, 2007). From the sample of the present study, 99.5% of the first year students indicated that they looked up their roommates on Facebook after receiving the roommate assignment from school in August 2010. It suggests Facebook is playing an important role in their daily lives, especially during the transition to college life. Being able to ―meet‖ a future roommate before moving in together helps to alleviate the unease students face related to the prospect of living with a stranger (Scissors, 2007); however, others have asserted that the decision to change roommates based on a negative Facebook impression would rob college freshmen of significant social learning opportunities to ―develop personal flexibility and learn to get along with people different from themselves‖ (Eberhardt, 2007). The arguments from both sides seem legitimate and they point out how Facebook can be both beneficial and harmful in the 2 development of roommate relationships among first year students. How does Facebook affect first year students’ decisions to keep or change their assigned roommates? This is the central question that has inspired the undertaking of the present study. Facebook, like other social network sites (SNSs), allows its members to present themselves in a variety of ways and forms in their personal profiles. Users can upload pictures and update personal information, such as their status, education, activities and interests, contact information and other personal background. It enables users to connect with friends, establish their friend networks, and interact with friends using a variety of applications available on the site. boyd and Ellison (2007) defined SNSs as web-based services that allow users to ―(1) construct a public or semi-public profile within a bounded system, (2) articulated a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system.‖ Although SNSs are primarily designed to maintain existing social links that are created offline, some people also use them as a tool to make new friends (Lenhart & Madden, 2007). For college freshmen, they are able to benefit from both during the transition from high school to college: on the one hand, they can maintain connections with friends from their past, and on the other hand, they can establish virtual connections with new friends they meet, including people in their dorms, classmates, and participants from social events, and so on. Whether students are connecting with existing friends or new ones, how they present themselves in their profiles has become an important issue. For instance, when a student joins a group which advocates some irresponsible behaviors, even if these behaviors do not reflect his actual belief, the self-image he presents in the profile will be associated with negative qualities. It can become unfavorable to him when he attempts to apply for undergraduate research assistant. 3 In his book, Self-Presentation: Impression Management and Interpersonal Behavior, Mark R. Leary (1996) indicated that it is never a simple task to produce the desired outcome when it comes to wanting others to form a particular impression. As he pointed out, ―many things can go wrong as people try to assess the most desirable impression to convey to particular targets in a given situation, adjust their behavior to foster the desired impression, and assess whether others have formed the impression they intended‖ (p.11). When SNSs users manage their impressions, they may face the dilemma of choosing between a façade that is intended to fit the norms shared by the public and a portrayal of a more genuine and playful image of self. Thus, for example, a student may upload party-related content and photos on Facebook, showing he was having fun at the party. Consequently, his friends know he is portraying a sociable side of himself; but for other people, such as the parents of his future roommate, such images might give an impression that he is involved with alcohol and drugs. The example demonstrates the importance of managing one’s online presence to convey the right impression to other people, because it sometimes can produce a false image for the unexpected audiences. Previous research (Lampe, Ellison, & Steinfield, 2006; Lampe, et al., 2008) showed that college students perceived their peers, such as friends and classmates, were their expected audience, as opposed to faculty, law enforcement, and future employers. However, there have been reports in the past where Facebook users got into trouble with school administrators or limited their opportunity with their future employers for not strategically managing their profiles for these unexpected viewers (e.g., Brandenburg, 2008; Du, 2007; Smith & Blanchard, 2008). Although Facebook interface has changed a lot since then and users have more control over who can see their profiles, strategically managing one’s image is equally critical. 4 Research on computer-mediated communication (CMC) and self-presentation in online dating (e.g., Ellison, Heino, & Gibbs, 2006; Gibbs, Ellison, & Heino, 2006; Toma, Hancock, & Ellison, 2008) suggested that users engaged in selective self-presentational behaviors in order to create a more favorable, but sometimes misrepresented, image to impress others. But at the same time, being aware of the possibility of future face-to-face (FtF) interactions with the target person, they would be less likely to completely lie about themselves in their profiles. Facebook users, unlike online daters, have even less incentive to engage in deceptive self-presentational activities, partly due to what boyd and Ellison (2007) suggested that SNSs ―enable users to articulate and make visible their (existing) social network,‖ but it does not mean impression management would be less important to them. CMC research has shown the process by which people form an impression of their online partner based on information available in CMC environments, despite the absence of nonverbal cues (e.g., Hancock & Dunham, 2001; Jacobson, 1999; Ramirez, 2007; Walther, Slovacek, & Tidwell, 2001). Compared with more traditional text-based CMC technologies, Facebook provides its users with a setting richer in cues where users are capable of forming more accurate impressions of other people (Antheunis, Valkenburg, & Peter, 2010; Gosling, Gaddis, & Vazire, 2007). Theories of uncertainty reduction (Berger, 1979; Berger & Calabrese, 1975; Sunnafrank, 1986, 1990) suggested that people will engage in information-seeking behaviors to reduce their level of uncertainty during the initial stage of developing interpersonal relationship. With the prevalence of Facebook among young adults, it is not surprising that freshmen use it to check out their future roommates’ profiles to get an idea about who he/she is and what it would be like to live with this person. 5 Although research in the past has shown that profile owners have less incentive to create a false image and thus viewers are likely to form a more accurate impression of other people, will the Facebook impressions effectively reduce first year students’ uncertainty about their roommates? Discrepancies between online image and offline reality can always happen. Social information processing (SIP) theory and the hyperpersonal model (Walther & Parks, 2002) suggested that people tend to either idealize or stereotype their online partners. For college freshmen, they could find out that their roommates in reality may not be the person they pictured earlier based on their initial impressions from Facebook. It is possible that freshmen will be disappointed to find out that their roommates are in fact unpleasant to live with; they may feel relieved that their roommates are actually very nice; or they are able to reinforce their expectation that their roommates are similar to what they pictured. Burgoon and Hale (1988) proposed a nonverbal expectancy violations theory to explain that when positive violations in expectation occur, it is likely to create a more favorable consequence; conversely, when negative violations occur, it is likely to result in a more unfavorable outcome. Hence, it is important to ask now whether the discrepancies between initial Facebook and offline impressions affect first year students’ degree of uncertainty and the relationship development with their future roommates. This study will build on the existing theories to explore the role that Facebook plays in the development of roommate relationships for first year students, including the following questions: (1) How do college freshmen form impressions of their previously unacquainted roommates based on the information obtained from Facebook? (2) How do students’ initial Facebook impressions and their information-seeking behaviors influence the degree of uncertainty they feel about their future roommates? (3) To what extent are there discrepancies between the impressions students develop through Facebook and their subsequent offline 6 impressions when they first meet their roommates? And (4) to what extent do impression discrepancies affect first year students’ uncertainty about their roommates once they meet in person, as well as the extent to which they develop close relationship with their roommates? By conducting interviews and a three-wave survey to observe the changes in first year students’ level of uncertainty, the changes in roommate impressions, and the degree of closeness in their relationship, the findings will be valuable for further understanding of the implications of SNSs for the uncertainty reduction process and the interpersonal relationship development for students during the transition to college. 7 CHAPTER 2 THEORETICAL FRAMEWORK The transition from high school to college is a very challenging process for the first year college students. During this transition, interpersonal relationship development is one of the several challenges they are facing. They are usually very concerned about who their roommates will be and whether they will be accepted by peers and make friends. Research showed that peer relationships and participation in social networks are critical for the first year students to reduce some of the stress involved in the transition process (e.g., Corwin & Cintron, 2011; Leafgren, 1989; Tinto, 1997). What do they do to reduce the stress caused by their uncertainty about meeting new people and living with their new roommates? During interpersonal interaction, people want to be able to predict and explain their partners’ behaviors. They would like to know their partners’ reactions in advance and why they react in a certain way. In order to do so, they will start gathering information that allows them to predict and make sense of their partners’ behaviors. Facebook happens to be a venue full of personal information where users can gain access to by adding/accepting profile owners to be their Facebook friends. Moreover, the plethora of information on roommates’ profiles, including the uploaded and tagged pictures, the shared and commented wall posts, and the different intensity of friends’ interactions, all contribute to the dynamic process of forming impressions of the roommates. It allows first year college students to be more informed as to what the future living situation would be, thus effectively reducing their uncertainty. Uncertainty Reduction To explain the interpersonal relationship development with roommates, especially at the early stage of acquaintance, Berger and Calabrese’s (1975) uncertainty reduction theory (URT) 8 and Sunnafrank’s (1986) predicted outcome value theory (POV) offered two viable explanations for the process of uncertainty reduction. Although the two theories were developed to explain the process of uncertainty reduction for interpersonal relationship development during face-to-face interaction, many researchers also extended the theories to examine computer-mediatedcommunication (CMC) interaction (e.g., Antheunis, et al., 2010; Ramirez, Walther, Burgoon, & Sunnafrank, 2002; Tidwell & Walther, 2002). Uncertainty Reduction Theory In their early discussion of uncertainty reduction theory (URT), Berger and Calabrese (1975) made a key assumption that during initial interactions with strangers, an individual’s primary concern was to reduce uncertainty about the other’s interaction behaviors. They asserted that uncertainty reduction involved both a ―proactive process of creating predictions‖ and a ―retroactive process of explaining the other’s behavior‖ (p. 101). The theory proposed seven axioms describing the relationship between uncertainty and the amount of verbal communication, 1 nonverbal affiliative expressiveness , information seeking behavior, intimacy level of 2 communication content, reciprocity rate , similarities between interaction partners, and liking. 1 Nonverbal affiliative expressiveness by definition means ―nonverbal expressions of affiliation‖ that a partner shows toward the person during the verbal communication. The indicators of nonverbal affiliative behaviors can be eye contact, head nods, arm gestures, pleasantness of vocal expressions, and facial expressions. 2 Reciprocity rate means the amount of reciprocal information exchange between partners. Reciprocal information disclosure helps assure both interacting parties that no one will gain ―information power‖ over the other. 9 According to the theory, in a situation where the uncertainty level is high, the amount of verbal communication, information seeking behavior, and reciprocity rates will increase in order to reduce the uncertainty; nonverbal affiliative expressiveness and the intimacy level of communication content, on the contrary, will decrease when uncertainty level is high; and similarities and liking to the interaction partners will reduce uncertainty. Berger (1979) later on suggested that people would most likely engage in uncertainty reduction activities under three circumstances: (a) there are incentive values to interact with the partners; (b) partners have deviant behaviors; and (c) there are anticipated future interactions with the partners. In order to reduce the uncertainty about the partners, people would use any combination or all three types of passive, active, and interactive information-seeking strategies to find out more about the partners. People who adopt passive information-seeking strategies will obtain information with the minimal or no direct effects on the target person. The target will have the least awareness of the fact that they are being observed. The passive strategies can be gathering information about the target person from online search engines, like Google; reading his blogs; following his tweets on Tweeter; or checking up his status updates on Facebook. Active information-seeking strategies require some efforts in order to find out something about the target, although they do not involve direct interaction with him. Berger (1979) suggested two active information strategies: the first is to ask the third parties who might be familiar with the target; the second is environmental structuring—by creating an environment in which they can unobtrusively observe the target. For first year students, they might find out that they and their future roommates have some common Facebook friends, so they can ask them about their roommates. Interactive strategies require a direct interaction with the target during which different tactics are used to elicit the desired information. They can be direct verbal interrogation or reciprocal self-disclosure. 10 Predicted Outcome Value Theory Subsequent research on URT has produced conflicting results, leading to important revisions to the theory. Kellermann (1986) found that the anticipated future interaction did not always lead to reduced uncertainty, and sometimes increased information-seeking behaviors would even result in a higher level of uncertainty (Planalp & Honeycutt, 1985). In predicted outcome value theory, Sunnafrank (1986) expanded URT and proposed that when a person expected greater predicted outcome values (POV) in the future relationship, he would be more attracted to the partner; more efforts would be made to ―extend interaction and establish future contact‖ (p.10) if the person anticipated positive predicted relational outcomes, but if the person anticipated negative outcomes in the future, he may ―terminate or curtail the conversation and future contact‖ (p.11); and finally, the person would choose conversation topics that would lead to the most positive outcomes. He contended that the relationship between uncertainty and the seven factors proposed in URT axioms would vary, depending on what kind of relational outcome the person would expect for the future relationship. According to POV, what Berger and Calabrese (1975) proposed would only be true under the circumstances where the person was expecting a positive relational outcome. For college freshmen preparing to move in with roommates they have never met before, there is a fairly short period between the day they receive their roommate assignment (normally in early August) to the day they move in together (in late August). One can only expect that they would experience a very high level of uncertainty about living with a stranger. Sunnafrank (1990) suggested that when people anticipate that they will be in ―close physical proximity‖ in the future—defined by Berger (1979) as ―anticipated future interaction‖—the process of reducing uncertainty will most likely happen. This is exactly the scenario faced by college freshmen, who 11 understandably rely on a social network site like Facebook to learn more about their future roommates, as they seek to develop a good and close roommate relationship. Both theories proposed that high levels of uncertainty lead to an increase in informationseeking behaviors. However, URT suggested that as uncertainty levels decrease, informationseeking behaviors decrease, too (Berger, 1979; Berger & Calabrese, 1975), while POV had an opposite proposition—reduced uncertainty would cause increased information-seeking behaviors when positive outcome values were expected; whereas when negative outcome values were expected, reduced uncertainty would produce decreased information-seeking behaviors (Sunnafrank, 1986, 1990). It is natural to assume that college freshmen would expect a positive relational outcome value with their future roommates. With the popularity of Facebook among students, they are able to engage in information-seeking behaviors to form a better idea of what their roommates would be like as well as what it would be like to live together. The characteristics of Facebook's design allow freshmen to adopt any combination of the three information-seeking strategies to help reduce their uncertainty about their future roommates, but adding/accepting the roommates as their Facebook friends is a prerequisite condition. They could use passive strategies with no direct intervention and minimal effects on the roommates’ behaviors, active strategies in which no direct contact is made with the roommates, and/or interactive strategies involving direct interaction with the roommates (Berger, 1979) on Facebook. For example, they may be able to obtain their roommates’ demographic and background information, observe their prospective roommates’ interactions with other Facebook friends, and initiate direct interactions with them on Facebook (not necessarily in this order, though). In other words, Facebook facilitates the information-seeking process by enabling incoming students to choose whatever methods they feel most useful and comfortable for 12 reducing uncertainty about their future roommates. With regard to information-seeking behavior and uncertainty, the first part of hypothesis 1 is proposed: Hypothesis 1 (A): Before moving in with the previously unacquainted roommates, (a) the frequency that freshmen invest in gathering information about their roommates on Facebook, (b) the interaction with the roommates on Facebook, and (c) the number of channels used for communication will have a positive impact on reducing uncertainty about their future roommates. Impression Formation Online Early research on CMC suggested that the absence of nonverbal cues would restrict peoples’ ability to form impressions and develop close interpersonal relationships online (Baym, 2001; Walther & Parks, 2002). However, subsequent research showed that if the limitation of time was removed, those interacting via CMC could achieve similar relational outcomes to a face-to-face (FtF) group (Walther, 1992). The social information processing (SIP) theory suggests that while it may take longer period of time, people are able to form impressions of others based on the social information available online even with the absence of nonverbal cues in the CMC environments (Tidwell & Walther, 2002; Walther, 1992, 1996, 1997). In line with the same consideration of the lack of nonverbal cues in CMC, the social identity/deindividuation (SIDE) theory focuses on the contextual cues and cues that ―indicate the common social categories of CMC group members‖ (Walther & Parks, 2002, p. 539): that is, instead of relying on the individual-level information, people form impressions of others based on the social identity shared by group members which sometimes resulted in over-idealized perceptions about others (Hancock & Dunham, 2001; Tidwell & Walther, 2002; Walther, 1996, 1997; Walther & Parks, 2002). A more recent hyperpersonal model suggests that CMC allows users to practice 13 ―selective self-presentation‖ to create a favorable image of self, and because of limited social and interpersonal cues, the receivers of other’s self-presentation tend to form more stereotyped or idealized impressions of other people (Hancock & Dunham, 2001; Walther, 1996). These three CMC theories point out the fact that developing interpersonal relationships would not be confined to FtF settings. Rather, people are able to form impressions about others from information available in CMC environments in order to reduce uncertainty if future interaction seems possible. Facebook, unlike traditional text-based CMC with limited nonverbal cues, allows the users to share a great deal of personal information which can then facilitate the process of managing impressions of self as well as forming impressions of others. As Antheunis, Valkenburg, and Peter (2010) suggested, cue-richer and more open CMC environments may affect the information-seeking strategies that people use to form impressions and to reduce uncertainty about a new acquaintance in initial interactions. Moreover, another characteristic of Facebook is that the personal information, especially other-generated information (e.g., friends’ comments and wall postings about the person), may have greater warranting value (Walther, Van Der Heide, Hamel, & Shulman, 2009; Walther, Van Der Heide, Kim, Westerman, & Tong, 2008). Walther and Park (2002) defined the warranting value of information as ―derived from the receiver’s perception about the extent to which the content of that information is immune to manipulation by the person to whom it refers‖ (p. 552). The warranting value of information on Facebook can be viewed from two aspects. First, not only the user himself but friends can contribute information about the profile owner. Information provided by friends was perceived to have higher warranting value (Walther, et al., 2009). Second, users still have offline connections with many of their Facebook friends which, as suggested by signaling theory, can ―provide 14 explicit or implicit verification of identity claims‖ and thus makes deception costly (Donath & boyd, 2004; Lampe, Ellison, & Steinfield, 2007). With these two constraints, Facebook users are believed to have less incentive to misrepresent themselves. For college freshmen, with relatively more reliable information about the roommates on Facebook, the impressions they get about their roommates before moving in with him/her should be able to decrease their uncertainty. Impressions of the Roommates’ Social, Physical, and Task Attractiveness Interpersonal attraction has long been found to be important in interpersonal communication and close relationship (Berscheid, 1985; Berscheid & Reis, 1998; J. C. McCroskey & McCain, 1974; L. L. McCroskey, McCroskey, & Richmond, 2006). McCroskey and McCain (1974) summarized the relationship between interpersonal attraction and interpersonal communication: ―the more people are attracted to one another, the more they will communicate with each other; and the more we are attracted to another person, the more influence that person has on us in interpersonal communication‖ (p.261). Previous research showed that physical appearance attractiveness has a strong association with positive personality impressions—―what is beautiful is good‖ (Dion, Berscheid, & Walster, 1972; Walther, et al., 2009); and social attractiveness and task attractiveness are also common in evaluating the impressions of the partners (Tong, Van Der Heide, Langwell, & Walther, 2008; Walther, et al., 2008). Facebook affords students information about future roommates that can influence their impressions of the roommates’ attractiveness on all three dimensions—social attractiveness represents liking, task attractiveness is associated with roommates’ general ability (or competence) and their sense of responsibility, and physical appearance attractiveness. For instance, Tong and her colleagues (2008) found that the number of Facebook friends had a curvilinear relationship with others’ perception of profile owner’s social attractiveness, which 15 suggesting that too few or too many Facebook friends will affect the ratings of the individual’s social attractiveness. In addition, Walther and his research team (2008) investigated the impact of the appearance and behavior of a user’s Facebook friends on others’ perceptions of the user’s physical attractiveness, social attractiveness, and task attractiveness. Their findings indicated that the presence of friends’ attractive photographs on a user’s profile would increase the user’s physical and social attractiveness; and that user’s task and social attractiveness were influenced by the wall postings made friends with regard to the user. And thus, the second part of the first hypothesis is proposed: Hypothesis 1 (B): Before moving in with the previously unacquainted roommates, in addition to the three factors described in H1(A), the initial Facebook impressions of the roommate’s social, appearance, and task attractiveness would also have a positive impact on reducing students’ uncertainty about their future roommates. Impression, Uncertainty, and Anticipated Future Interaction in the Initial Interactions Berger and Calabrese (1975) proposed that a decreased uncertainty level would lead to an increase in liking. Sunnafrank (1986) modified the proposition by arguing that the predicted relational outcomes would affect this relationship. That is, if a negative relational outcome is anticipated in the future, even when the uncertainty level has decreased, the impression of the partner’s attractiveness would not improve. On the contrary, if a positive relationship outcome is expected in the future, the impression of the partner’s attractiveness would be more positive, and the person would be more certain about the partner. Previous research has indicated that there is a positive relationship between interpersonal attraction and reduced uncertainty (Antheunis, et al., 2010; L. L. McCroskey, et al., 2006). The first hypothesis for this study addresses the impact of 16 impressions on uncertainty, while the second hypothesis would like to address the relationship between impressions and predicted outcome values. Would initial Facebook impression of the roommate affect an incoming student's predicted outcome value for the relationship? The assumption is that if incoming students have a positive first impression about their roommates, their expectation for a positive future interaction would be positive, and therefore, they would make more efforts in order to reduce the uncertainty. This should be the case for freshmen who hold positive impressions of their roommates, but it would be equally important to consider those who have less positive impressions about their roommates. From the above reasoning, negative impressions may produce a negative expectation about future interaction, and therefore, little or no effort would be invested to decrease uncertainty. However, it may also be reasonable to assume that even though incoming students have negative impressions about their roommates in the beginning, they may still expect to develop a positive relationship with the roommates, because, unless they request to make a new roommate arrangement, they have no choice but to live with this person; and thus, they would still make an effort to reduce uncertainty. It would be interesting to see whether there is any difference in the amount of efforts they make to decrease uncertainty and whether there are differences in degrees of uncertainty between freshmen with positive and negative impressions of their roommates: Hypothesis 2: Before moving in with the previously unacquainted roommates, freshmen with more positive Facebook impressions of their roommates would (a) spend more time gathering roommates’ information on Facebook, (b) have more Facebook interactions with the roommates, (c) use more channels to communicate with the roommates, and (d) 17 have greater certainty about their roommates, compared to freshmen whose impressions of the roommates were negative. Impressions and Uncertainty in Offline Interactions Upon move-in day, for most incoming students, it is the first time they get to meet their new roommates FtF. The initial impressions they formed on Facebook can affect their offline attitudes towards the roommates. A study on impressions and relationship development (Ramirez, 2007) showed that although anticipated future interaction has an impact on initial relationship formation, continuous relationship development occurs as ―a function of the valence of the initial impression formed via CMC‖ (p.66). The findings from Ramirez’s study indicated a positive impact of initial impression on interpersonal relationship development. With a foreseeable future of living together with their roommates, students who formed positive impressions and those who formed negative impressions may already decide on different approaches to face their new roommates before their first offline meeting. However, these initial impressions and approaches are then subject to change upon meeting their roommates in person for the first time on move-in day. The offline impressions can be different from the Facebook impressions, depending on whether there was misrepresentation from roommates’ profiles or students’ incorrect perceptions of the roommates. Before investigating the impact of discrepant Facebook and offline impressions on students’ uncertainty and the future relationship development, the following two hypotheses focus on the impact of first year students’ impressions of the roommates formed right after first meeting them face-to-face: first on their uncertainty level; and second on whether degrees of the information-seeking efforts and level of uncertainty about their roommates would differ between positive and negative offline impressions. 18 Hypothesis 3: Students’ impressions of their roommates’ offline (a) social attractiveness, (b) appearance attractiveness, and (c) task attractiveness would have a positive impact on their degrees of certainty about the roommates right after meeting them FtF. Hypothesis 4: Right after meeting the roommates FtF, freshmen with positive impressions of their roommates would (a) spend more time gathering roommates’ information on Facebook, (b) have more Facebook interactions with the roommates, (c) use more channels to communicate with the roommates, and (d) have greater certainty about their roommates. Impression Discrepancies Uncertainty reduction is an ongoing process in interpersonal relationship development. While Facebook offers college freshmen a place to get information about their new roommates, the impressions they previously formed on Facebook can still be different from what the roommates are really like. As hyperpersonal interaction theory explained, people are able to engage in selective self-presentation to the optimization of the ideal self, and the receivers tend to idealize their partners in CMC environments (Walther, 1996). Jacobson (1999) also made a similar comment about online impressions after interviewing members of four different textbased virtual communities: ―these impressions are based not only on cues provided, but also on the conceptual categories and cognitive models people use in interpreting those cues.‖ Most participants Jocobson interviewed agreed that offline experiences did not match their online expectations about their partners. More studies dealing with online and offline impression discrepancies in recent years have focused on the dating relationship (e.g., Ellison, et al., 2006; Gibbs, et al., 2006; Shaw Taylor, Fiore, Mendelsohn, & Cheshire, 2010; Toma, et al., 2008). Similar to Facebook, the online dating sites allow the users to create their own profiles with 19 personal details and photographs. In order to make their self-presentations attractive to other members of the sites, users need to decide what information to disclose and what to ignore in the profiles, and sometimes they may even need to decide whether they will lie about themselves. For example, Toma and her colleagues (2008) found that deception in online dating profiles was ―frequent yet small in magnitude‖ (p.1033). Shaw Taylor, Fiore, Mendelsohn, and Cheshire (2010) investigated how impressions change during the transition from online to offline dating and how the discrepancies in impressions influence the relationship longevity. Among several factors being examined in their study, surprisingly, impressions of partners’ physical attractiveness formed online were mostly confirmed when partners meet face-to-face, and the most important factor that influenced relationship longevity was how well participants felt they had gotten to know their partners during the online-offline transition. One of the reasons for this outcome was that online daters were aware of the possibility of developing a close offline relationship with the partner in the future, so they did not want to ―polish‖ too much about themselves and then be accused of lying later. Although users of SNSs mostly use the services to maintain existing friend network and sometimes make new friends, they strategically manage their self-presentation for reasons very similar to those of online daters. On the one hand, they want to make their profile appealing to their existing friends and new friends; on the other hand, they are aware of the fact that the friends actually know them in real life. Gosling, Gaddis, and Vazire (2007) found that personality impressions that were formed based on Facebook profiles showed some accuracy, while profile owners did enhance their self-presentations. Therefore, what would college freshmen do while gathering information about their roommates on Facebook? The previous section in this chapter has discussed the warranting value of information on Facebook. Walther 20 and his colleagues (2009) found partial support for the notion that Facebook users do place more confidence in the information that is immune to a profile owner’s manipulation. Table 1 Four Different Possibilities for Impression Discrepancies Real Life Positive Negative 3. Positive 1. Unpleasant Surprise (Impression confirmed) (From positive to negative: Facebook Positive Negative increase in impression) Negative 4. Pleasant Surprise 2. Negative (From negative to positive: Positive (Impression confirmed) increase in impression) Regardless of the accuracy in self-presentation or impressions formed on Facebook, one can expect that the initial face-to-face encounter with the roommate may produce four possible conditions (see Table 1). The first situation is that an incoming student forms a positive impression of his/her roommate on Facebook, but changes to a negative impression after moving in with the roommate. The second situation is that an incoming student develops a negative Facebook impression of the roommate, and the impression remains negative after FtF interaction. The third situation is that a student's Facebook and offline impressions remain positive. Finally, the last possibility is that a student's initial impression of his/her roommate is negative, but changes to a positive offline impression after meeting the roommate in person. The second and the third situations in which the online and offline impressions remain the same reinforce first 21 year students' impressions of their roommates; however, the first and the last situations where the online and offline impressions were conflicted will require students to make some adjustment. Based on these four conditions, the following hypothesis investigates the impact of impression discrepancies on first year students' uncertainty levels following their move-in on campus. Hypothesis 5: Discrepancies in initial Facebook and offline impressions would influence first year students' degree of uncertainty about their roommates following move-in. H5 (a): Where impression discrepancies are positive (condition 4) or there is no discrepancy but the impressions remain positive (condition 3), students will be more certain about their roommates. H5 (b): Where the impression discrepancies are negative (condition 1) or there is no discrepancy but the impressions remain negative (condition 2), students will be less certain about their roommates. Relationship Development with the Roommates The transition from high school to college can be very challenging for college freshmen. It may be their first experience living with a stranger, making new friends, absorbing advanced knowledge, and learning to live a more independent college life. It would be beneficial to have peer groups and close friends who are experiencing this transitional period at the same, offering a strong possible supports for each other (Corwin & Cintron, 2011). These supports can come from their roommates. The room they share together can be viewed as their home at school. Developing a good relationship with one's roommate is without doubt an extremely important task. One characteristic of roommate situations is that roommates have a potential to develop a very close relationship. The familiarity that follows from daily interactions provided by living together is likely to create positive affective ties between roommates (Van Laar, Levin, Sinclair, 22 & Sidanius, 2005). In addition, a wide range of activities that roommates can share provide them with the opportunities to discover previously unnoticed similarities or counter-stereotypic characteristics or behaviors. However, the degree of closeness developed in their relationship may still be influenced by both the Facebook and offline impressions. It may take longer time for students who have negative discrepant impressions of their roommates to realize that their roommates are not the kind of people they have pictured. Will different impression discrepancies result in different relationship closeness? The next hypothesis will explore the impact of discrepant impressions on the degrees of closeness they develop in the roommate relationship. The measure of closeness of interpersonal relationships (Berscheid, Snyder, & Omoto, 1989) was developed based on the concept of interdependence proposed by Kelly and the colleagues (1983). They suggested that a close relationship is characterized by high interdependence between two people’s interconnected activities in which (a) they have frequent impact on each other, (b) the impact is strong, and (c) the impact involves diverse activities. Therefore, the time that the roommates spend together, the degree of influence on each other’s behaviors, decisions, and goals, and the number of different activities they do together can be indicators of their interdependence and how close their relationship is. Expectancy Violations Theory As mentioned in the first chapter, one of the goals for this study aims to investigate the influence of impression discrepancies on roommates' relationship development. The theory of expectancy violations (Burgoon & Hale, 1988) provides a viable framework to explain how people deal with the inconsistency in expectations and realities, and to predict the impact of the violation in expectations on the relational outcomes. In explaining the expectancy violations model, Burgoon and Hale indicated that ―if the actual (violation) behavior is more positively 23 valenced than the expected behavior(s), a positive violation occurs and should produce more favorable communication outcomes than conforming to the expected (normative) pattern. Conversely, if the actual behavior is more negatively valenced than the expected behavior, a negative violation is said to occur and should yield more negative consequences than conforming to expectations‖ (P.65). For freshmen who had matched Facebook and offline impressions of their roommates (i.e., no violation in their expectations), their anticipated relational outcome with the roommates would probably remain the same; whereas for freshmen who had discrepant Facebook and offline impressions, their anticipated relational outcome would change—positive changes in impression discrepancies (i.e., positive violation) may lead to a more positive relationship in the future, while the negative changes (i.e. negative violation) may produce a more negative relationship with the roommates. Based on this expectancy violations model, the sixth hypothesis is proposed: Hypothesis 6: Discrepancies between initial Facebook and subsequent offline impressions will affect the relationship development with the roommates. H6 (a): Negative impression discrepancies (condition 1) will lead to the lowest degree of closeness in the relationship. H6 (b): Impressions that remained negative (condition 2) would lead to less development of relationship closeness with the roommates. H6 (c): Impressions that remained positive (condition 3) would lead to a closer relationship with the roommates. H6 (d): Positive impression discrepancies (condition 4) would lead to the closest roommate relationship among the four conditions. 24 Changes in Uncertainty and Impressions over Time In order to observe the changes in the first year students’ levels of uncertainty and impressions of the roommates over time, the study designs a three-wave survey to record students’ uncertainty and impressions of their roommates at each stage—before moving in, upon moving in, and after living together for a while. They provide the researcher with an opportunity to investigate whether there is any significant increase or decrease in first year students’ uncertainty and their impressions of the roommates among three periods of time. Whether students’ uncertainty about their roommates will decrease as they have more opportunities to get to know each other, as the theories of uncertainty reduction predicted, can be examined here. Moreover, whether students’ uncertainty will differ among groups (i.e. four groups with different types of impression discrepancies) and the interaction effect of time and groups can be tested. Hypothesis 7 (A): First year students’ uncertainty level will decrease over time. Hypothesis 7 (B): First year students’ uncertainty level will be different among groups. Hypothesis 7 (C): There will be an interaction effect of time and groups on students’ uncertainty level. In addition to the changes in uncertainty level, whether the changes in impressions will follow the prediction by theories of impression formation on CMC that students’ initial impressions of social attractiveness, appearance attractiveness, and task attractiveness are scored higher than offline impressions will be examined here. Also, whether there are group differences and whether there is an interaction effect of time and groups will be investigated in the following hypothesis: Hypothesis 8 (A): First year students’ impressions of roommates’ (a) social attractiveness, (b) physical appearance attractiveness, and (c) task attractiveness will decrease 25 significantly from Time 1 (i.e., before moving in) to Time 2 (i.e., upon moving in), and no significant changes will be found between Time 2 and Time 3. Hypothesis 8 (B): First year students’ impressions of roommates’ (a) social attractiveness, (b) appearance attractiveness, and (c) task attractiveness will be different among groups. Hypothesis 8(C): There will be an interaction effect of time and groups on students’ impressions of roommates’ (a) social attractiveness, (b) appearance attractiveness, and (c) task attractiveness. 26 CHAPTER 3 METHODS A combination of survey methods and in-depth interviews with a subset of survey respondents was used for this study. To test the hypotheses, quantitative data were collected at three different time points: (1) the first-wave survey was launched on August 16, 2010, two weeks before college freshmen moved in with their roommates; (2) the second-wave survey was distributed on September 6, 2010, a week after they moved into the dormitory; and (3) the lastwave survey was sent on October 18, 2010, seven weeks after they lived together. The data collection period was two weeks long for each survey. The First-Wave Study: Two Weeks before Move-In Procedure In August 2010, two weeks before first year students’ move-in day, a random sample of 2,000 students plus 773 students from two introductory classes at a large Midwestern university received an email invitation from the researcher. A brief description about the research project, an incentive for participation, and a link to the survey were included in the invitation. Participants from the random sample were given a chance to receive one of the five $100 Amazon gift cards from the drawing, and participants recruited from the classes were compensated with extra credit points from their instructors. To be eligible to receive either the gift cards or extra credit points, participants needed to complete all three waves of the survey. The surveys were hosted on Zoomerang, a commercial online survey software site. Students were asked to answer questions about their roommate assignment, the level of acquaintance, roommate-related Facebook experience, ratings of the importance of the Facebook information for impression formation, uncertainty level about the roommates, impressions of roommates’ 27 social, appearance, and task attractiveness, and their demographic information (see Appendix A for the complete first-wave questionnaire). Participants A total of 606 students completed the first-wave online survey, with a response rate of 21.9%. Table 2 summarizes the demographic information. The sample consisted of 226 (37.5%) males and 376 (62.5%) females, with an average age of 18.09 years (SD = .37). The racial/ethnic distribution of the sample consisted of 77.7% White, 7.9% African American, 8.4% Asian, and 4.8% Hispanic. Among the respondents who reported their family total household income, 55.9% had more than $75,000. About 99% of the participants had Facebook profiles. Students who did not have a Facebook profile provided reasons for not having one, including: (a) not beneficial; (b) privacy concerns; (c) time consuming and causes many altercations among people; and (d) access issue because of the censorship in some country. In response to time spent on Facebook on a typical day, 202 (34.1%) participants spent less than an hour, and 235 (39.7%) answered that they spent one to two hours per day. The average amount of time spent on Facebook each day was 101.35 minutes (SD = 83.63). The average number of their Facebook friends is 550.75, ranging from 4 to 2,987. Compared with the previous result that undergraduate students in 2007 spent 63.57 minutes on Facebook and had 302.08 Facebook friends on average (Steinfield, et al., 2008), there has been a huge increase in amount of Facebook use time and number of Facebook friends since then. 28 Table 2 Age, Gender, Ethnicity, Family Income, and Facebook Use for the Sample of the First-Wave Study (N = 606) Variable Mean (SD) Age % 18.09 (.37) N 599 Gender Male 37.5 226 Female 62.5 376 White 77.7 471 African American 7.9 48 Native American 1.2 7 Asian 8.4 51 Pacific Islander 0.3 2 Hispanic/Latino 4.8 29 Multiracial 1.7 10 $75,000 or more 55.9 231 $50,000 to $74.999 16.0 66 $35,000 to 49,999 9.4 39 $20,000 to $34,999 8.2 34 Under $20,000 10.4 43 Race/Ethnicity Family Income 29 Table 2 (cont’d) Variable Mean (SD) % N Yes 98.5 597 No 1.5 9 Less than 1 hour 34.1 202 1 – 2 hours 39.7 235 2 – 3 hours 14 83 3 – 4 hours 5.4 32 4 – 5 hours 3.0 18 5 – 6 hours 0.8 5 6 – 7 hours 1.7 10 7 – 8 hours 0.8 5 More than 8 hours 0.3 2 Facebook Profile Time spent on Facebook on a typical day Daily minutes Facebook use* 101.35 (83.63) Facebook Friends 550.75 (357.05) *. Minutes of Facebook use were converted from an ordinal scale by assigning the mid-point of each response category, where less than 1 hour = 30 min, 1-2 hours = 90 min, 2-3 hours = 150 min, 3-4 hours = 210 min, 4-5 hours =270 min, 5-6 hours = 330 min, 6-7 hours = 390 min, 7-8 hours =450 min, more than 8 hours = 480 min. 30 Roommate Assignment Five hundred and five (83.3%) students had one roommate, 20 (3.3%) shared the room with two roommates, and 77 (12.7%) with three roommates. Students with more than one roommate were asked to choose one in order to answer the survey questions. In response to roommate assignment, more than two-thirds (69.1%) had their roommates randomly assigned to them, and 30.9% said that they chose their own roommates and that they had known him/her for an average of 4.25 years, ranging from 1 month to more than 18 years. Some students pointed out that they first met their roommates at school’s academic orientation program (AOP). When asked how well they knew the roommates, more than half of the participants (58.4%) reported their roommates were complete strangers, 10.1% nodding acquaintance, 10.6% acquaintance, 4.2% close acquaintance, 7.1% friend, and 8.6% close friend (see Table 3). Table 3 Descriptive Statistics for Students’ Roommate Assignment (N = 606) Variable % N One 83.3 505 Two 3.3 20 Three 12.7 77 More than three 0.7 4 My roommate was randomly assigned to me. 69.1 419 I chose my roommate. 30.9 187 Number of roommate(s) Roommate assignment 31 Table 3 (cont’d) Variable % N Complete stranger 58.4 347 Nodding acquaintance 10.1 60 Acquaintance 10.6 63 Close acquaintance 4.2 25 Friend 7.1 42 Close friend 8.6 51 How well do you know your roommate? Months for knowing the roommate M = 51.02 (SD = 59.06) Roommate-Related Facebook Experience Table 4 summarizes the information of roommate-related Facebook experience. Ninetythree percent of respondents reported that their roommates had Facebook profiles. Almost all the participants (99.5%) looked up their roommates on Facebook. More than one-fifth (21%) of the freshmen indicated that their parents looked at their roommates’ profiles too. Eighty-eight percent of the participants looked up their roommates’ profiles in 2010, and among them, 80.2% did this in August for the first time. Almost all the participants (99.5%) added or accepted their roommates as their Facebook friends. Nineteen percent of the freshmen logged on to Facebook to learn more about their roommates several times a week, and about one-third (32.7%) learned more about the roommates once a week. About 28% of the respondents interacted with their roommates on Facebook several times a week, and 27% said they interacted with them on 32 Facebook once a week. Among other communication channels students used to interact with their roommates, text messaging (60%) was the most used communication channel, followed by cell phone (45%) and email (40%). Some students also reported that they’ve talked to the roommates in person. In addition to Facebook, the average number of other communication channels they used to interact with their roommates is 1.68 out of 6. Table 4 Summary of Roommate-Related Facebook (FB) Experience for the Sample of the First-Wave Study (N = 606) Variable Mean (SD) % N Yes 92.6 553 No 2.6 16 I don’t know. 4.7 28 Yes 99.5 550 No 0.5 3 Yes 21.1 116 No 78.9 435 Yes 99.5 547 No 0.5 3 Roommate on FB Looked the roommate up on FB Parents looked at the roommate’s FB profile Roommate as a FB friend 33 Table 4 (cont’d) Variable Mean (SD) % N Several times per day 1.1 6 Once a day 6.1 33 Several times a week 19.2 104 Once a week 32.7 177 Several times a month 10.1 55 Monthly or less 20.7 112 Never 10.1 55 Several times per day 2.6 14 Once a day 7.1 39 Several times a week 27.8 152 Once a week 26.7 146 Several times a month 11.4 62 Monthly or less 14.5 79 Never 9.9 54 Email 39.8 241 Landline phone 5.0 30 Frequency of learning more about the roommate on FB Time interacting with roommate on FB Other communication channels used 34 Table 4 (cont’d) Variable Mean (SD) % N Cell phone 45.4 275 Text messaging 60.1 364 Instant messaging 7.4 45 Internet phone 10.4 63 Number of communication channels used* 1.68 (1.23) *FB was excluded. Information for Impression Formation on Facebook Among the information available on Facebook, pictures (73.6%), wall postings and comments (63.6%), favorites and interests (61.1%), status updates (53.7%), and background information (48.7%) were the top five items chosen by the freshmen that they believed to be most critical when forming an impression of their roommates (see Table 5). Some participants also mentioned other important information for impression formation, including their roommates’ (1) Facebook friends’ appearance, (2) interaction with their friends on Facebook, (3) friends’ comments about them, (4) language/conversation style (e.g., grammar, curse words, slang, etc.), (5) how frequently they update their Facebook profile, (6) how active they are on Facebook, (7) whether they smoke, drink, or party a lot, and (8) sexual orientation. More than 83.8% of the respondents agreed that they liked what they saw about their roommates on Facebook, and about 10% of the students agreed that they would prefer another roommate after looking them up on Facebook. 35 Table 5 Information for Impression Formation on Facebook Item* %** Mean SD Pictures 73.6 5.10 1.61 Wall 63.6 4.69 1.69 Favorites and Interests 61.1 4.57 1.72 Status 53.7 4.26 1.66 Background information 48.7 4.01 1.78 Notes 39.4 3.55 1.74 Groups 36.3 3.44 1.75 Religious views 35.8 3.50 1.88 Relationship status 33.5 3.44 1.78 Political views 25.0 3.11 1.72 Bookmarked pages 20.2 2.88 1.61 Facebook friends in common 15.5 2.57 1.67 Facebook friend amount 13.8 2.49 1.58 Website link 13.0 2.65 1.57 *. Students were asked to rate these individual items from 1 to 7, 1 as not at all important, and 7 as extremely important when they determine what their roommates would be like. **. Percentage of students who believe the individual items are more than moderately important (>4) to determine what their roommates would be like. 36 The Second-Wave Study: A Week after Living Together Procedure A second email invitation was sent to the participants who completed the first-wave survey on September 6, 2010, a week after college freshmen moved in with their roommates. A link to the second-wave online survey and an emphasis on the rules of receiving incentives—gift card drawing or extra credit—were included in the invitation. The second-wave survey was designed to record students’ offline uncertainty and offline impressions of the roommates. Students were asked to answer questions including whether they made requests to change roommates, perceived accuracy of roommates’ Facebook information, uncertainty level upon move-in, impressions of the roommates’ social, appearance, and task attractiveness upon movein (see Appendix B for the complete second-wave questionnaire). Participants There were 435 freshmen who completed the second-wave survey with a return rate of 72.8% from the first-wave sample (N = 606). Fifteen respondents (3.4%) reported that they had requested a roommate change. Six of them answered that the information they found about their roommates on Facebook had an influence on their decision; even so, thirteen of them still accepted to have their new roommates randomly assigned to them. For the purpose of this study, the responses from these fifteen participants were dropped in the following section (N = 420). At the time when they took the second survey, a total of 415 participants answered that they still maintained a Facebook profile, and 96.6% of them said that their roommates had Facebook profiles too. 61.7% of the respondents said they were on Facebook several times per day, and 28.3% were on Facebook once a day. As for the amount of time they spent on Facebook on a typical day, 36.7% of the freshmen replied they were on Facebook 1 to 2 hours, and 29.3% 37 said they spent less than an hour. On average, they spent 109.95 minutes (SD = 82.70) each day on Facebook upon the time they took the second-wave survey, which was similar to 101.35 minutes in the first-wave survey. The average number of their Facebook friends is 552.77 (SD = 352.28), ranging from 10 to 2,313, similar to the result of 550.75 friends in the first-wave study. Three hundred and ninety-six (95.4%) freshmen were their roommates’ Facebook friends. About four-fifths (83%) logged on to Facebook to learn more about their roommates and 77.5% interacted with their roommates on Facebook at least on a monthly basis in the past month at the time they took the survey. Two-thirds of the participants replied that they did not meet their roommates until the day they moved in together. In addition to Facebook, among the communication channels the participants used to interact with their roommates in the past month when they completed the second-wave survey, text messaging (76.2%) was the most used communication channel, followed by cell phone (62.4%), face-to-face (59.0%), and email (33.6%). The average number of communication channels they used to interact with their roommates is 3.27 (SD = 1.49) out of 8 (see Table 6). Table 6 Summary of Facebook (FB) Experience for the Sample of the Second-Wave Study (N = 420) Variable Mean (SD) % N Yes 98.8 415 No 1.2 5 Several times per day 61.7 253 Once a day 28.3 116 Facebook Profile Frequency on Facebook 38 Table 6 (cont’d) Variable Mean (SD) % N Several times a week 6.8 28 Once a week 2.0 8 Several times a month 0.5 2 Monthly or less 0.7 3 Less than 1 hour 29.3 120 1 – 2 hours 36.7 150 2 – 3 hours 18.3 75 3 – 4 hours 9.0 37 4 – 5 hours 2.7 11 5 – 6 hours 1.7 7 6 – 7 hours 1.2 5 7 – 8 hours 0.5 2 More than 8 hours 0.5 2 Yes 96.6 400 No 1.7 7 Time spent on Facebook on a typical day Daily minutes Facebook use* 109.95 (82.70) Number of FB friends 552.77 (SD = 352.28) Roommate on FB 39 Table 6 (cont’d) Variable Mean (SD) % N 1.7 7 Yes 95.4 396 No 4.6 19 Several times per day 2.0 8 Once a day 6.9 27 Several times a week 14.5 57 Once a week 21.6 85 Several times a month 11.2 44 Monthly or less 26.6 105 Never 17.3 68 Several times per day 4.8 19 Once a day 4.6 18 Several times a week 17.0 67 Once a week 17.2 68 Several times a month 14.2 56 Monthly or less 19.7 78 Never 22.5 89 I don’t know. Roommate as a FB friend Frequency of learning more about the roommate on FB Time interacting with the roommate on FB 40 Table 6 (cont’d) Variable Mean (SD) % N Yes 35.5 146 No 64.5 265 Face-to-face 59.0 248 Email 33.6 141 Landline phone 4.0 17 Cell phone 62.4 262 Text messaging 76.2 320 Instant messaging 7.4 31 Internet phone 11.9 50 Facebook 72.9 306 Meeting with the roommate before move-in Communication channels used in the past month 3.27 (1.49) *. Minutes of Facebook use were converted from an ordinal scale by assigning the midpoint of each response category, where less than 1 hour = 30 min, 1-2 hours = 90 min, 2-3 hours = 150 min, 3-4 hours = 210 min, 4-5 hours =270 min, 5-6 hours = 330 min, 6-7 hours = 390 min, 7-8 hours =450 min, more than 8 hours = 480 min. 41 Information Accuracy for Impression Formation on Facebook In the second-wave survey, students were asked to rate how accurate they felt about their roommates’ self-presentation on Facebook on a seven-point Likert-type scale after they had the chance to observe their roommates in real life. Table 7 summarizes the information of their perception about the accuracy of their roommates’ self-presentation on Facebook. The mean value for each item was above the mid-point value, suggesting that on average students perceived the information about their roommates on Facebook to be accurate. The top five items deemed to be more accurate were background information (83.4%), pictures and bookmarked pages (both were 72.6%), relationship status (71.1%), wall postings (68.0%), and status update (65.2%). Table 7 The Perception of the Accuracy of Roommates’ Self-Presentation on Facebook Item* %** Mean SD Background information 83.4 5.72 1.27 Pictures 72.6 5.21 1.38 Bookmarked pages 72.6 4.65 1.30 Relationship status 71.1 5.47 1.54 Wall 68.0 5.04 1.33 Status 65.2 4.95 1.37 Favorites and Interests 62.2 4.98 1.39 Groups 57.1 4.90 1.30 Religious views 55.3 4.90 1.37 Notes 50.8 4.81 1.30 42 Table 7 (cont’d) Item* %** Mean SD Facebook friend amount 50.1 4.52 1.47 Facebook friends in common 48.0 4.54 1.75 Website link 36.9 4.53 1.33 Political views 36.8 4.41 1.40 *. Students were asked to rate these individual items from 1 to 7, 1 as not at all accurate, and 7 as very accurate after they compared their roommates’ online and offline self-presentation. **. Percentage of students who believed the individual items are more than somewhat accurate (>4) after comparing their roommates’ online and offline self-presentation. The Third-Wave Study: Seven Weeks after Living Together Procedure The last email invitation was sent to the students who completed the second-wave survey on October 18, 2010, after they lived with their roommates for seven weeks. A link to the lastwave survey and the reminder of receiving incentives were included in the invitation. In this survey, students were asked to answer questions regarding their roommate relationships, including how much time they spent together, how many activities they did together, and how influential their roommates were on their daily decisions and future goals. Their levels of uncertainty and impressions of roommates’ attractiveness in the three dimensions were also measured in the questionnaire (see Appendix C for the complete third-wave questionnaire). 43 Participants A total of 349 students completed the last-wave survey, with a return rate of 80.2% from the previous sample (N = 435). Among them, 98.9% still had a Facebook profile, and 96.8% said their roommates were still on Facebook too. About 78% of the participants logged into Facebook several times a day, and 37.0% said they spent one to two hours each day on Facebook, while 24.6% spent two to three hours a day. The average minutes they spent on Facebook was 127.65 minutes (SD = 87.69). The mean value of their Facebook friends was 580.16 (SD = 358.50), with a huge difference from zero to 2,474. 78.2% of the respondents replied that they still got news/updates about their roommates from Facebook, and 58.5% said that they still interacted with their roommates on Facebook. The average number of their Facebook friends in common was 54.98 (SD = 111.97), ranging from zero to 733 (see Table 8). Table 8 Summary of Facebook (FB) Experience for the Sample of the Third-Wave Study (N = 349) Variable Mean (SD) % N Yes 98.9 345 No 1.1 4 Several times per day 77.8 267 Once a day 16.3 56 Several times a week 2.9 10 Once a week 1.2 4 Several times a month 0.6 2 Facebook Profile Frequency on Facebook 44 Table 8 (cont’d) Variable Mean (SD) % N 1.2 4 Less than 1 hour 19.1 65 1 – 2 hours 37.0 126 2 – 3 hours 24.6 84 3 – 4 hours 11.4 39 4 – 5 hours 2.9 10 5 – 6 hours 2.1 7 6 – 7 hours 1.2 4 7 – 8 hours 0 0 1.8 6 Yes 96.8 338 No 3.2 11 Monthly or less Time spent on Facebook on a typical day More than 8 hours Daily minutes Facebook use* 127.65 (87.69) Number of FB friends 580.16 (SD = 358.50) Roommate on FB FB friends in common with the roommate 54.98 (SD = 111.97) 45 Table 8 (cont’d) Variable Mean (SD) % N Yes 78.2 273 No 21.8 76 Yes 58.5 204 No 41.5 145 Getting news/updates about the roommate on FB Interact with the roommate on FB *. Minutes of Facebook use were converted from an ordinal scale by assigning the midpoint of each response category, where less than 1 hour = 30 min, 1-2 hours = 90 min, 2-3 hours = 150 min, 3-4 hours = 210 min, 4-5 hours =270 min, 5-6 hours = 330 min, 6-7 hours = 390 min, 7-8 hours =450 min, more than 8 hours = 480 min. Measurement Uncertainty Clatterbuck (1979) operationalized uncertainty through measures of attributional confidence. He argued that the process of reducing uncertainty can be divided into two interactive processes—retroactive explanation and proactive prediction. In the case of a first year student who has not met his or her prospective roommate, information available on Facebook seems to be a good source to get to know who his/her roommate is and what it would be like to live with him/her in the future. Previous research by Berger and Calabrese (1975) and Clatterbuck showed that retroactive attribution processes of uncertainty reduction were closely linked to proactive attribution process, and thus either one could be used to test any hypothesis involving uncertainty. Because the instrument for retroactive attributional confidence measures a 46 person’s confidence in how confident he is in giving specific facts about one target person, such as ―has he/she ever wanted a date and could not get it?‖ and ―what does he/she value more— friendship or money?‖, it does not suit for college freshmen who have not met their prospective roommates when they participated in the first-wave survey. For this reason, this research focuses on the proactive attribution of uncertainty reduction processes that the freshmen engaged in, also known as proactive attributional confidence by Clatterbuck. A seven-item CLUES scale (CL7; Clatterbuck, 1979) was revised to measure first year students’ proactive attributional confidence in predicting their future roommates. Participants were asked to answer the extent to which they agree with statements such as ―I am confident of my general ability to predict how my roommate will behave,‖ and ―I can predict my roommate’s feelings and emotions‖ in a seven-item seven-point Likert-type scale (see Table 25 and Table 26 in Appendix E for complete scale items, descriptives, and intercorrelations). High internal consistency reliability was found for this measure across all three stages (αs > .91). The value of the scale mean was obtained by taking average of the item means (see Table 9). Impressions of the Roommates’ Social, Appearance, and Task Attractiveness A fifteen-item interpersonal attraction scale (J. C. McCroskey & McCain, 1974) measuring social attractiveness, appearance attractiveness, and task attractiveness was revised to measure students’ impressions of their roommates at each stage. Instead of using the commonlyused big five personality test, the interpersonal attraction scale was able to capture more individual variances in how freshmen liked or disliked their roommates. Students rated on sevenpoint Likert-type scales about their impressions of their roommates’ social attractiveness, appearance attractiveness, and task attractiveness (see Table 27-32 in Appendix E for complete scale items, descriptives, and intercorrelations). The reliability tests for three subscales at three 47 stages—social, appearance, and task attractiveness—revealed moderate to high internal consistency (.79 < αs < .89; see Table 9). Table 9 Summary of Scale Reliability, Means, and Standard Deviations for CL7, Social, Appearance, and Task Attractiveness for the Three-Wave Study First Wave Second Wave Third Wave Items α M SD α M SD α M SD CL7 7 .94 3.82 1.46 .91 4.81 1.10 .92 4.98 1.27 Social 5 .79 5.54 .98 .87 5.34 1.16 .89 5.15 1.45 Appearance 5 .79 4.35 1.10 .82 4.30 1.10 .83 4.23 1.15 Task 5 .82 5.13 .94 .83 5.19 1.03 .89 5.02 1.29 Scale Relationship Closeness Inventory The relationship closeness inventory (RCI) developed by Berscheid, Snyder, and Omoto (1989) has three subscales: frequency, diversity, and strength. The frequency dimension measures the amount of time that students spend interacting with their roommates face to face per day (see Table 33 in Appendix E for complete scale items). The respondents were asked to estimate the amount of time they spent alone with their roommates in the morning, afternoon, and evening on a typical day within the past week. By dissecting a typical day into three time estimates, the researcher was able to obtain a more accurate estimate of total amount of time they spent together. Berscheid and her colleagues treated these time estimates as a separate indicator of the total frequency score and performed a reliability test on this frequency subscale. The alpha for the frequency scale in this study was .75, and the mean value for the total amount of time 48 spent together was 384.76 minutes (SD = 236.32), with minimum of 0 minutes to maximum of 1,140 minutes (the ceiling value was 1,200 min). Before converting the total amount of time that roommates spent together into a 10-point Frequency scale created by Berscheid and her colleagues (see Table 36 in Appendix E for the conversion table), missing data in each of the three time estimates were replaced with the item mean, and then the three time estimates were summed up to get the total amount of time spent together. After the transformation procedure, the mean value for the frequency was 5.96 (SD = 1.98). The diversity scale measures the number of different activities that the roommates do together (see Table 34 in Appendix E for complete scale items). Two inapplicable items (i.e., went to an auction/antique show and engaged in sexual relations) were dropped from the original 38-item scale. Students were asked to answer whether they had done each of the 36 different activities alone with their roommates in the past week. Because the responses were dichotomous, a Kuder-Richardson reliability test was performed to check for internal consistency. The alpha is .91. The mean value for the total amount of the activities that the participants performed together with their roommates was 8.29 (SD = 5.99). The missing data were then replaced with the mean before transforming the data into a 10-point scale created by Berscheid and her colleagues. The mean value for the diversity was 4.81 (SD = 1.72) after the transformation procedure (see Table 36 in Appendix E for the conversion table). The strength dimension measures the extent to which the participants are influenced by their roommates on both current and future goals, decision-making, behaviors, and plans (see Table 35 in Appendix E for complete scale items). Respondents were asked to rate the extent of influence from their roommates on 34 items, such as ―My roommate influences which parties and other social events I attend,‖ ―My roommate influences the way I feel about the future,‖ and 49 ―My roommate influences how I decorate my dorm room,‖ using a seven-point Likert-type scale with a higher score indicating greater perceived influences. After reverse-coding some items in the scale, a raw total strength value was obtained by summing up all items, ranging from 34 to 238. The alpha is .92. The mean for the raw total strength value was 93.29 (SD = 28.33). Again, the missing data were replaced with the mean before converting the data into a 10-point scale that Berscheid and her colleagues (1989) created (see Table 36 in Appendix E for the conversion table). The mean for the converted strength scores was 3.46 (SD = 1.40). The last step to create an overall index of relationship closeness was to sum across the scores of converted frequency, diversity, and strength (Berscheid, et al., 1989), with higher scores indicating greater closeness between roommates. The mean for the overall closeness scores was 14.24 (SD = 4.04), ranging from 3 to 30. The Cronbach’s alpha for this three-item RCI index was .69 (see Table 10 for descriptive statistics for the three subscales of RCI). Table 10 Descriptive Statistics for the Raw and the Converted Total Frequency, Diversity, Strength, and RCI Scores (α = .69) Raw Value Converted Value α M SD N M SD N Frequency .75 384.76 236.32 294 5.96 1.98 349 Diversity .91 8.29 5.99 337 4.81 1.72 349 Strength .92 93.29 28.33 322 3.46 1.40 349 RCI .69 14.24 4.04 349 Variable 50 Research Design One goal of this study was to investigate the role Facebook played in affecting first year students’ impressions of their roommates and their relationship development in situations where roommates were not previously acquainted. Two filter questions based on the roommate assignment (i.e., randomly assigned or not) and the degree of their acquaintance with the roommates (i.e., complete stranger or not) were developed to filter out the students who already knew their roommates. The initial analysis of the data showed that the degree of the acquaintance was able to better capture the characteristics of the targeted subjects. In the end, there were two hundred and two participants who satisfied the criteria and finished all three sets of survey. Several hypotheses in the study predicted that the differences in Facebook and real-life impressions of the roommates would affect students’ certainty level about their roommates as well as their relationship closeness. Four possible conditions as discussed in the previous chapter were derived from positive or negative initial Facebook impressions against positive or negative offline impressions, a 2 x 2 design. The initial Facebook impressions could be either confirmed (i.e., impressions remained positive or negative) or conflicted (i.e., impressions got positive or negative) with the impressions formed after living together with the roommates. In order to separate the targeted sample into four groups based on the types of impressions they formed about the roommates on Facebook and in real life, the researcher performed a median split on both online (Mdn = 4.83 on a 7-point scale) and offline (Mdn = 4.87 on a 7-point scale) overall impressions. The overall impressions were obtained by taking average on impressions of roommates’ social, appearance, and task attractiveness scales. Freshmen in Group 1 were those who had positive initial Facebook impressions of their roommates which turned into negative impressions after moving in; people in Group 2 were those who had 51 negative impressions of their roommates both online and offline; people in Group 3 had the same positive online and offline impressions of their roommates; and finally, people in Group 4 were those whose impressions of their roommates changed from negative to positive (see Table 11). Table 11 Impression Discrepancies in Four Different Conditions Real Life Positive Negative Group 3 Group 1 Impressions remained positive Impressions from positive to N = 68 negative Facebook Positive N = 38 Negative Group 4 Group 2 Impressions from negative to Impressions remained negative positive N = 63 N = 33 Interviews Nineteen in-depth phone interviews were conducted from November 9 to November 22 in 2010. These students were recruited from the participants who completed the three-wave survey. Eleven of them are Females. One has known her roommate for over 10 years. The rest did not know their roommates at all when they received their roommate assignments. As for their changes in initial Facebook and offline impressions, five students were in Group 1, one in Group 2, five in Group 3, and eight in Group 4. 52 Interview questions are structured into three parts: (1) the first part focuses on their perceptions of roommates’ profiles and how information on the profiles helped to reduce/increase their uncertainty; (2) the second part focuses on whether there were any surprises or differences about the roommates, how they dealt with these surprises, and whether their relationship was affected by these surprises; and (3) the last part focuses on their relationship by the time they participated in the interviews. These interviews are able to provide detailed information about what they thought about their roommates’ profiles, which piece of information on Facebook was particularly important for them to form impressions, how they interacted with their roommates before move-in, how they dealt with the discrepant impressions, how they interacted with their roommates after living together, and how close they were. The wealth of detail in these responses is able to provide supplementary supports for this study (see Appendix D for interview questions). 53 CHAPTER 4 RESULTS Initial analyses examined whether participants’ information-seeking behaviors and impressions of the roommates affected their uncertainty. Hypothesis 1 and 3 were tested through multiple linear regression analyses performed in two steps. In the first step, information-seeking behaviors were entered. Respondents’ impressions of roommates’ social, appearance, and task attractiveness were entered in the second step. Multivariate analyses of variance (MANOVA) were performed to test Hypothesis 2 and 4 to look at all dependent variables at once to avoid Type I error caused by conducting several univariate tests while examining group differences in information-seeking behaviors and impressions. One-way ANOVAs were performed to test Hypothesis 5 and 6 to examine the extent of influence that impression discrepancies have on participants’ levels of uncertainty upon move-in together and the degrees of the closeness in the relationship with their roommates. Hypothesis 7 and 8 were testing through the mix-design ANOVAs in order to examine the effects of time (i.e., before, upon move-in, and move-in for a while) and groups (i.e., four groups according to types of impression discrepancies) on participants’ uncertainty and impressions at three different stages. Hypotheses Testing for the First-Wave Study The analyses for the first-wave data focused on the participants who did not know their roommates before move-in and had added or accepted their roommates as their Facebook friends. After filtering out the unsuitable cases, the first-wave sample size was N = 302. The intercorrelations between the frequency of checking a roommate’s information on Facebook, the frequency of roommate interaction on Facebook, the total number of channels used (up to seven channels) to communicate with the roommate, overall initial Facebook 54 impressions as well as the specific initial Facebook impressions of the social, appearance, and task attractiveness, and the level of certainty about the roommate before moving in are shown in Table 12. The analyses suggested that information-seeking frequency (r = .14, p < .05), interaction frequency (r = .24, p < .01), number of communication channels (r = .23, p < .01), and overall initial impressions (r = .30, p < .01) had a weak positive relationship with respondents’ certainty level about their roommates. Table 12 Intercorrelations for Information-Seeking Frequency, Interaction Frequency, Number of Channels, Facebook Impressions, and Certainty about the Roommates in the First-Wave Study Variable 1. FB Info-seeking 2. FB Interaction 1 2 3 4 a b c 5 -.58** -- 3. Number of Channels .02 .17** -- 4. First Impression: .14* .19* .29** -- a. Social .13* .19** .22** .88** -- b. Appearance .11 .12 .26* .86** .61** -- c. Task .08 .15* .18** .77** .63** .42** -- 5. Certainty .14* .24** .23** .30** .24** .22** .28** *. Correlation is significant at the .05 level (2-tailed). **. Correlation is significant at the .01 level (2-tailed). 55 -- Hypothesis 1 H1 predicted that before moving in with the roommate, (A) the frequency that a student invests in gathering information about the roommate on Facebook, the interaction with the roommate on Facebook, the number of channels used, and (B) the impression of the roommate’s social, appearance, and task attractiveness would have a positive impact on reducing a student’s uncertainty about the future roommate. A multiple linear regression was calculated to predict students’ level of certainty about the roommates based on H1 (A): their information-seeking frequency, interaction frequency, the number of channels, and H1 (B): the impression of roommates’ social, appearance, and task attractiveness. A significant regression equation was found (see step 1 in Table 13), F (3, 211) = 8.89, p < .001, with adjusted R² of .11. The analysis for H1 (A) showed that interaction frequency (β = .21, p < .01) and the number of communication channels used (β = .23, p < .01) were significant predictors, while informationseeking frequency (β = -.001, p > .05) was not. Thus, more Facebook interaction and more channels used to communicate with the roommates appeared to increase students’ level of certainty. Hence, H1 (A) was partially supported. Adding students’ impressions of their roommates’ social, appearance, and task attractiveness into the analysis resulted in a significant regression equation (see step 2 in Table 13), F (6, 208) = 8.64, p < .001, with adjusted R² of .18. The change in R² from step 1 to step 2 was significant, R² change = .09, p < .001. This result indicates that among the six variables being examined, Facebook interaction (β = .17), number of communication channels (β = .14), impressions of roommates’ appearance attractiveness (β = .15), and task attractiveness (β = .23) are significant predictors of students’ uncertainty level. Therefore, H1 (B) was partially supported. 56 Table 13 Regression Coefficients for Students’ Certainty about their Roommates in the First-Wave Study Variable Step 1 Step 2 Info-Seeking Frequency -.001 -.03 Interaction Frequency .21** .17* Number of Channels .23** .14* Social Attractiveness -- -.02 Appearance Attractiveness -- .15* Task Attractiveness -- .23** F 8.89*** 8.64*** df 3, 211 6, 208 SE 1.06 1.02 Adjusted R² .10 .18 .11*** .09*** R² Change Note: Standardized regression coefficients were shown. *. p < .05; **. p < .01; ***. p < .001 Hypothesis 2 H2 predicted that before moving in with the roommates, students with positive impression of their roommates would go on to Facebook more frequently gathering the roommates’ information, have more Facebook interactions with the roommates, use more channels to communicate with the roommates, and have greater certainty about their roommates, compared to students whose impressions of the roommates were negative. A one-way MANOVA was calculated to examine the effect of initial Facebook impressions on information- 57 seeking frequency, interaction frequency, number of communication channels used, and respondents’ certainty about their roommates. The researcher performed a median split on the overall first impressions (Mdn = 4.93 on a seven-point scale) to divide participants into two groups (positive vs. negative impressions). Table 14 Summary of Means and Standard Deviations for Variables by the Types of the Initial Facebook Impressions and Cronbach’s Alphas for Scales in the First-Wave Study Overall Variable α items Positive Negative M SD M SD M SD FB Info-Seeking Frequency 3.70 1.35 3.75 1.39 3.63 1.30 FB Interaction Frequency 3.77 1.50 3.86 1.50 3.67 1.49 Number of Channels 2.19 1.05 2.36 1.08 1.98 .97 Certainty .89 7 3.19 1.12 3.43 1.10 2.90 1.08 Overall First Impressions .77 3 4.90 .85 5.57 .47 4.18 .53 a. Social .78 5 5.36 .94 5.97 .62 4.65 .74 b. Appearance .83 5 4.43 1.28 5.23 .87 3.52 1.02 c. Task .79 5 4.96 .84 5.48 .66 4.38 .60 A significant effect was found (Lambda (4, 292) = 6.06, p < .001), indicating there was an overall significance of the model. The differences in degrees of information-seeking behaviors and levels of uncertainty simultaneously existed between groups. Follow-up univariate ANOVAs indicated that a respondent’s number of communication channels used (F (1, 295) = 11.47, p < .01) and certainty about the roommate (F (1, 295) = 17.05, p < .001) were 58 influenced by his/her initial Facebook impression; however, information-seeking frequency on Facebook (F (1, 295) = .93, p =.34) and the frequency of roommate interaction on Facebook (F (1, 295) = 1.69, p = .20) were not (see Table 14). Thus, H2 was partially supported. Hypotheses Testing for the Second-Wave Study In addition to the filter criteria (i.e., roommate was a stranger and a Facebook friend) for the first-wave study, to conduct hypotheses testing for the second-wave study, participants who had changed their initial roommate assignment were also filtered out from the sample. The resulting sample size in the second-wave study was thus N = 203. Table 15 Intercorrelations for Information-Seeking Frequency, Interaction Frequency, Number of Channels, Offline Impressions, and Certainty about the Roommates in the Second-Wave Study Variable 1. FB Info-seeking 2. FB Interaction 1 2 3 4 a b c 5 -.47** -- 3. Number of Channels .08 .29** -- 4. Offline Impression: .16* .36** .25** -- a. Social .11 .33** .30** .87** -- b. Appearance .16* .33** .28** .79** .58** -- c. Task .12 .18** .003 .72** .49** .26** -- .25** .34** .31** .52** .52** .46** .25** 5. Certainty *. Correlation is significant at the .05 level (2-tailed). **. Correlation is significant at the .01 level (2-tailed). 59 -- The correlation coefficients indicated that respondents’ impressions of their roommates’ social attractiveness (r = .52, p < .01), appearance attractiveness (r = .46, p < .01), and task attractiveness (r = .25, p < .01) had a moderate positive relationship with their degree of certainty about their roommates at the second stage. The analysis also showed that the correlation between the overall offline impressions and the degree of certainty was significant, r = .52, p < .01 (see Table 15). Hypothesis 3 The third hypothesis predicted a positive relationship between a first year student’s offline impression of the roommate’s (1) social, (2) appearance, and (3) task attractiveness and certainty about the roommate. A multiple linear regression analysis was used to analyze the impact of information-seeking frequency on Facebook, Facebook interaction frequency, number of channels, and offline impressions on students’ uncertainty level about their roommates after moving in together. The result was significant, F(6, 190) = 17.94, p < .001, adjusted R² = .34. The change in R² from step 1 to step 2 was significant, R² change = .18, p < .001.The analysis showed that after controlling for information-seeking on Facebook (β = .15, p < .05), Facebook interaction (β = .07, p > .05), and number of channels used to communicate with roommates (β = .13, p < .05), students’ offline impressions of roommates’ social attractiveness (β = .34, p < .001) and appearance attractiveness (β = .18, p < .05) were significant predictors of students’ certainty about the roommates, while task attractiveness (β= .003, p > .05) was not (see step 2 in Table 16). H3 was partially supported. 60 Table 16 Regression Coefficients for Students’ Certainty about their Roommates in the Second-Wave Study Variable Step 1 Step 2 .14 .15* Interaction Frequency .21** .07 Number of Channels .24*** .13* Social Attractiveness -- .34*** Appearance Attractiveness -- .18** Task Attractiveness -- .003 F 14.13 17.94 df 3, 193 6, 190 SE .89 .79 Adjusted R² .17 .34 .18*** .18*** Info-Seeking Frequency R² Change Note: Standardized regression coefficients were shown. *. p < .05; **. p < .01; ***. p < .001 Hypothesis 4 The fourth hypothesis predicted that students who had positive offline impressions of their roommates would (a) gather their roommates’ information more often on Facebook, (b) have more Facebook interaction, (c) use more communication channels, and (d) have greater level of certainty about their roommates than those who had negative offline impressions. A 61 median split (Mdn = 4.93 on a 7-point scale) was performed on students’ overall offline impressions. Table 17 Summary of Means and Standard Deviations for Variables by the Types of the Offline Impressions and Cronbach’s Alphas for Scales in the Second-Wave Study Overall Variable α Negative M SD M SD M SD FB Info-Seeking Frequency 3.29 1.55 3.47 1.59 3.09 1.50 FB Interaction Frequency 3.13 1.78 3.62 1.74 2.61 1.67 Number of Channels 3.03 1.31 3.28 1.24 2.78 1.34 7 4.64 .97 5.08 .82 4.16 .89 .71 3 4.95 .78 5.59 .42 4.30 .45 a. Social .85 5 5.33 .99 6.00 .60 4.63 .81 b. Appearance .83 5 4.33 1.03 5.02 .73 3.63 .79 c. Task .80 5 5.20 .91 5.73 .65 4.64 .82 Certainty Overall Offline Impressions items Positive A one-way MANOVA was then used to examine the effect of offline impression (positive or negative) on students’ information-seeking and roommate interaction on Facebook, number of communication channels used to communicate with the roommates, and finally, the certainty about their roommates after moving in. The analysis showed a significant effect, Lambda (4, 197) = 16.29, p <.001. Follow-up univariate ANOVAs indicated that except for information-seeking on Facebook (F (1, 200) = 2.64, p = .11), the other three variables— Facebook interaction with roommates (F (1, 200) = 16.95, p < .001), number of communication 62 channels used (F (1, 200) = 7.14, p <.01), and degree of certainty about the roommates after moving in (F (1, 200) = 59.21, p <.001)—were significantly different between students with a positive impression and those with a negative impression (see Table 17). Thus, H4 was partially supported. Hypothesis 5 H5 predicted that freshmen with different types of impression discrepancies (i.e., Group 1: impressions changed from positive to negative, Group 2: impressions remaining negative, Group 3: impressions remaining positive, and Group 4: impressions changed from negative to positive) would have different levels of uncertainty about their roommates in the beginning of their living together. An analysis of variance revealed a significant result, F (3, 198) = 20.60, p < .001. Table 18 Summary of Means and Standard Deviations for Scores on Students’ Certainty at the Second Stage by Different Impression Discrepancies Impression Discrepancies M SD N Group 1: Positive to Negative 3.96 1.00 32 Group 2: Remaining Negative 4.25 .82 65 Group 3: Remaining Positive 5.09 .86 78 Group 4: Negative to Positive 5.07 .70 27 Total 4.64 .97 202 Post hoc comparisons using Tukey’s HSD procedure (p < .05) indicated that students in Group 3 (M = 5.09, SD .86) and Group 4 (M = 5.07, SD = .70) had a greater degree of certainty about their roommates than students in Group 1 (M = 3.96, SD = 1.00) and Group 2 (M = 4.25, 63 SD = .82), whereas students in Group 3 and Group 4 were not significantly different in their scores of certainty, and students’ level of uncertainty about their roommates in Group 1 and Group 2 were not significantly different, either (see Table 18). H5 was supported. Hypotheses Testing for the Third-Wave Study The analyses in this section focus on the participants who completed all three waves of the survey and did not know their roommates before they moved in together (N = 202). Hypothesis 6 H6 predicted that discrepancies in initial Facebook and offline impressions upon move-in would affect the degrees of closeness in first year students’ roommate relationship. Four types of impression discrepancies were obtained after comparing the initial Facebook impressions (positive or negative after performing median split, Mdn = 4.88 on a 7-point scale) and the offline impressions (positive or negative, after median split was performed, Mdn = 4.86). A oneway ANOVA was calculated comparing the relationship closeness with roommates by the four different types of impression discrepancies. A significant difference was found among the groups, F (3, 198) = 8.12, p < .001. Tukey’s HSD (p < .05) was used to determine the nature of the differences between the groups. The analysis indicated that freshmen in Group 3 (M = 14.72, SD = 4.21) and Group 4 (M = 14.85, SD = 3.32) developed a much closer relationship with their roommates than those in Group 1 (M = 12.00, SD = 3.08) and Group 2 (M = 12.21, SD = 3.95). Neither freshmen in Group 3 and Group 4 nor those in Group 1 and Group 2 were significantly different from each other (see Table 19). Hence, H6 was partially supported. However, the mean value for each group did show that Group 4 had the highest level of relationship closeness, followed by Group 3, Group2, and then Group 1, as predicted in H6. 64 Table 19 Summary of Means and Standard Deviations for Scores on Students’ Relationship Closeness Index (RCI) by Different Impression Discrepancies Group M SD N Group 1: Positive to Negative 12.00 3.08 38 Group 2: Remaining Negative 12.21 3.95 63 Group 3: Remaining Positive 14.72 4.21 68 Group 4: Negative to Positive 14.85 3.32 33 Total 13.45 4.00 202 Note: RCI index ranges from 3 to 30, after combining the converted subscales of frequency, diversity, and strength. Hypothesis 7: Changes in Certainty H7 predicted that (A) first year students’ uncertainty level would decrease over time; (B) their uncertainty level would be different among groups (i.e., four types of impression discrepancies); and (C) there would be an interaction effect of time and groups on students’ levels of uncertainty. A 3 x 4 mixed-design ANOVA was calculated to examine the hypothesis. Mauchly’s test indicated that the assumption of sphericity had been violated, χ² (2) = 26.56, p < .001, therefore degrees of freedom were corrected using Huynh-Feldt estimates of sphericity (ɛ = .91). The analysis revealed a significant main effect for time, F (1.82, 352.11) = 190.97, p < .001. Bonferroni corrected post hoc tests at p < .05 showed that students’ certainty levels about their roommates at the three different stages were significantly different from each other. The degree of certainty increased from the first stage (M = 3.10) to the second stage (M = 4.52), and then from the second stage to the third stage (M = 4.81). H7 (A) was supported (see Figure 1). 65 A significant main effect was also found for groups, F (3, 194) = 15.83, p < .001. Freshmen in Group 3 (M = 4.59) and Group 4 (M = 4.36), that is, students who had a positive or a positive change in impressions, had a greater degree of certainty about their roommates than students who had a negative or a negative change in impressions in Group 1 (M = 3.64) and Group 2 (M = 3.87); however, students in Group 4 and Group 3 were not significantly different from each other, and the same situation applied to students in Group 1 and Group 2 (see Figure 2). H7 (B) was supported. Moreover, the analysis showed that a significant Time x Groups interaction was present, F (5.45, 352.11) = 2.32, p < .05. Students’ levels of certainty at three different stages were significantly different based on whether their impressions of the roommates were getting better, worse, remaining positive, or negative (see Figure 3 and Table 20). H7(C) was supported. Upon examination of the data, it appeared that although students in each group had similar pattern of increase in certainty level, students in Group 4 had the most increase in their certainty, especially from Time 1 to Time 2. Figure 1 Figure 2 Changes in Certainty at the Three Stages Changes in Students’ Certainty by Groups 6 4 4.52 5 4.81 3.64 3.87 4.59 3.1 0 2 Certainty 0 Certainty Stage 1 Stage 2 Group 1 Group 3 Stage 3 66 Group 2 Group 4 4.36 Figure 3 Changes in Students’ Certainty at the Three Stages by Groups Certainty First Stage Second Stage Third Stage 5.02 4.19 3.88 4.11 2.93 Group 1 5.27 4.9 4.57 3.48 3.03 2.85 Group 2 Group 3 Group 4 Table 20 Summary of Descriptive Statistics for Students’ Scores of Certainty Time Changes in Impression First Wave M SD N Group 1 2.93 1.15 37 Group 2 2.85 1.08 63 Group 3 3.48 1.13 65 Group 4 3.03 1.06 33 Total Second Wave 5.15 3.10 1.13 198 Group 1 3.88 1.02 37 Group 2 4.19 .86 63 Group 3 5.02 .81 65 67 Table 20 (cont’d) Time M SD N Group 4 4.90 .89 33 Total 4.52 .99 198 Group 1 4.11 1.46 37 Group 2 4.56 1.18 63 Group 3 5.27 1.03 65 Group 4 5.15 1.06 33 Total Third Wave Changes in Impression 4.81 1.24 198 Hypothesis 8: Changes in Impressions H8 predicted that: (A) students’ impressions of the roommates’ social, appearance, and task attractiveness will decrease significantly from Time 1 (i.e., before moving in) to Time 2 (i.e., upon moving in), and no significant changes will be found between Time 2 and Time 3; (B) their impressions of the roommates’ attractiveness in the three dimensions will be different among groups; and (C) there will be an interaction effect of time and groups on students’ impressions of the roommates’ attractiveness in the three dimensions. Impressions of the Social Attractiveness A 3 x 4 mixed factorial ANOVA was calculated to examine the effects of time and groups on students’ impressions of their roommates’ social attractiveness. Mauchly’s test indicated that the assumption of sphericity was violated, χ² (2) = 42.34, p < .05, therefore degrees of freedom were corrected using Huynh-Feldt estimates of sphericity (ɛ = .85). The analysis revealed a significant main effect for time, F (1.70, 322.07) = 8.85, p < .001. Follow-up 68 protected dependent t tests using Bonferroni procedure at p < .05 revealed a significant decrease from the first stage (M = 5.33) to the second (M = 5.16), while there was no significant changes between the second and the third stage (M = 5.01). H8 (A) was supported for social attractiveness. See Figure 4. A significant main effect was also found for the four different groups, F (3, 189) = 3 56.168 , p < .001. Post hoc tests using Bonferroni procedure (p < .05) showed that except for Group 1 (M = 4.94) and Group 4 (M = 5.31) which were not significantly different from each other, other differences among groups were significant (mean for Group 2 = 4.36, and mean for Group 3 = 5.99). H8 (B) was supported for social attractiveness (see Figure 5). In addition, the result indicated that there was a significant Time x Groups interaction, F (5.11, 322.07) = 9.70, p < .001. Students’ impressions of their roommates’ social attractiveness at three different stages were significantly different among groups (see Figure 6 and Table 21). Upon examination of the data, it appears that students in Group 1 showed the most decrease in their impressions of roommates’ social attractiveness from Time 1 to Time 2; conversely, students in Group 4 had an increase in their impressions of roommates’ social attractiveness upon moving in; while students in Group 2 and Group 3 had relatively steady impressions of their roommates’ social attractiveness across three stages. H8 (C) was supported for social attractiveness. 3 The accuracy of this F-test ratio was compromised, because Levene’s tests indicated that variances were not homogeneous. Attempts to stabilizing the variances between groups by performing log transformation of the data did not succeed. Levene’s test still indicated the variances were significantly different. 69 Figure 4 Figure 5 Changes in Impressions of the Roommates’ Changes in Students’ Impressions of the Social Attractiveness at the Three Stages Roommates’ Social Attractiveness by Groups 5.33 5.4 8 6 4 2 0 5.16 5.2 5.01 5 4.8 4.94 Social Attractiveness Stage 1 Stage 2 5.99 4.36 5.31 Social Attractiveness Stage 3 Group 1 Group 2 Group 3 Group 4 Figure 6 Changes in Students’ Impressions of the Roommates’ Social Attractiveness at the Three Stages by Groups Social Attractiveness First Stage Second Stage Third Stage 6.02 6.09 5.87 5.78 5.68 5.02 4.53 4.51 Group 1 5.22 4.53 4.28 4.28 Group 2 Group 3 70 Group 4 Table 21 Summary of Descriptive Statistics for Impressions of Social Attractiveness Time Changes in Impression First Wave N Group 1 5.78 .90 35 4.53 .70 61 Group 3 6.02 .57 64 Group 4 5.02 .56 33 Total 5.33 .94 193 Group 1 4.53 1.13 35 Group 2 4.28 .96 61 Group 3 6.09 .63 64 Group 4 5.68 .54 33 Total 5.16 1.16 193 Group 1 4.51 1.34 35 Group 2 4.28 1.46 61 Group 3 5.87 1.08 64 Group 4 5.22 1.32 33 Total Third Wave SD Group 2 Second Wave M 5.01 1.46 193 Impressions of the Appearance Attractiveness The same two-way mixed-design procedure was performed to examine the effects of time and groups on impressions of roommates’ appearance attractiveness. Mauchly’s test showed that the assumption of sphericity was violated, χ² (2) = 11.34, p < .01, therefore degrees of freedom 71 were corrected using Huynh-Feldt estimates of sphericity (ɛ = .97). A significant main effect was found for time, F (1.94, 348.39) = 3.54, p < .05. Follow-up protected dependent t tests using Bonferroni procedure (p < .05) showed that students’ impressions of their roommates’ appearance attractiveness decreased from the first stage (M = 4.38) to the second stage (M = 4.21) when p = .06, while there was no significant changes from the second stage to the third stage (M = 4.21). H8 (A) was marginally supported for appearance attractiveness (see Figure 7). The analysis also showed there was a significant main effect for groups, F (3, 180) = 4 69.81 , p < .001. Group comparisons using Bonferroni procedure at p < .05 suggested significant differences among groups (Group 1 = 4.38, Group 2 = 3.38, Group 3 = 5.19, and Group 4 = 4.11), except for Group 1 and Group 4 which did not differ from each other (see Figure 8). H8 (B) was supported for appearance attractiveness. A significant interaction between Time x Groups was found, F (5.81, 348.39) = 10.56, p < .001. Like social attractiveness, students’ impressions of their roommates’ appearance attractiveness at three different stages were significantly different in the types of impression changes they had towards the roommates (see Figure 9 and Table 22). Upon examining the data, it appears that students in Group 1, 2, and 3 had a similar pattern of changes in their impressions of roommates’ appearance attractiveness, but students in Group 1 had the most decrease, while students in Group 4 had an increase in their perceptions of roommates’ appearance attractiveness. H8(C) was supported for appearance attractiveness. 4 However, the accuracy of this F-test ratio was compromised, because Levene’s test indicated that variances were only homogeneous at the first stage. Attempts to stabilizing the variances between groups by performing log transformation of the data did not succeed. 72 Figure 7 Figure 8 Changes in Students’ Impressions of the Changes in Students’ Impressions of the Roommates’ Appearance Attractiveness at Roommates’ Appearance Attractiveness by the Three Stages Groups 4.38 4.4 4.3 4.2 4.1 6 4.21 4.21 5.19 4.38 4.11 3.38 4 2 0 Appearance Attractiveness Stage 1 Stage 2 Appearance Attractiveness Stage 3 Group 1 Group 2 Group 3 Group 4 Figure 9 Changes in Students’ Impressions of the Roommates’ Appearance Attractiveness at the Three Stages by Groups Appearance Attractiveness First Stage Second Stage 5.39 4.97 3.94 Third Stage 5.09 5.08 4.48 4.22 3.72 3.41 3.33 3.4 Group 1 Group 2 Group 3 73 4.13 Group 4 Table 22 Summary of Descriptive Statistics for Impressions of Appearance Attractiveness Time Changes in Impression First Wave N Group 1 4.97 .96 25 3.41 1.00 62 Group 3 5.39 .79 64 Group 4 3.72 .99 33 Total 4.37 1.28 184 Group 1 3.94 .62 25 Group 2 3.33 .95 62 Group 3 5.09 .73 64 Group 4 4.48 .58 33 Total 4.23 1.08 184 Group 1 4.22 .70 25 Group 2 3.40 1.06 62 Group 3 5.08 .82 64 Group 4 4.13 .92 33 Total Third Wave SD Group 2 Second Wave M 4.23 1.14 184 Impressions of the Task Attractiveness A two-way mixed factorial ANOVA was calculated to examine the effects of time and groups on students’ impressions of their roommates’ task attractiveness. Mauchly’s test indicated that the assumption of sphericity was violated, χ² (2) = 30.62, p < .001, therefore degrees of 74 freedom were corrected using Huynh-Feldt estimates of sphericity (ɛ = .89). The analysis revealed a significant main effect for time, F (1.78, 322.97) = 5.77, p < .01. Follow-up protected dependent t tests using Bonferroni procedure at p < .05 revealed a significant increase from the first stage (M = 4.90) to the second stage (M = 5.19), while there was no significant changes between the second and the third stage (M = 4.99). The result showed a different direction in changes in students’ impressions of roommates’ task attractiveness (see Figure 10). H8 (A) was not supported for task attractiveness. 5 A significant main effect was also found for the four different groups, F (3, 182) = 36.94 , p < .001. Post hoc comparisons using Bonferroni procedure (p < .05) showed that Group 1 (M = 4.79) and Group 2 (M = 4.44) were not significantly different from each other, and Group 1 and Group 4 (M = 5.20) were not different, either. Other than those two comparisons, significant differences among groups were found (Group 3 = 5.67). H8 (B) was supported for task attractiveness. See Figure 11. In addition, the result indicated that there was a significant Time x Groups interaction, F (5.32, 322.97) = 6.40, p < .001. Students’ impressions of their roommates’ task attractiveness at three different stages were significantly different in the types of impression changes (see Figure 12 and Table 23). Upon examination of the data, students in Group 2, 3, and 4 followed a similar pattern of changes that their impressions of roommates’ task attractiveness increased upon moving in, but then decreased after living for a while; students in Group 4 had the most increase 5 The accuracy of this F-test ratio was compromised. Levene’s tests indicated that variances were not homogeneous at the third stage (p < .05). Attempts to stabilizing the variances between groups by performing log transformation of the data did not succeed. 75 from Time 1 to Time 2; students in Group 1 had the most decrease in their impressions of roommates’ task attractiveness upon moving in. H8(C) was supported for task attractiveness. Figure 10 Figure 11 Changes in Impressions of the Roommates’ Changes in Students’ Impressions of the Task Attractiveness at the Three Stages Roommates’ Task Attractiveness by Groups 5.4 5.2 5 4.8 4.6 10 5.19 4.99 4.9 4.79 5 5.67 4.44 5.2 0 Task Attractiveness Task Attractiveness Stage 1 Stage 2 Group 1 Group 3 Stage 3 Group 2 Group 4 Figure 12 Changes in Students’ Impressions of the Roommates’ Task Attractiveness at the Three Stages by Groups Task Attractiveness First Stage Second Stage Third Stage 5.55 5.81 5.64 5.15 4.61 4.61 Group 1 4.31 4.6 4.42 5.75 5.28 4.56 Group 2 Group 3 76 Group 4 Table 23 Summary of Descriptive Statistics for Impressions of Task Attractiveness Time Changes in Impression First Wave N Group 1 5.15 .98 32 4.31 .63 62 Group 3 5.55 .77 61 Group 4 4.56 .61 31 Total 4.91 .91 186 Group 1 4.61 .86 32 Group 2 4.60 .97 62 Group 3 5.81 .66 61 Group 4 5.75 .53 31 Total Third Wave SD Group 2 Second Wave M 5.19 .99 186 Group 1 4.61 1.12 32 Group 2 4.42 1.40 62 Group 3 5.64 .96 61 Group 4 5.28 1.34 31 Total 4.99 1.32 186 77 A summary of the results of hypothesis testing is shown in Table 24 below. Table 24 Hypotheses Testing and Results Hypothesis Results H1 (A): Information-seeking behaviors reduce uncertainty before Partially supported moving in. H1 (B): Facebook impressions of roommates’ social, appearance, Partially supported and task attractiveness reduce uncertainty before moving in. H2: There are group differences in information-seeking behaviors Partially supported and uncertainty before moving in. H3: Offline impressions of roommates’ social, appearance, and Partially supported task attractiveness reduce uncertainty upon moving in. H4: There are group differences in information seeking behaviors Partially supported and uncertainty upon moving in. H5: Discrepant Facebook and offline impressions have an impact Supported on uncertainty upon moving in. H6: Discrepant Facebook and offline impressions have an impact Partially supported on roommates’ relationship closeness. H7 (A): Students’ uncertainty changes over time. Supported H7 (B): There are group differences among students’ uncertainty. Supported H7 (C): There are interaction effect of Time x Groups on students’ Supported uncertainty. 78 Table 24 (cont’d) Hypothesis Results H8 (A): Impressions of roommates’ social, appearance, and task Supported for social and attractiveness decrease from Time 1 to Time 2. appearance attractiveness H8 (B): There are group differences among students’ impressions Supported of roommates’ social, appearance, and task attractiveness. H8 (C): There are interaction effect of Time x Groups on students’ impressions of roommates’ social, appearance, and task attractiveness. 79 Supported CHAPTER 5 INTERVIEW SUMMARY There were nineteen first year students participating in the phone interviews between November 9 and 22, 2010. The interviewees were recruited from the respondents who finished the three-wave survey. About fifty-eight percent of the participants were female, and 42% were male. Five of them were in Group 1—the unpleasant surprise group—whose impressions changed from positive to negative; one student was in Group 2, the impression stayed negative after move-in; another five students belonged to Group 3, whose impressions of their roommates remained positive after living together; and eight students were in Group 4—the pleasant surprise group—that their impressions of their roommates changed from negative to positive. These interview data provide supplementary supports for the study and offer insights into first year students’ Facebook use associated with uncertainty reduction and roommate relationship development. The following section will focus on the summary of the interview data. Before Moving in with the Roommates Among students being interviewed, only one out of 19 students lived with three roommates; the other 18 students (95%) shared the rooms with only one roommate. About 80% of them received the roommate assignments two to three weeks before moving in; 10% in July; and one student received his assignment a week before moving in due to some scheduling conflict with one of the resident halls. Only one student had known her roommate since childhood; 95% of the interviewees did not know their roommates at all at the time they received the assignments. All of them looked up their roommates’ Facebook profiles, and most of them did so right after receiving the emails from school. Fifty percent of the students initiated the contact on Facebook with their roommates; while two students mentioned that they either called or emailed their roommates first, and then added them as their Facebook friends a few days later. One student said that it took him a day or two to ―track down‖ his roommate, because there were multiple people using the same name. 80 Roommates’ Facebook Profiles When discussing their roommates’ Facebook profiles, many students emphasized the importance of the information, including pictures, wall posts, interests and activities, on their roommates’ profiles that helped them form a basic idea of who their roommates were. About 80% of the students mentioned that the pictures were very helpful. For example, a student said, ―My roommate seemed really nice. Her pictures on Facebook indicated that, especially those tagged by her friends. I think you can really see what kind of person she is from the pictures, and you can get the raw real impression from the pictures tagged by her friends.‖ For this student, the pictures provided an immediate and strong first impression. Besides, these pictures seemed to be genuine and thus offered a reliable access to the roommate’s personality. Another student also pointed out that, ―She [My roommate] seems really cool, nice and smart. I thought we can get along. I like what I saw. Her pictures, interests, and wall postings are important, and her tagged pictures give me more credible information.‖ More than half of the participants mentioned that they could also learn something about their roommates from their wall posts written by roommates themselves and their friends. One student said, ―I kind of looked on the wall. I just wanted to see who he is. It determines what kind of personality he has. For me the most important thing is what kind of person he is, so I just judged by what he wrote on the wall posts, and what his friends wrote about him and the updates. The first post I got was about the online game, World of Warcraft, and I never knew that before, and I thought, hey, it is going to be a cool experience to move in with someone that I don’t have much in common with.‖ For students who relied heavily on pictures and wall posts to form impressions about their future roommates, several of them did specifically point out that the tagged pictures and wall posts written by their roommates’ friends were more credible, as suggested by previous research (Walther and Park, 2002; Walther, et al., 2009) that this type of other-generated information did possess higher warranting value. 81 More than 50% of the students looked into their future roommates’ interests (e.g., interests in music and movies) and activities to see if they had anything in common with each other. For example, some students checked out their roommates’ music tastes (e.g., musicians, bands, and music genre) and some looked at their roommates’ activities (e.g., sports, video games, and social activities). A student said, ―his [My roommate’s] profile was pretty good. He had some interests that were similar to mine. His interests of music, movies, books, games, sports, and hobbies in his profile were important to me.‖ A female participant mentioned that, ―She seemed to be normal and that comforted me. She has a group of friends from high school from where I was. She has a lot of pictures, activities and school stuff. It made me feel that she did a lot of activities and we have something in common to talk about.‖ Roommates’ political and/or religious views were important to a few interviewees. A student said that she was concerned about her roommate’s religious view, ―I noticed [it] on the Facebook right away, and we talked about it. It did not make me uncomfortable. We just agreed to disagree.‖ Students with different sexual orientation tended to rely on roommates’ political/religious views as indicators to whether their future roommates would be open-minded to their choice. A student said, ―I went through his political and religious views, and was very relieved that he is someone I could get along. I am gay, and rooming with a random guy would be difficult, 50-50 chance it is find or not ok. Based on what people believe from their views, you can kind of get a general view of who he is. If he is conservative, it could be a red flag. Usually I don’t mind people’s responses, but I have to live with him, so it is important.‖ Another student made a similar comment that, ―I was mostly concerned with his religious and political views. I am gay and I was worried how that whole situation was going to be.‖ This student said that finding out that he and his future roommate were different politically from Facebook became his main worry before moving in; however, after getting into direct interaction through text messaging with his 82 future roommate, he was very happy to find out that they are a lot alike, although he still decided that ―I couldn’t tell him about me until I had met him in person.‖ There were a few participants indicating that they in fact didn’t rely too much on Facebook to get to know their roommates. A student explained that he did not look into his roommate’s profile that much, because ―I didn’t look into his profile that much because Facebook profiles aren’t always reflective of the people. You can’t really get to know how a person acts or behaves. I didn’t want to be biased before I met him. Pictures were not really important. I just wanted to wait until school started to actually get to know my roommate.‖ For these students, they got a basic idea of who their roommates were, and then decided to use a more interactive information-seeking strategy, such as using Facebook message, Facebook chat, emails, text messages, and even direct phone calls. A student said he learned a lot about his roommate by sending Facebook messages back and forth: ―I messaged him and asked him about himself. I got most of the information about him from that.‖ Another student mentioned that most specific information was hard to gather from Facebook because ―not everyone keeps everything completely up to date.‖ He and his roommate relied on text messaging to communicate. Unlike interviewees who appeared to think less about the information on Facebook, there were also students who perceived Facebook as an important source for getting to know their roommates. However, a few students mentioned that they wanted to learn more about their roommates, but there was nothing much on their roommates’ Facebook profiles. A student said, ―she didn’t have anything written in her biography [About Me section], so I couldn’t see what she says about herself.‖ Another student who thought at first that his roommate was a strange-looking kid said, ―From his photo album, he is a strange looking kid at first. He did not have detailed sections other than photos. I looked at the wall posts as well. He did not have a lot either. He seems to be social, but he did not have a lot of friends and wall posts.‖ 83 A student complained that his roommate’s profile was really shabby that, ―He did not even have a profile picture. I think he does not really want to show his face. He did list something about himself. Basically I can sum him up by looking at his profile. He would fall into a category of a geek. He did not have any pictures of outdoor activites, and he is very introverted.‖ When discussing whether there was anything on Facebook about their future roommates that concerned them, a student held a completely opposite view of Facebook reducing her uncertainty about the roommate: ―I think it made me more anxious, to be honest. I don’t want to say scared, but I was nervous being in such a close quarter when we don’t have similar interests. After reading her profile, I was slightly concerned that our interests weren’t the same. She came from a different background than I did. I was concerned that we won’t get alone as well. Her status updates were really negative. She said negative things about her life. I was really worried about her negative personality, and she has a complete opposite religious view.‖ As Planalp & Honeycutt (1985) indicated in their findings, the information that the student obtained from Facebook was so disturbing and unexpected that her uncertainty about the future roommate was increased. She became more nervous about living with her after checking her up on Facebook. In addition to different interests and negative status updates, another student was concerned about her roommate’s photos: ―At the time I was a little scared. She was a lot different. She likes to party and I don’t. She had photos about her drinking and party, and she posted crappy and graphic things about herself on her wall. I don’t think she cares about her self-presentation. But overall, Facebook helps reduce my uncertainty. I get to brace myself for what is going to be.‖ Two students discussed their concerns about their roommates’ language proficiency after they found out some information on Facebook about their future roommates was written in Korean or Chinese, but for both of them, it turned out that both of their roommates speak English pretty well. 84 Communication Channels for Roommate Interaction before Move-in All nineteen interviewees engaged in interactive information-seeking behaviors to communicate with their future roommates before move-in, although most of them preferred less ―intimate‖ (i.e., asynchronous) communication channels, like email, Facebook message, and text messaging. A student pointed out that ―phone call would be too intimate.‖ Students who called their roommates answered that they only used it once or twice; one student who did not use phone said, ―I mostly sent her Facebook messages and emails. I never used my cell phone. Because I am scared by what I saw on her profile, I would never give her my number.‖ Two other students believed that indirect interaction can avoid the awkwardness during the initial relationship development. They said, ―we never used phones. We haven’t met each other yet, so if I called her, it would be weird,‖ and ―phone call would be too intimate. Emails and Facebook messages would not be as awkward.‖ Interviewees who interacted more often with their roommates appeared to form more positive impressions about their roommates. In addition, the quality of their interaction appeared to be another important factor influencing their impressions and uncertainty. A student said that she was texting her roommate every other day, but mainly talking about what to bring, not many personal details. However, she also made this comment, ―At least I knew that we could carry a conversation although I didn’t really learn much about her through text messaging.‖ Thus, future research can also take the quality, or the valence, of interaction between students into consideration (Antheunis, et al., 2010; Ramirez, 2007; Ramirez & Burgoon, 2004; Tidwell & Walther, 2002) As for the discussion of the content of their interaction, one of the common themes was what things to bring to their dorm rooms. Some of them talked about their majors and classes, and some of them shared more personal details about themselves with their roommates, including who they are, how many brothers and sisters they have, what they like to do in general, their routines, and life goals. A student said, ―I feel comfortable talking to him [my roommate]. He is into music, into band, and he is a really smart kid. I know he is in the school orchestra, although I am not sure of his instrument.‖ 85 Upon Moving in: Impression Discrepancies Students were asked whether there were any surprises for them after they moved in with their roommates and whether there were things they wished to find out earlier before moving in. About 70% of the participants indicated that there was no surprise after meeting their roommates in person. However, a few students said that they were surprised to find out that their roommates were slightly different from what they pictured. A student said, ―I thought he was a little bit quieter from the messages he sent me, so I was a little surprised to find out that he was nice and outgoing. I would say you can’t exactly get the whole feel about what a person is on Facebook without actually meeting him.‖ Moreover, a student said that his roommate looked older and taller on Facebook; one commented that his roommate was more sociable in real life; and a female student pointed out her roommate is ―more outgoing than I thought from her profile.‖ But overall, these students agreed that these differences did not have too much of an effect on their initial impressions of their roommates. While many students admitted that Facebook was helpful for getting a basic idea about what their roommates would be like, there were more details about their roommates that Facebook cannot convey. Among things that a few students mentioned they wished they could have found out earlier were their roommates’ sleep time, tidiness, and aspects of their family background. However, in the end, most of them were able to communicate and work things out with their roommates, as illustrated in the following quote: ―We work out our sleep time. We are both open minded. I feel comfortable about talking to her about everything.‖ A student talking about his roommate’s tidiness said, ―I would say you can’t exactly get the whole feel about what a person is on Facebook without actually meeting him. He is a bit messy. I encourage him to take out the garbage. It does not work really. Pretty much I clean the room. Yes, he does appreciate my effort. Sometimes he eats up my 86 food that I save for later time. I told him and he will replace it as quickly as he can. Since we talk about it, we get over it and it does not bother me.‖ Another student said that he was worried about the way his roommate portrayed himself on Facebook, but after living together, he found out that his roommate was in fact very nice and considerate. He admitted that, ―I would not have been able to know this through Facebook.‖ A student mentioned that she was not aware of her roommate’s background, and wished that if she had known, she would be less judgmental about the roommate: ―There wasn’t any surprise for me when I met her in person. I was somewhat aware of her religious view. The fact that she is an atheist was very concerning me, and to my family as well. She has a cynical view on her profile. She dressed different from me, so I was very anxious about the fact that we are not going to get along. We have different life styles. She is not the most social person. I was sure how things are going to work out. The fact that she can’t communicate with me made me anxious about how we are going to get along. But, I kind of wish I would have known where she is from. She lived with her grandparents, and she is not very close to her parents. She never met her father, and she did not see her mother. I am very close to my parents. That was a little bit of a shock. If I had known, I would have welcomed her into my family as one of my friends. If I had known her background earlier, I would not be this judgmental, and I would have tried to accommodate her more, to make sure I made her more comfortable.‖ During the interview, the student showed her regret at not being able to be more considerate to her roommate. The prejudgment she made about the roommate based on her initial Facebook impression discouraged her from making further efforts to interact with her roommate. However, she said that after she found out her roommate’s background: ―I tried to make things easier. I made a lot of effort of being friendly and welcoming. She [My roommate] did try to be more social and more conversational. It was less awkward after a few weeks, and she said she was happy to converse with me.‖ 87 In addition, a student mentioned that he was very surprised at his roommate’s sexual orientation after move-in: ―One big thing is his sexual orientation. I found out he is homosexual, so it is interesting. I mean I don’t have any homosexual friend before. I kind of wish I had known it before but at the same time it is his profile. There is not much information on it. Initially, I was worried because I never knew anyone homosexual. But I just kind of got used to it. It does not bug me as much as earlier. I was worried about life style. I am from a conservative family. Where I am from, there is only one [person] who came out of the closet.‖ There were extreme cases that the negative discrepancies, or negative expectancy violations, were so grave in magnitude that two students made decisions to move to another dorm when they participated in the interviews. One student complained about her ex-roommate that, ―She looks more open on Facebook, but mean in reality. She said she would not be drinking and late [before moving in], but she did. I was very surprised that she did not keep her promise in the first week. She did not respect my feeling. She keeps repeating her problems so I moved to another dorm. I wish I would have found out how different we are. I know we are different, but not THAT different. Our relationship is seriously affected by it. We don’t talk at all and she was deleted from my friend list.‖ The other student expressed how unsatisfied he was with his roommate’s personality, being impolite and racist. He said, ―When I first met him, he made all these rules. He said that I can’t have any friends in our room past midnight and stuff, which is very surprising to me as a college student, not to interact with other people. He doesn’t mind if my friend is here, but then he will leave. He is not sociable. I don’t want to be rude, but he is a borderline racist. He and I are like polar-opposite. I just don’t think someone in my age would really have any problem with a different race or religion. I mean he does not like black people. He said that black people can’t say proper English. I could not believe what I heard when he said that. He thinks he can say whatever he wants to say. He is way 88 too self-confident. I have tried to talk to him with all these problems. And then he would get out of it and made me uncomfortable. I filled out the application and will move out soon. It is not what I expected, we are just so different. It is his personality that affects the relationship and kind of kills everything.‖ Developing a Close Relationship with the Roommates In addition to the findings showing that Facebook can influence students’ impressions of their future roommates, the interview data also suggest that Facebook use can impact how roommate relationships develop over time. Interviewees who did develop a closer relationship with their roommates reported more frequent usage of Facebook both to keep track of their roommates and interact with them. The survey data support this finding from the interviews, by showing that only some respondents continued to use Facebook extensively with their roommates, and this usage was positively associated with relationship closeness. There were 99.5% of participants logging onto Facebook to get information about their roommates in the first-wave survey, but in the third-wave survey the percentage dropped to 78.2%, and only 58.5% of respondents still interacted with their roommates on Facebook. During the interviews, interviewees were asked to rate their roommate relationship on 10-point scale, with 1 as the least close, and 10 as the closest. The mean value was 5.89 (SD = 2.65). Interview findings below are organized according to the level of closeness reported. Relationship Closeness = 1~2 The following three cases describe students who did not get along with their roommates, and no longer use Facebook to interact with or track roommates. One student said that since the semester started, the only activity that she and her roommate ever did was one dinner together. Other than that, they had no interaction at all. Their relationship had gone from bad to worse. By the time she had the interview with the researcher, she had already moved to another dorm. She mentioned that when she moved out, the roommate acted hostile to her and her family. However, she thought Facebook prepared her well, stating that ―Facebook prepared me to watch out, showing what to look out for.‖ 89 Another student who planned to move out said he barely saw or interacted with his roommate at all. They had a very different daily schedule, and even when they were both in the room, his roommate would rather play video games than having any interaction with him. He tried to talk to him several times, but all he got was sarcastic comments in return. Another student said he and his roommate never did activities together. All his roommate did was playing World of Warcraft all day long. But he made an interesting analogy between Facebook and a book cover, ―It is the whole thing about not judging a book by its cover. The thing with Facebook is that it is the cover and once you get to know him a little better, the whole perception can change. You usually do need to talk to the person to know what he is really like and judge by that.‖ Relationship Closeness = 4~5 Students in this category only had occasional interactions with their roommates, either on Facebook or face-to-face. Most of them would have lunch or dinner together, and once in a while they would check out their roommates’ updates on Facebook. One student thought that her roommate did not portray herself correctly on Facebook, so there was a huge difference between her Facebook image and real self. Another student said that Facebook was only useful to know who her roommate was in the beginning, but ―it was not as important in her roommate relationship.‖ Relationship Closeness = 6~7 Students had a much more and positive interactions with their roommates in this category. They did more activities together (e.g., studying, hanging out on weekends, partying, having meals, playing video games, attending social events, etc.). Many of them had common friend circles. They all agreed that they would stay friends even after they move out next year, and two said that they plan to continue to live together in their sophomore year. Facebook still played some parts in their interaction—they checked up each other’s status updates and interacted with each other on Facebook (e.g., Facebook chat and wall postings). Although Facebook was not their primary mode of communication now, it facilitated the development of their roommate relationship. As one student said, 90 ―Facebook is really helpful to begin the relationship with my roommate. It makes it less awkward to meet for the first time, and I get to know her better. It keeps us connected, and I sometimes check up her status updates to see if she has a crappy day or not.‖ Relationship Closeness = 8~10 Students in this category have developed a very close relationship with their roommates. They spent a lot more time together (e.g., some said that they are inseparable), and many of them had the same circle of friends. A few mentioned that they signed up to live together for another year. They used Facebook to interact with their roommates on a daily basis, like getting updates, writing on each other’s wall, and using Facebook chat (sometimes even when they were in the same room!). From the information that students provided during the interviews, one can see the tendency that those who had a more successful experience with roommates viewed Facebook as beneficial for developing a closer roommate relationship. Conclusion The goal for conducting the phone interviews for this study was to find more detailed information about how first year students used Facebook to form impressions of their future roommates, whether there were any discrepancies between initial Facebook and offline impressions, how they dealt with it, and how close their roommate relationships have become. The interview data suggested that while all students passively obtained roommates’ information on Facebook, they also engaged in interactive informationseeking behaviors to communicate with their roommates to reduce their uncertainty. Several of them acknowledged that Facebook has its limitation to get the whole picture of who their roommates are, but most of them appeared to agree that Facebook is beneficial in initiating their roommate relationship. Interviewees relied on pictures, wall posts, interests and activities to form impressions of their future roommate, which were consistent with the findings from the survey data. Moreover, a few students from the interviews emphasized the importance of their roommates’ political and religious views when they were trying to figure out what their roommates would be like in the beginning of the relationship development. Although Facebook played an influential role in the initial stage of students’ roommate 91 relationships for impression formation and uncertainty reduction, the interview data suggested that offline experience and impression discrepancies have a greater influence later on in their relationship. For students who developed a closer roommate relationship, Facebook became one of the communication channels they used to maintain their strong tie (Haythornthwaite, 2002; Haythornthwaite & Wellman, 1998). 92 CHAPTER 6 DISCUSSION AND CONCLUSION This study investigated first year college students’ Facebook use associated with their relationship development with their previously unacquainted roommates. The preliminary analysis of the data showed that looking up roommates’ profiles before moving in together is very common among first year students. This chapter first summarizes findings from the survey and the interviews by discussing each research hypothesis. Then the limitations and future research are discussed. Finally, the conclusion and the implications of the study are presented. Hypothesis 1 The first hypothesis aimed to test the impact of information-seeking behaviors and the initial Facebook impressions of the roommates on first year students’ uncertainty levels. As predicted by URT (Berger, 1979; Berger & Calabrese, 1975) and POV (Sunnafrank, 1986, 1990), the results indicated that increased interactions with roommates on Facebook, increased number of channels used to communicate with the roommates, impressions of roommates’ appearance and task attractiveness were able to reduce their uncertainty about the roommates before moving in. The frequency of information-seeking on Facebook and the impressions of roommates’ social attractiveness, however, did not show significant effects on reducing uncertainty. After careful examination of the survey item measuring the information-seeking frequency, it may be the time constraint and the question design that produced this result. The first-wave survey was distributed on August 16, only a few days after they received their roommate assignments from school. At the same time, they need to wait for their roommates to add or accept them as their Facebook friends before getting a full access to view each other’s profile. In addition, the 93 question asked about the frequency, not the amount of time, which also influenced the outcome. It is possible that students may simply go onto Facebook and go over their roommates’ profiles in depth, spending hours rather than repeatedly visiting his or her profile page. Therefore, the amount of time may be a more appropriate measure here. As Joinson (2008) suggested that Facebook users have different usage patterns to meet their needs—for users who want to fulfill their needs for content, which is the case for first year students who want to know more about their future roommates, they would spend more time on the site; for users who want to fulfill their need for social connections, such as reconnecting with old friends from high school, they would increase their frequency of Facebook use. There was another possibility indicated by a student who participated in the interview that because it was less than a month before moving in with his roommate, he believed the best way to reduce his uncertainty was using a more direct method, like text messaging, to interact with his roommate. He said, ―I think that trying to find specific information is very hard on Facebook, like specific interests and such. Mostly because not everyone keeps everything completely up to date. He [My roommate] and I mostly relied on text messaging to communicate. That way we could ask pointed questions privately. I sent him probably about 25 texts a week, about who is bringing what, majors, classes, interests, and even life goals. He and I are a lot alike.‖ What this student suggested has an implication for students’ information-seeking strategies before moving in with their roommates that it seems interactive strategies such as direct interaction on Facebook and through other communication channels are more effective in reducing their uncertainty (Tidwell & Walther, 2002) than passively learning more about their 94 roommates on Facebook. Antheunis and the colleagues (2010) had the same finding that the interactive uncertainty reduction strategy was the only strategy that effectively reduced the information seeker’s level of uncertainty on SNSs. And it can be the reason that the frequency of learning more about the roommate on Facebook did not appear to be a valid predictor in the analysis. H1 (B) was partially supported that initial Facebook impressions of roommates’ appearance and task attractiveness will have a positive influence on reducing students’ uncertainty, but not for social attractiveness. Why do students’ impressions of roommates’ social attractiveness not have an influence on reducing their uncertainty, while theories of URT (Berger, 1979; Berger & Calabrese, 1975) and POV (Sunnafrank, 1986, 1990) suggested that with a positive anticipated future interaction, social attraction is positively related to lower levels of uncertainty in initial interaction? A study (Antheunis, et al., 2010) found that on social network sites, a positive perception of the target person’s profile was a relatively less important determinant of uncertainty level than a negative perception (i.e., the negativity effect), which may be the reason for the finding in this study. The scores for students’ impressions of their roommates’ social attractiveness were relatively positive even between groups (overall M = 5.36 on a 7-point scale; M = 5.97 for positive impression group; M = 4.65 for negative impression group), and thus, as suggested by previous research, cannot effectively reduce their uncertainty level. Future research should therefore focus on the moderating effect of the valence of the impressions on uncertainty reduction. Hypothesis 2 The second hypothesis predicted that freshmen who had positive impressions of their roommates would score higher in (1) information-seeking frequency on Facebook, (2) frequency 95 of interaction on Facebook, (3) number of communication channels used, and (4) certainty. The results suggested that students with positive impressions of their roommates did use more communication channels to interact with their roommates, and they were more certain about what their roommates would be like. However, both information-seeking and interaction frequencies on Facebook did not differ between groups (positive vs. negative impressions). It can be the same reason discussed in H1 that students may spend hours going through their roommates’ profiles rather than visiting their profile pages repeatedly. Thus, the amount of time may be a more appropriate measure than the frequency of checking up roommates on Facebook. It can be the same reason for no significant group differences in frequency of Facebook interaction with roommates. Instead of asking how often they interacted on Facebook, the amount of time they spent interacting with the roommates on Facebook may be more suitable to measure their patterns of Facebook use. Future research should focus on whether students with positive initial Facebook impressions will differ from students with negative initial Facebook impressions in the amount of time they spend on checking up their roommates’ profiles as well as interacting with their roommates. Findings from the Interviews: Initial Impression Formation on Facebook Preliminary analysis of the survey data suggested that the top five types of the information that students relied on to form impressions of their roommates were pictures, wall posts, interests and activities, status updates, and background information. The findings through interview data indicated similar results. Roommates’ pictures, wall posts, and their activities and interests were the three most mentioned items during the interviews. Many interviewees believed that they were able to form relatively accurate impressions of their roommates from the pictures and wall posts; several even pointed out the pictures and wall posts either tagged or written by 96 other people were more credible source for impression formation on Facebook. In addition, interviewees looked through future roommates’ interests and activities in order to find whether they had anything in common. Common tastes in music, movies, and activities are important indicators for their perception in whether they would get along in the future. Several interviewees mentioned the importance of political and religious views that their future roommates hold, which was not shown from the survey data. They pointed out that through their future roommates’ liberal or conservative belief, they would be able to decide their roommates’ attitude toward some sensitive issues like sexual orientation. In addition to adopting passive information-seeking strategies to form roommates’ impressions, all interviewees used other communication channels for direct interactions with their future roommates before moving in. Most of them preferred asynchronous communication channels, like text messaging, emails, and Facebook message; while several of them talked to their future roommates through phone calls. Discussion of items to bring to the dorm rooms was the major topic in their interaction. Several students mentioned that they also exchanged more detailed personal information with their roommates, and they felt rather comfortable sharing the personal background with their future roommates. Most interviewees acknowledged the usefulness of Facebook in helping them reduce the uncertainty about their future roommates, although some students were also aware of the limitations of Facebook for not able to provide more in-depth pictures about their future roommates. Among nineteen students being interviewed, only one student had the opposite view of Facebook reducing her uncertainty. She mentioned that what she saw on Facebook about her roommate was in fact made her more anxious about living with a very different person in the future. 97 Hypothesis 3 The third hypothesis investigated the impact of offline impressions on uncertainty reduction in the beginning of the semester. The outcome revealed that after controlling for the influence of information-seeking frequency on Facebook (β = .15), frequency of interaction on Facebook, and number of channels used (β = .13), both the offline impressions of the roommates’ social attractiveness (β = .34) and appearance attractiveness (β = .18) had a positive impact on reducing uncertainty. However, the offline impressions of roommates’ task attractiveness did not lead to the same prediction. It can be due to the fact that students’ initial Facebook impressions of their roommates’ task attractiveness were pretty accurate (M = 4.96 from initial Facebook impression, M = 5.20 from FtF impression upon moving in), and thus produced less impact on reducing their uncertainty. It can also be the similar reason discussed in H1 (B) that negative information is considered to have significantly more influence on impressions and levels of uncertainty (Antheunis, et al., 2010; Kellermann, 1984; Walther, et al., 2009; Yoo, 2003, 2009). And because students’ offline impressions of roommates’ task attractiveness were relatively positive (M = 5.20 on a seven-point scale; M = 5.73 for positive offline impression group; M = 4.64 for negative offline impression group), thus, cannot effectively reduce their uncertainty level. As suggested in H1, more research on the effect of the impression valence is critical for better understanding of this finding. Hypothesis 4 The fourth hypothesis suggested that freshmen with positive offline impressions of their roommates would gather their roommates’ information on Facebook more frequently, interact with their roommates more often, use more communication channels, and have a greater certainty level about their roommates. The comparisons revealed that the frequency of obtaining 98 information about roommates from Facebook (overall M = 3.29, positive M = 3.47, negative M = 3.09) was the only variable that didn’t show group difference. There are two reasonable explanations for this finding. The first reason is that the amount of time spending on getting more information about the roommates will be more meaningful than the frequency of time, as Joinson (2008) indicated that getting information about the roommates is more content-based activities, so they would spend more time on Facebook, like previously discussed in this chapter. The second reason is that after moving in with the roommates, students get to observe their roommates on a daily basis through FtF interaction and other sources, such as the space they share together. Gosling, Gaddis, and Vazire (2008) indicated that people are able to form impressions of others on the basis of the environment they create and inhabit, and their findings suggested that impressions formed based on bedrooms were more accurate than impressions formed on office settings. Some students from the interviews indicated that they were not aware of their roommates’ messiness until they moved in together, and that was something they could not find out from Facebook. The findings of more interactions between roommates, more communication channels used to interact with the roommates, and greater level of certainty about the roommates for the students who formed positive offline impressions of their roommates appeared to be consistent with POV expectations (Sunnafrank, 1986, 1990). Upon meeting roommates FtF for the first time, students’ impressions of their roommates may or may not change, and as Ramirez (2007) suggested, the valence of impressions will influence individual’s expected relational outcome. Students who formed less positive offline impressions about their roommates may adjust their expectations about their future roommate interaction, and thus engage in less uncertainty reduction efforts (Ramirez, 2007). 99 Findings from the Interviews: Impression Discrepancies During the interviews, many students mentioned that their initial Facebook impressions of their roommates were fairly accurate. There were some slightly differences in their physical appearances (e.g., younger and taller in real life) and personalities (e.g., more outgoing or considerate), but they indicated that those differences were minor, and did not affect their initial impressions much. A few students admitted that they wished they had known their roommates’ sleep time, tidiness, background, and sexual orientation earlier before moving in, but these were the detailed personal information that one rarely revealed on Facebook. Fortunately, for most of the interviewees they were able to worked things out with their roommates—either having a positive interaction or they agree to disagree. There were two extreme cases where students decided to move to other dorms. Their relationship with the roommates became so intolerable that they believed moving out was the only solution to the problem. Both students were already aware of the possible outcomes with their roommates before moving in based on the impressions formed on Facebook, but they all mentioned that they never thought it would be worse than they already expected. Even so, they agreed that Facebook helped them get prepared for what could happen after moving in with the roommates. Hypothesis 5 & 6 Both H5 and H6 aimed to explore the extent of influence that discrepancies between initial Facebook and subsequent offline impressions have on students’ level of uncertainty about their roommates right after moving in as well as the relationship closeness with the roommates after living for a while. The results showed a support for H5 that where impression discrepancies are positive or there is no discrepancy but the impressions remain positive, students do show more certainty about their roommates; conversely, where impression discrepancies are negative 100 or there is no discrepancy but the impressions remain negative, students have lower certainty about their roommates. The analysis of the data also show partially support for H6 that students with positive discrepant impressions have developed the closest relationship with the roommates (RCI for Group 4 = 14.85), followed by students with no discrepant impressions but the impressions remain positive (RCI for Group 3 = 14.72), then students with no discrepant impressions but the impressions remain negative (RCI for Group 2 = 12.21), and finally, students with negative discrepant impressions have the lowest degree of closeness in their roommate relationship (RCI for Group 1 = 12.00). While the order of RCI scores followed the prediction for each group, RCI scores for Group 3 and Group 4 were not significantly different from each other, and neither were Group 1 and Group 2, which may due to the smaller cell size in Group 1 and Group 4. The implication of the findings provides more insights into the theories of uncertainty reduction as well as the theories of CMC by identifying the influence of impression discrepancies on reducing individual’s uncertainty and interpersonal relationship development. While research on URT and POV suggested that positive impressions have a positive association with reduced uncertainty (e.g., Clatterbuck, 1979; Douglas, 1990), it focused on the interpersonal interaction in FtF settings. The theories of CMC that deal with impression formation online—SIP and hyperpersonal theories—expanded URT and POV into online environments and suggested that, even without nonverbal cues, online users are able to form impressions of others based on the information available online to reduce uncertainty; however, the impressions they formed about others tend to be stereotyped or over-idealized (Antheunis, et al., 2010; Ramirez, 2007; Tidwell & Walther, 2002; Walther, 1996). Neither of these theories considered the impact of 101 discrepant impressions on uncertainty reduction and the development of interpersonal relationship. The current study then seeks to remedy their oversight by recognizing the impact of discrepant impressions. Indeed, there has been little research in this area. In Planalp and Honeycutt’s research (1985), they indicated that people’s uncertainty will increase when they are confronted with events that are not expected, that are not coherent with their existing knowledge, and to which the adjustment are required. The findings in this study not only support Planalp and Honeycutt’s position regarding how students’ uncertainty increase when they disappointedly find out that their roommates are not what they pictured based on initial Facebook impressions; the findings also advance their theoretical assumptions regarding how students’ uncertainty decrease when they surprisingly find out that their roommates are actually better than they expected. In addition, the study applies expectancy violations theory to predict the possible relational outcomes derived from different types of discrepancies in impressions. Therefore, future research dealing with mixed-mode (e.g., online to offline) relationship development, such as online dating, would be greatly benefited from the theoretical framework proposed by this study. Findings from the Interviews: Roommate Relationship Development The findings from the interviews provide further supports for the hypotheses of the impact of impression discrepancies on relationship development. Interviewees with negative impression discrepancies developed the least close roommate relationship, including the two students who either moved out or applied to move out to the other dorms. Interviewees with confirmed negative offline impressions about their roommates had very few interactions with their roommates. As one student pointed out, ―we take the same classes, but we don’t sit together. I was kind of hoping to room with someone that I can hang out with, but it really does not bother 102 me.‖ Students in Group 1 (Unpleasant Surprise) and Group 2 (Negative Impression Confirmed) appeared to agree that ―Facebook can help to know who the roommates are, but it is not as important in the roommate relationship,‖ as one student pointed out. Interviewees in Group 3 (Positive Impression Confirmed) or Group 4 (Pleasant Surprise) developed a much closer relationship with their roommates as suggested by the survey data. One student with a confirmed positive impression mentioned that she and her roommate are inseparable, ―We walk to class together and sit together, we go to lunch together, and then do homework together. On the days when we don’t have the same classes, we wait for each other till 6pm and go to dinner together. We hang out in the room studying. Sometimes we talk to each other through Facebook when we are in the same room.‖ A student with positive impression discrepancies rated her roommate relationship with a score ―seven‖ on a ten-point scale. She said, ―After I got to know her background, and after we set up the boundary, we get along pretty well. We have the same major, which I was not aware until the second week we started to have class, being in the same room. We go to those meetings together. We do meals together. Most likely we will remain friends after we move out next year.‖ Another student with a positive discrepant impression of his roommate also pointed out that he and his roommate get along pretty well. He said, ―We pretty much eat and go to parties together. We do our homework together in the afternoon, then play a little bit of games, and then eat dinner together.‖ An interesting observation made by the researcher during the interviews seemed to support the notion of media multiplexity that the ―stronger the friendship tie, the more people 103 communicate, the more kinds of information they exchange, and the more media they used‖ (Haythornthwaite & Wellman, 1998, p. 1112). Interviewees who developed closer roommate relationships appeared to use Facebook more to interact with their roommates, and they pointed out Facebook plays a beneficial role in their roommate relationship development. Hypothesis 7 Hypothesis 7 explored the impact of time (i.e., Time 1: before moving in, Time 2: upon moving in, and Time 3: living together for while) and group difference (i.e., four groups based on different types of impression discrepancies) on students’ certainty levels. The results suggested that certainty level did increase significantly from Time 1 to Time 2, and then from Time 2 to Time 3; and that certainty level was different across groups—students in Group 3 and Group 4 were more certain about their roommates than those in Group 1 and Group 2. The finding is consistent with the predictions of URT and POV. It provides support for a positive correlation between the certainty level and the time individuals have known one another (Clatterbuck, 1979). Students’ certainty levels in each group increased over time; however, students with negative impression discrepancies showed the least increase in certainty about their roommates, suggesting the process of impression formation on Facebook has produced unreliable initial impressions of the roommates which may due to roommates’ misrepresentation, students’ idealization of the roommates, or very limited information available on Facebook. While the significant interaction effect was present too (see Figure 3 and Table 20 in Chapter 4), the findings were contradictory to the prediction of expectancy violations theory. According to the theory, violations in expectation should increase uncertainty (Afifi & Burgoon, 2000). For students with either a positive or negative impression discrepancies, uncertainty levels should increase after meeting their future roommates upon moving in. Instead, the findings 104 showed that their levels of uncertainty actually decreased from Time 1 to Time 2—from Figure 3, one can see that Group 4 became more certain to a greater extent than other groups - hence uncertainty decreased rather than the expected increase. Similarly, Group 1's certainty also increased, but to a lower extent than the other groups. This is also inconsistent with expectancy violations theory. The results suggest that there may be other factors which have more prominent impacts on students’ uncertainty upon meeting FtF. It can be the valence of the offline impressions or perceived roommate attractiveness confounding with the impact of discrepant impressions on uncertainty, or whether the students’ discrepant impressions constitute an expectancy violation which lead to the change in uncertainty—the same question addressed by Ramirez and Wang (2008) in their recent study. It can also be a measurement problem, because the scales of CL7 and interpersonal attraction were correlated with each other, suggesting that when students were more certain about their roommates, they were more attracted to them. Future research should continue to explore the relationship between uncertainty reduction and discrepant online and offline impressions. Hypothesis 8 Hypothesis 8 examined the impact of time and groups on students’ impressions of roommates’ social, appearance, and task attractiveness. The findings suggested that students’ impressions of roommates’ social and appearance attractiveness did decrease significantly from Time 1 to Time 2, providing support for the predictions of SIP and hyperpersonal interaction theories that the receivers tend to have idealized perception of others while the senders will engage in optimized selective self-presentational behaviors (Walther, 1996), which can easily result in forming more positive online impressions of others. However, the change in students’ impressions of roommates’ task attraction is the opposite of what SIP and hyperpersonal 105 interaction theory would predict—instead of decreasing from Time 1 to Time 2, it increased significantly. Students formed even more positive impressions of roommates’ task attractiveness right after moving in (M = 5.19). One explanation for this is that the initial Facebook impressions of roommates’ task attractiveness were reinforced by their actual performance in real life. While freshmen formed relatively positive impressions of their roommates’ task attractiveness based on Facebook (M = 4.90, above scale’s mid-point), they were able to get a first-hand observation of how competence their roommates were (i.e., having confidence in roommates’ ability to get the job done and depending on him/her to get things done) after living together, especially during the most hectic time when they were facing new challenges, such as moving into a new place and adapting to a new academic environment. The finding also suggested that information contributing to task attractiveness on Facebook may have a greater warranting value (Walther & Parks, 2002; Walther, et al., 2009). For instance, a student’s list of professional activities or participation in a particular organization on Facebook often has a straightforward indication of her actual ability. As for the main effect for groups, the findings indicated that there were significant group differences for students’ impressions of roommates’ social, appearance, and task attractiveness. The analysis led to the same outcomes that students in Group 3 and Group 4 gave higher scores on average than students in Group 1 and Group 2. However, the analysis also showed that there was no significant difference between Group 1 (negative expectancy violations) and Group 4 (positive expectancy violations) for roommates’ social, physical, and task attractiveness. It may seem contradictory at first. But after taking a closer look at the data (see Table 21, 22, and 23 in Chapter 4), one can realize that it was because students in Group 1 gave relatively high scores to their roommates for their online social, appearance, and task attractiveness, and this high score 106 increased the overall mean scores across the three stages. The same rationale is also applicable to the result for the students in Group 4. The low scores at the first stage brought down the overall mean scores for Group 4. Another contributing factor can be the relative smaller cell size in Group 1 and Group 4 as mentioned in previous section in this chapter. The analysis of the interaction effects of Time x Groups on students’ impressions of roommates’ social, appearance, and task attractiveness showed significant results (see Figure 6, 9, and 12 in Chapter 4). The changes in impressions of roommates’ social and appearance attractiveness from Time 1 to Time 2 were noticeable for students in Group 1 (significant decrease) and Group 4 (significant increase); however, students in Group 1 showed a greater influence from discrepant impressions than students in Group 4, suggesting the possible presence of negativity effect (Kellermann, 1984). Research (e.g., Gibbs, et al., 2006; Walther, et al., 2009; Yoo, 2003) showed that negative information is weighted more heavily than positive information in impression formation. Thus, students who find out their roommates are not what they previously expected (after meeting them FtF for the first time upon moving in) are forced to deal with the conflicting discrepancies which result in the greatest drop in their impressions of roommates’ attractiveness. Limitations There are some limitations for this study. First of all, the research instrument is a threewave self-administered questionnaire. A large proportion of the measurements, such as attributional confidence and interpersonal attraction scales, were reappearing in each of the three questionnaires. It may either desensitize the respondents or allow respondents to figure out what the purpose of this study is, and resulted in biased responses. The researcher was aware of this problem and took precautions while arranging the order of the survey items, but biased responses 107 were still likely present in the data. Second, the present study focused on first year college students who were going to live with previously unacquainted roommates. The generalizability of the results may be problematic. Thirdly, although the researcher started with inviting more than 2,700 first year students to participate in this project, there were only about 200 qualified respondents for the final analysis. The statistic power may not enough for some statistical analysis. Also, the number of respondents in Group 1 (i.e., negative increase in impressions) and Group 4 (i.e., positive increase in impressions) were less than 50, and it may affect the outcome of the analysis. And, the decision to perform median split for positive and negative impression groups was arbitrary, and it is inevitable that some subjects who did have positive impressions of their roommates were put in the negative impression group. The problem can be fixed in the future by adding measures in the survey for respondents to self-rate their impression of their roommates. Lastly, the relationship between discrepant impressions and uncertainty reduction was not consistent with the prediction of expectancy violations theory. The uncertainty level for students with positive and negative impression discrepancies was significantly decreased, although with different magnitude, after they moved in with the roommates. The findings suggest that students’ levels of uncertainty about their roommates may not be solely a function of violations in expectation. The predicted outcome value and hyperpersonal effect may offer some explanation for these results. Students may be more aware that their roommates were strategically managing their self-presentation on Facebook, and therefore, upon meeting them FtF, they are less surprised by discrepancies and are able to make use of new information to reduce uncertainty. The fact that impressions formed on Facebook did not provide a complete picture of their roommates once they met, in this case would not necessarily increase uncertainty. 108 It is also possible that the problems with measuring uncertainty led to this outcome. Knobloch and Solomon (2002) suggested that it may not be able to address the full range of questions salient within close relationships. Instead, they developed a three dimensional relationship uncertainty to measure the level of confidence people have in their perceptions of involvement in an interpersonal relationship. By measuring self uncertainty, partner uncertainty, 6 and relationship uncertainty , the scale may be more appropriate with better face validity to capture the types of uncertainty that first year students have towards their future roommates. Future Research The study examined the impact of first year college students’ initial Facebook impressions, subsequent offline impressions upon moving in, and the discrepancies between Facebook and offline impressions of their previously unacquainted roommates on their uncertainty reduction. The findings showed partially support for the predictions of the theoretical framework proposed by URT, POV, SIP, and hyperpersonal interaction theory. Students’ initial Facebook impressions of roommates’ social attractiveness (H1A) and their subsequent offline 6 According to Knobloch, Satterlee, and DiDomenico (2010), self uncertainty measures individuals’ confidence in their own involvement in a relationship (e.g., how certain am I about how important this relationship is to me?); partner uncertainty measures the perceptions that individuals experience about their partner’s involvement in a relationship (e.g., how certain am I about how important this relationship is to my partner?); and relationship uncertainty measures the ambiguity individuals experience about the dyad as a whole (e.g., how certain am I about the current status of this relationship?). The researchers suggested that relational uncertainty shares strong ties with people’s communication behavior, which can be thus applicable for this study. 109 impressions of roommates’ task attractiveness (H3) could not effectively reduce their uncertainty level. It can be the result of relatively positive ratings for roommates’ Facebook social attractiveness and their offline task attractiveness as discussed in previous sections. Future research, therefore, can focus on the moderating effect of the impression valence on uncertainty reduction by experimental design of Facebook profiles that generate positive and negative impressions. The investigation of information-seeking strategies that students use to reduce their uncertainty about their roommates in this study indicated that the amount of time they spend on Facebook to gather roommates’ information and interact with them would be a more appropriate measure than the frequency of time due to their usage patterns are more content-based. In addition, the quality of their interaction is as important as the amount of their interaction which would affect their uncertainty level (e.g., Antheunis, et al., 2010; Clatterbuck, 1979). Future research can also explore the role of the quality, or the valence, of the information on Facebook and information exchanged plays during the interaction in reducing students’ uncertainty about their roommates. In addition, the present research investigated the impact of Facebook on the development of first year students’ roommate relationship. It focused on the relationship between impression discrepancies, uncertainty reduction, and relationship closeness. Other factors which may affect roommate relationship after moving in together were not discussed in this research. Future research with different theoretical perspectives should be able to add more insights into this area of study. Finally, the causal relationship between the formation of initial Facebook impressions and level of uncertainty before moving in with the roommates is difficult to justify, since they 110 can happen and affect each other at the same time. A more thorough research design in the future can solve this problem by measuring students’ levels of uncertainty twice, one before freshmen looking up their roommates on Facebook, and then one after that. Conclusion The goal of this research was to examine (1) how college freshmen formed an initial impressions of their previously unacquainted roommates based on the information available on Facebook; (2) how their Facebook impressions and their information-seeking behaviors influenced their level of uncertainty about who the roommates were and what it would be like to live with him/her; (3) whether there were discrepancies between initial Facebook impressions and subsequent offline impressions; and (4) to what extent the impression discrepancies affected students’ uncertainty and the development of their relationship closeness with their roommates. The findings did support the idea of the role Facebook plays in facilitating the process of impression formation and reducing students’ uncertainty about living with new roommates. To prevent creating false impressions for new friends while promoting one’s online self-image in front of existing network, it is imperative for Facebook users to strategically engage in impression management. This study on Facebook, therefore, highlights the benefit of wellmanaged impressions as well as the potential disadvantage of misrepresentation that comes to affect individual’s relationship development with new friends. Given the fact that in 2007 there were already some reports regarding first year college students’ use of Facebook to check up their future roommates (e.g., Collura, 2007; Eberhardt, 2007; Walsh-Sarnecki, 2007), this research is probably one of the earliest attempts to provide a theoretical foundation for explaining the impact of Facebook use on roommate relationship for college freshmen. It has become a rather common phenomenon for college freshmen to ―friend‖ 111 their roommates on Facebook after receiving the roommate assignment from school. Almost all students participated in this research project checked up their roommates on Facebook before moving in with them. Indeed, Facebook is changing the way students seek out information about their future roommates. The findings provide support that Facebook is beneficial in reducing their uncertainty about their future roommates. It also provides a new venue for students to get a better idea of who their roommates would be before initiating interpersonal communication. Moreover, it helps reduce the uneasiness that is often experienced by college freshmen before getting in touch with a complete stranger. With the current findings, students would therefore be more equipped when it comes to managing a positive self-image, developing a healthy interpersonal relationship with roommates, and making the transition to college more smoothly. This investigation provided further support for the notion that impression discrepancies can affect individual’s uncertainty reduction process and the development of close interpersonal relationship, which can be a useful addition to the existing CMC and uncertainty reduction literature. First year college students’ use of Facebook to initiate the contact with their future roommates has become a unique scenario for SNSs. Unlike the situation where they make new friends or keep contact with existing friends on Facebook, it is the combination of both—starting online as a complete stranger and then developing an intense relationship with each other after the semester starts. The study finds that students with negative discrepancies between initial Facebook impressions and subsequent offline impressions of their roommates will be less certain about their roommates and consequently develop the least close relationship with their roommates; on the contrary, students with positive impression discrepancies will be more certain about their roommates and develop the closest roommate relationship. The findings suggest that impression discrepancies, which extend the theories of uncertainty reduction and impression 112 formation on CMC, can provide a useful theoretical foundation for future research on the development of mixed-mode (e.g., from online to offline) relationship. 113 APPENDICES 114 APPENDIX A First-Wave Survey Question 1: Which residential hall are you going to live in?  Armstrong Hall  Bailey Hall  Bryan Hall  Butterfield Hall  Emmons Hall  Rather Hall  Akers Hall  Holmes Hall  Hubbard Hall  Abbot Hall  Mason Hall  Phillips Hall  Snyder Hall  McDonel Hall  Owen Graduate Hall  Van Hoosen Apartments  Shaw Hall  Case Hall  Holden Hall  Wilson Hall  Wonders Hall  Campbell Hall  Gilchrist Hall  Landon Hall  Mayo Hall  Williams Hall  Yakeley Hall  Other Question 2: How many roommates do you have? [Mandatory]  one [Skip to Q7]  two [Skip to Q5]  three [Skip to Q5]  more than three [Skip to Q5] Question 3: Did you choose your roommate, or was he/she randomly chosen for you? [Mandatory]  I chose my roommate. [Skip to Q7]  My roommate was randomly chosen for me. [Skip to Q7] 115 Question 4: If you chose your roommate, please answer the following question. How long have you known your roommate? Please indicate the number of years and/or months (for example, 3 years, 8 months). years _____________ months ____________ If you have more than one roommate, please choose the roommate who will share the bunk bed with you in order to answer the rest of this questionnaire. If you are not sure yet, please pick one that you feel comfortable with to complete the rest of this questionnaire. Question 5: Did you choose your roommate, or was he/she randomly chosen for you? [Mandatory]  I chose my roommate.  My roommate was randomly chosen for me. Question 6: If you chose your roommate, please answer the following question. How long have you known your roommate? Please indicate the number of years and/or months (for example, 3 years, 8 months). years _____________ months ____________ Questions 7: How well do you know him/her?  Complete stranger  Nodding acquaintance  Acquaintance  Close acquaintance  Friend  Close Friend  Other, please specify _________________ Question 8: Do you have a Facebook profile? [Mandatory]  Yes  No [Skip to Q26]  If not, can you tell us why? _________________ This section will ask you about your general Facebook use. Question 9: On average, how often do you spend time on Facebook?  Several times per day  Once a day  Several times a week  Once a week  Several times a month  Monthly or less  Other, please specify _________________ 116 Question 10: How much time do you typically spend on Facebook each day during a typical week day (Monday to Friday)?          less than 1 hour 1 ~ 2 hours 2 ~ 3 hours 3 ~ 4 hours 4 ~ 5 hours 5 ~ 6 hours 6 ~ 7 hours 7 ~ 8 hours more than 8 hours Question 11: How much time do you typically spend on Facebook each day during a typical weekend day?          less than 1 hour 1 ~ 2 hours 2 ~ 3 hours 3 ~ 4 hours 4 ~ 5 hours 5 ~ 6 hours 6 ~ 7 hours 7 ~ 8 hours more than 8 hours Question 12: How many friends do you have on Facebook? To find out, please log on to your Facebook profile. _______________ Question 13: Please indicate the extent to which you agree with the following statements by choosing whether you (7) strongly agree; (6) agree; (5) somewhat agree; (4) neither agree nor disagree; (3) somewhat disagree; (2) disagree; (1) strongly disagree. Facebook is part of my everyday activities. 1 2 3 4 5 6 7 I am proud to tell people I'm on Facebook. 1 2 3 4 5 6 7 Facebook has become part of my daily routine. 1 2 3 4 5 6 7 I feel out of touch when I haven't logged onto Facebook for a while. 1 2 3 4 5 6 7 I feel I am part of the Facebook community. 1 2 3 4 5 6 7 I would be sorry if Facebook shut down. 1 2 3 4 5 6 7 117 Question 14: Does your roommate have a Facebook profile? [Mandatory]  Yes  No [Skip to Q26]  I don't know. [Skip to Q26] Question 15: Have you looked up your roommate's Facebook profile? [Mandatory]  Yes  No [Skip to Q26]  If not, can you tell us why? _______________ Question 16: Did your parent(s) also look up your roommate's Facebook profile?  Yes  No Question 17: When was the first time you looked up your roommate's Facebook profile? (e.g., if it's on Aug. 6, 2010, please put 8 in Month, 6 in Day, and 2010 in Year.) Month _________ Day ___________ Year ___________ Question 18: Did you add or accept him/her as your Facebook friend? [Mandatory]  Yes  No Question 19: On average, how often do you spend time on Facebook learning more about your roommate?  several times per day  once a day  several times a week  once a week  several times a month  monthly or less  never Question 20: On average, how often do you interact/communicate with your roommate via Facebook?  several times per day  once a day  several times a week  once a week  several times a month  monthly or less  never 118 Question 21: Below is a list of the information you may find about your roommate on Facebook. How critical are these items for you to determine what your roommate would be like? Please indicate the extent to which you think the following items by choosing whether they are (7) extremely important; (6) very important; (5) moderately important; (4) neutral; (3) slightly important; (2) low importance; (1) not at all important. His/her status updates in News Feed (i.e., the first page you see after you log in) His/her pictures (e.g., profile photos, albums, etc.) His/her Wall postings and comments The number of his/her Facebook friends The number of your friends in common His/her favorite movies/music/books/TV shows/quotes, etc. His/her political views 1 2 3 4 5 6 7 n/a 1 1 1 1 1 1 2 2 2 2 2 2 3 3 3 3 3 3 4 4 4 4 4 4 5 5 5 5 5 5 6 6 6 6 6 6 7 7 7 7 7 7 n/a n/a n/a n/a n/a n/a Question 22: How critical are these items for you to determine what your roommate would be like? Please indicate the extent to which you think the following items by choosing whether they are (7) extremely important; (6) very important; (5) moderately important; (4) neutral; (3) slightly important; (2) low importance; (1) not at all important. His/her religious views His/her relationship status His/her website link His/her background information (e.g., hometown, high school, etc.) The Notes he/she wrote in the profile Groups he/she joined Pages he/she bookmarked 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 n/a n/a n/a n/a 1 2 3 4 5 6 7 n/a 1 2 3 4 5 6 7 n/a 1 2 3 4 5 6 7 n/a Question 23: What other critical information not mentioned above is important for you to determine what your roommate would be like? (please specify) __________________________ Now, we would like to know what you think about your roommate after you’ve learned something about him/her on Facebook. Please indicate the extent to which you agree with the following statements by choosing whether you (7) strongly agree; (6) agree; (5) somewhat agree; (4) neither agree nor disagree; (3) somewhat disagree; (2) disagree; (1) strongly disagree. Question 24: Overall, I liked what I saw about my roommate on Facebook. Question 25: After looking up my roommate on Facebook, I would prefer another roommate. 119 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Question 26: Besides Facebook, what communication channels have you used to interact/communicate with your roommate? (Check all that apply)  emails  landline phone  cell phone  text messaging  instant messaging like AIM or MSN  Internet phone like Skype  Other, please specify ____________________ Question 27: How certain are you about your roommate at this moment? Please indicate the extent to which you agree with the following statements by choosing whether you (7) strongly agree; (6) agree; (5) somewhat agree; (4) neither agree nor disagree; (3) somewhat disagree; (2) disagree; (1) strongly disagree. I am confident of my general ability to predict how my roommate will behave. 1 2 3 4 5 6 7 I am certain that he/she likes me. 1 2 3 4 5 6 7 I am accurate at predicting the values he/she holds. 1 2 3 4 5 6 7 I am accurate at predicting his/her attitudes. 1 2 3 4 5 6 7 I can predict his/her feelings and emotions. 1 2 3 4 5 6 7 I can empathize with (share) the way my roommate feels about himself/herself. 1 2 3 4 5 6 7 I know him/her very well. 1 2 3 4 5 6 7 Question 28: Now, we would like to ask what you think about your roommate at this moment. Please indicate the extent to which you agree with the following statements by choosing whether you (7) strongly agree; (6) agree; (5) somewhat agree; (4) neither agree nor disagree; (3) somewhat disagree; (2) disagree; (1) strongly disagree. I think he/she could be a friend of mine. 1 2 3 4 5 6 7 n/a It would be difficult to meet and talk with him/her. 1 2 3 4 5 6 7 n/a He/She just wouldn’t fit into my circle of friends. 1 2 3 4 5 6 7 n/a We could never establish a personal friendship with each other. 1 2 3 4 5 6 7 n/a I would like to have a friendly chat with him/her. 1 2 3 4 5 6 7 n/a I think he/she is quite handsome/pretty. 1 2 3 4 5 6 7 n/a 120 He/She is very sexy looking. 1 2 3 4 5 6 7 n/a I find him/her very attractive physically. 1 2 3 4 5 6 7 n/a I don’t like the way he/she looks. 1 2 3 4 5 6 7 n/a He/She is somewhat ugly. 1 2 3 4 5 6 7 n/a He/She is a typical goof-off when assigned a job to do. 1 2 3 4 5 6 7 n/a I have confident in his/her ability to get the job done. 1 2 3 4 5 6 7 n/a If I wanted to get things done, I could probably depend on him/her. 1 2 3 4 5 6 7 n/a I couldn’t get anything accomplished with him/her. 1 2 3 4 5 6 7 n/a He/She would be a poor problem solver. 1 2 3 4 5 6 7 n/a Your demographic information Question 29: Your sex  Male  Female Question 30: Your age ________ Question 31: Your major (Please specify your department, otherwise choose "undecided")  undecided  Please specify your major _____________ Question 32: Family's Total Household Income  Under $20,000  $20,000 to $34,999  $35,000 to $49,000  $50,000 to $74,999  $75,000 or more  I don’t want to disclose Question 33: Ethnicity (Choose all that apply)  Caucasian/White  African American  Native American  Asian  Pacific Islander 121     Hispanic/Latino Multiracial I don’t want to disclose Other, please specify ______________ 122 APPENDIX B Second-Wave Survey Hi, Welcome back! Thank you for helping me with this research project. Your participation is very important to me. This is the second questionnaire which focuses on how you feel about your roommate now that you just moved in with him/her. It will take about 10-15 minutes to complete this survey. Question 1: Since last time you took the survey, have you requested any change in your initial roommate assignment? [Mandatory]  Yes  No [Skip to Q12] Question 2: To what extent did you think the information you find on Facebook about your roommate contribute to your decision to request such change? [Mandatory] 1 2 3 4 5 6 7 not at all A great extent Question 3: Would you like to leave your email address here so the researcher can contact you for questions regarding your decision to change roommate? Question 4: Did you choose your current roommate, or was he/she randomly chosen for you?  I chose my roommate.  My roommate was randomly chosen for me. Question 5: If you chose your current roommate, please answer this question. How long have you known your roommate? Please indicate the number of years and/or months (for example, 3 years, 8 months). years _____________ months ___________ Question 6: How well do you know your current roommate?  Complete stranger  Nodding acquaintance  Acquaintance  Close acquaintance  Friend  Close Friend  Other, please specify ______________ Question 7: Do you still have a Facebook profile? [Mandatory]  Yes  No [Skip to Q24] 123 Question 8: Does your current roommate have a Facebook profile? [Mandatory]  Yes  No  I don't know. Question 9: Have you looked up your current roommate's Facebook profile?  Yes  No Question 10: If you have looked up your current roommate's Facebook profile, when was the first time? (e.g., if it's on Aug. 26, 2010, please enter 8 in Month, 26 in Day, and 2010 in Year.) Month _________ Day ___________ Year __________ Question 11: Did you add or accept him/her as your Facebook friend? [Mandatory]  Yes [Skip to Q15]  No [Skip to Q19] Question 12: Do you still have a Facebook profile? [Mandatory]  Yes  No [Skip to Q24] If you have more than ONE roommate, please choose the SAME roommate you picked last time to complete this questionnaire. Question 13: Does you roommate still have a Facebook profile?  Yes  No  I am not sure. Question 14: Is he/she your Facebook friend? [Mandatory]  Yes  No [Skip to Q19] Question 15: On average, how often did you spend time on Facebook learning more about your roommate in the past month?  several times per day  once a day  several times a week  once a week  several times a month  monthly or less  never 124 Question 16: On average, how often did you interact/communicate with your roommate via Facebook in the past month?  several times per day  once a day  several times a week  once a week  several times a month  monthly or less  never Question 17: In the last questionnaire you probably rated the importance of the information on Facebook for you to determine what your roommate would be like. Now after move-in with him/her, of the following information, which one do you think is accurate about him/her in real life? Please indicate the extent to which you agree with the following items by choosing (7) very accurate; (6) accurate; (5) somewhat accurate; (4) Neutral; (3) somewhat inaccurate; (2) inaccurate; (1) not at all accurate. His/her status updates in News Feed (i.e., the first page you see after you log in) His/her pictures (e.g., profile photos, albums, etc.) His/her Wall postings and comments The number of his/her Facebook friends The number of your friends in common His/her favorite movies/music/books/TV shows/quotes, etc. His/her political views 1 2 3 4 5 6 7 1 1 1 1 1 1 2 2 2 2 2 2 3 3 3 3 3 3 4 4 4 4 4 4 5 5 5 5 5 5 6 6 6 6 6 6 7 7 7 7 7 7 Question 18: Of the following information, which one do you think is accurate about him/her in real life? Please indicate the extent to which you agree with the following items by choosing (7) very accurate; (6) accurate; (5) somewhat accurate; (4) Neutral; (3) somewhat inaccurate; (2) inaccurate; (1) not at all accurate. His/her religious views His/her relationship status His/her website link His/her background information (e.g., hometown, high school, etc.) The Notes he/she wrote in the profile Groups he/she joined Pages he/she bookmarked 125 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Your general Facebook use Question 19: On average, how often do you spend time on Facebook?  Several times per day  Once a day  Several times a week  Once a week  Several times a month  Monthly or less  Other, please specify _______________ Question 20: How much time do you typically spend on Facebook each day during a typical week day (Monday to Friday)?  less than 1 hour  1 ~ 2 hours  2 ~ 3 hours  3 ~ 4 hours  4 ~ 5 hours  5 ~ 6 hours  6 ~ 7 hours  7 ~ 8 hours  more than 8 hours Question 21: How much time do you typically spend on Facebook each day during a typical weekend day?  less than 1 hour  1 ~ 2 hours  2 ~ 3 hours  3 ~ 4 hours  4 ~ 5 hours  5 ~ 6 hours  6 ~ 7 hours  7 ~ 8 hours  more than 8 hours Question 22: How many friends do you have on Facebook now? To find out, please log on to your Facebook profile. ___________ Question 23: Please indicate the extent to which you agree with the following statements by choosing whether you (7) strongly agree; (6) agree; (5) somewhat agree; (4) neither agree nor disagree; (3) somewhat disagree; (2) disagree; (1) strongly disagree. Facebook is part of my everyday activities. 1 2 3 4 5 6 7 I am proud to tell people I'm on Facebook. 1 2 3 4 5 6 7 126 Facebook has become part of my daily routine. 1 2 3 4 5 6 7 I feel out of touch when I haven't logged onto Facebook for a while. 1 2 3 4 5 6 7 I feel I am part of the Facebook community. 1 2 3 4 5 6 7 I would be sorry if Facebook shut down. 1 2 3 4 5 6 7 Question 24: Did you meet your roommate face-to-face before moving into the dorm?  Yes  No Question 25: Besides Facebook, what other communication channels have you used to interact/communicate with your roommate in the past month? (Check all that apply)  face-to-face communication  emails  landline phone  cell phone  text messaging  instant messaging like AIM or MSN  Internet phone like Skype  Other, please specify ___________________ Question 26: In this section we would like to know how certain you are about your roommate now after move-in with him/her. Please indicate the extent to which you agree with the following statements by choosing whether you (7) strongly agree; (6) agree; (5) somewhat agree; (4) neither agree nor disagree; (3) somewhat disagree; (2) disagree; (1) strongly disagree. I am confident of my general ability to predict how my roommate will behave. 1 2 3 4 5 6 7 I am certain that he/she likes me. 1 2 3 4 5 6 7 I am accurate at predicting the values he/she holds. 1 2 3 4 5 6 7 I am accurate at predicting his/her attitudes. 1 2 3 4 5 6 7 I can predict his/her feelings and emotions. 1 2 3 4 5 6 7 I can empathize with (share) the way my roommate feels about himself/herself. 1 2 3 4 5 6 7 I know him/her very well. 127 Question 27: In this section, we would like to know what you think about your roommate now after move-in with him/her. Please indicate the extent to which you agree with the following statements by choosing whether you (7) strongly agree; (6) agree; (5) somewhat agree; (4) neither agree nor disagree; (3) somewhat disagree; (2) disagree; (1) strongly disagree. I think he/she could be a friend of mine. 1 2 3 4 5 6 7 It would be difficult to meet and talk with him/her. 1 2 3 4 5 6 7 He/She just wouldn’t fit into my circle of friends. 1 2 3 4 5 6 7 We could never establish a personal friendship with each other. 1 2 3 4 5 6 7 I would like to have a friendly chat with him/her. 1 2 3 4 5 6 7 I think he/she is quite handsome/pretty. 1 2 3 4 5 6 7 He/She is very sexy looking. 1 2 3 4 5 6 7 Question 28: What do you think about your roommate now after move-in with him/her? Please indicate the extent to which you agree with the following statements by choosing whether you (7) strongly agree; (6) agree; (5) somewhat agree; (4) neither agree nor disagree; (3) somewhat disagree; (2) disagree; (1) strongly disagree. I find him/her very attractive physically. 1 2 3 4 5 6 7 I don’t like the way he/she looks. 1 2 3 4 5 6 7 He/She is somewhat ugly. 1 2 3 4 5 6 7 He/She is a typical goof-off when assigned a job to do. 1 2 3 4 5 6 7 I have confident in his/her ability to get the job done. 1 2 3 4 5 6 7 If I wanted to get things done, I could probably depend on him/her. 1 2 3 4 5 6 7 I couldn’t get anything accomplished with him/her. 1 2 3 4 5 6 7 He/She would be a poor problem solver. 1 2 3 4 5 6 7 128 APPENDIX C Third-Wave Survey Thank you so much again for your continuous participation. This will be the last questionnaire for this research project. In this survey, we would like to learn more about the relationship between you and your roommate. If you have more than one roommate, we would like you to answer the following questions with regard to the particular roommate you chose in the earlier surveys you participated in. Question 1: Do you still get news/updates about your roommate from Facebook?  Yes  No Question 2: Do you still interact with each other on Facebook (e.g., leave messages or comments on the Wall)?  Yes  No Now, we would like you to estimate the amount of time you typically spend alone with your roommate during the day. We would like to make these time estimates by breaking the day into morning, afternoon, and evening, although you should interpret each of these time periods in terms of your own typical daily schedule. Think back over the past week and write in the average amount of time, per day, that you spent alone with him/her, with no one else around, during each time period, write ___ hour(s) ___minutes. Question 3: During the past week, what is the average amount of time, per day, that you spent alone with your roommate in the morning (e.g., between the time you wake up and 12 noon). hours _________ minutes _______ Question 4: During the past week, what is the average amount of time, per day, that you spent alone with your roommate in the afternoon (e.g., between 12 noon and 6 pm)? hours _________ minutes _______ Question 5: During the past week, what is the average amount of time, per day, that you spent alone with your roommate in the evening (e.g., between 6 pm and bedtime)? hours _________ minutes _______ 129 Question 6: Compared with the "normal" amount of time you usually spend alone with your roommate, how typical was the past week? (Check one)  typical  not typical Question 7: If not typical, please tell us why: ____________________________________ The following is a list of different activities that you may engage in over the course of one week. For each of the activities listed, please check all of those that you have engaged in alone with your roommate in the past week. Check only those activities that were done alone with your roommate and not done with him/her in the presence of others. Question 8: In the past week, I did the following activities alone with my roommate: (Check all that apply)          We did laundry together. We prepared a meal together. We watched TV together. We attended a non-class lecture or presentation together. We went to a restaurant together. We went grocery shopping together. We went for a walk/drive together. We discussed things of a personal nature together. We went to a museum/art show together. Question 9: The following questions concern the amount of influence your roommate has on your thoughts, feelings, and behavior. Please indicate the extent to which you agree or disagree with the following statements. My roommate will influence my future financial security. My roommate does not influence everyday things in my life. My roommate influences important things in my life. My roommate influences which parties and other social events I attend. My roommate influences the extent to which I accept responsibilities in our relationship. My roommate does not influence how much time I spend doing household work. My roommate does not influence how I choose to spend my money. My roommate influences the way I feel about myself. My roommate does not influence my moods. 130 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 1 2 2 3 3 4 4 5 5 6 6 7 7 Question 10: Now we would like you to tell us how much your roommate affects your future plans and goals. Using the 7-point scale from 1-not at all to 7-a great extent, please indicate the degree to which your future plans and goals are affecting by your roommate. My vacation plans My marriage plans My plans to have children My plans to make major investments (house, car, etc.) My plans to join a club, social organization, church, etc. My school-related plans My plans for achieving a particular financial standard of living. 1 1 1 1 1 1 1 2 2 2 2 2 2 2 3 3 3 3 3 3 3 4 4 4 4 4 4 4 5 5 5 5 5 5 5 6 6 6 6 6 6 6 7 7 7 7 7 7 7 Question 11: In the past week, I did the following activities alone with my roommate: (Check all that apply)          We planned a party/social event together. We attended class together. We went on a trip (e.g., vacation or weekend) together. We cleaned our dorm room/apartment together. We went to church/religious function together. We worked on homework together. We discussed things of a non-personal nature together. We went to a clothing store together. We talked on the phone. Question 12: Please indicate the extent to which you agree or disagree with the following statements. My roommate influences the basic values that I hold. My roommate does not influence the opinions that I have of other important people in my life. My roommate does not influence when I see, and the amount of time I spend with my family. My roommate influences when I see, and the amount of time I spend with, my friends. My roommate does not influence which of my friends I see. My roommate does not influence the type of career I have. My roommate influences or will influence how much time I devote to my career. My roommate does not influence my chances of getting a good job in the future. My roommate influences the way I feel about the future. 131 1 1 2 2 3 3 4 4 5 5 6 6 7 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6 7 7 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Question 13: In the past week, I did the following activities alone with my roommate: (Check all that apply)          We played cards/board game together. We attended a sporting event together. We exercised (e.g., jogging, aerobics) together. We went on an outing (e.g., picnic, beach, zoo, winter carnival) together. We did wilderness activity (e.g., hunting, hiking, fishing) together. We went to a concert together. We went dancing together. We went to a park together. We played music/sang together. Question 14: Please indicate the extent to which you agree or disagree with the following statements. My roommate does not have the capacity to influence how I act in various situations. My roommate influences and contributes to my overall happiness. My roommate does not influence my present financial security. My roommate influences how I spend my free time. My roommate influences when I see him/her and the amount of time the two of us spend together. My roommate does not influence how I dress. My roommate influences how I decorate my dorm room. My roommate does not influence where I live. My roommate influences what I watch on TV 1 2 3 4 5 6 7 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7 7 7 7 Question 15: In the past week, I did the following activities alone with my roommate: (Check all that apply)          We went to a movie together. We ate a meal together. We participated in a sporting activity together. We did outdoor recreation (e.g., sailing) together. We went to a play together. We went to a bar together. We visited family together. We visited friends together. We went to a department, book, hardware store, etc. together. 132 Question 16: Now, we would like to know how certain you are about your roommate NOW after living with him/her for more than a month. Please indicate the extent to which you agree with the following statements. I am confident of my general ability to predict how my roommate will behave. 1 2 3 4 5 6 7 I am certain that he/she likes me. 1 2 3 4 5 6 7 I am accurate at predicting the values he/she holds. 1 2 3 4 5 6 7 I am accurate at predicting his/her attitudes. 1 2 3 4 5 6 7 I can predict his/her feelings and emotions. 1 2 3 4 5 6 7 I can empathize with (share) the way my roommate feels about himself/herself. 1 2 3 4 5 6 7 I know him/her very well. Question 17: If you're interested in sharing more detailed information about your Facebook use and your relationship with the roommate, please leave your email address here. We would like to set up a time to talk to you. As a result of your participation in this interview, you will receive a $25 Amazon gift certificate. Thank you. Your email address: ___________________________ Question 18: What you think about your roommate now after living with him/her for more than a month? Please indicate the extent to which you agree with the following statements. I think he/she could be a friend of mine. 1 2 3 4 5 6 7 It would be difficult to meet and talk with him/her. 1 2 3 4 5 6 7 He/She just wouldn’t fit into my circle of friends. 1 2 3 4 5 6 7 We could never establish a personal friendship with each other. 1 2 3 4 5 6 7 I would like to have a friendly chat with him/her. 133 1 2 3 4 5 6 7 Question 19: Please indicate the extent to which you agree with the following statements. I think he/she is quite handsome/pretty. 1 2 3 4 5 6 7 He/She is very sexy looking. 1 2 3 4 5 6 7 I find him/her very attractive physically. 1 2 3 4 5 6 7 I don’t like the way he/she looks. 1 2 3 4 5 6 7 He/She is somewhat ugly. 1 2 3 4 5 6 7 Question 20: Please indicate the extent to which you agree with the following statements. He/She is a typical goof-off when assigned a job to do. 1 2 3 4 5 6 7 I have confident in his/her ability to get the job done. 1 2 3 4 5 6 7 If I wanted to get things done, I could probably depend on him/her. 1 2 3 4 5 6 7 I couldn’t get anything accomplished with him/her. 1 2 3 4 5 6 7 He/She would be a poor problem solver. 1 2 3 4 5 6 7 Question 21: Do you still have a Facebook profile? [Mandatory]  Yes  No [Skip to End] Question 22: Does your roommate still have a Facebook profile?  Yes  No Question 23: How many Facebook friends do you and your roommate have in common? (To find out, please go to your roommate's Facebook profile page. It will show in "Mutual Friends" column on your left hand side.) _________________ Question 24: On average, how often do you spend time on Facebook?  Several times per day  Once a day  Several times a week  Once a week  Several times a month 134  Monthly or less  Other, please specify _______________________ Question 25: How much time do you typically spend on Facebook each day during a typical week day (Monday to Friday)?  less than 1 hour  1 ~ 2 hours  2 ~ 3 hours  3 ~ 4 hours  4 ~ 5 hours  5 ~ 6 hours  6 ~ 7 hours  7 ~ 8 hours  more than 8 hours Question 26: How much time do you typically spend on Facebook each day during a typical weekend day?  less than 1 hour  1 ~ 2 hours  2 ~ 3 hours  3 ~ 4 hours  4 ~ 5 hours  5 ~ 6 hours  6 ~ 7 hours  7 ~ 8 hours  more than 8 hours Question 27: How many friends do you have on Facebook? ________________ Question 28: Please indicate the extent to which you agree with the following statements. Facebook is part of my everyday activities. 1 2 3 4 5 6 7 I am proud to tell people I'm on Facebook. 1 2 3 4 5 6 7 Facebook has become part of my daily routine. 1 2 3 4 5 6 7 I feel out of touch when I haven't logged onto Facebook for a while. 1 2 3 4 5 6 7 I feel I am part of the Facebook community. 1 2 3 4 5 6 7 I would be sorry if Facebook shut down. 1 2 3 4 5 6 7 135 APPENDIX D Interview Questions 1. How many roommates do you have? If you have more than one, please choose the one you chose when answering the surveys. 2. When did you receive your roommate assignment? How well did you know him/her at that time? 3. Did you look him/her up on Facebook? a. When did you add him/her as your friend? b. Who initiates the interaction first? 4. What did you think about his/her profile? a. Did you like it? b. Which part of information in his/her profile was more important to you? The pictures? Background information? His/her interests? His/Her friend network? Wall posts? c. Is there any information which leads to more questions about him/her? 5. In addition to FB, what other communication channels did you use to communicate with him/her? (Phone, emails, text messages, Skype, etc.) a. How often do you communicate? b. Which channel did you use most often? 6. Does FB help reduce your uncertainty about him/her? Questions about in the beginning of living together: 7. Were there any surprises for you after moving in with him/her when the semester started? a. What was the biggest surprise/difference from the Facebook profile? b. How did you deal with it? c. Was there anything you wish that you’d find out earlier before move-in? d. Do you think your relationship with him/her is affected by it? Questions about your relationship with him/her now: 8. How close are you with him/her now? From 1 to 10, 10 as very close. a. Do you do activities together? b. What kind of activities do you do together? c. Do you spend a lot of time together? 136 d. Tell me about your ordinary daily schedule, and the interaction with her during the day. e. Do you want him/her to be your roommate in the future? 9. Do you still get news/updates about him/her on FB? a. Do you still interact with each other on FB? 10. How would you describe your impression about him/her? Getting better? Getting worse? Pretty much the same in a good/bad way? 11. Anything you want to add regarding how FB affects your relationship with the roommate? Thank you. Here is your Amazon gift card code. Do you mind if I contact you in the future if I have some more questions? Thank you. 137 APPENDIX E Scale Items, Desicriptives, and Intercorrelations Table 25 CL7 Attributional Confidence Items CL 7 Items*: 1 I am confident of my general ability to predict how my roommate will behave. 2 I am certain that he/she likes me. 3 I am accurate at predicting the values he/she holds. 4 I am accurate at predicting his/her attitudes. 5 I can predict his/her feelings and emotions. 6 I can empathize with (share) the way my roommate feels about himself/herself. 7 I know him/her very well. *. Response categories range from 1 = strongly disagree to 7 = strongly agree 138 Table 26 Summary of Means, Standard Deviations, and Intercorrelations of Items for CL7 Attributional Confidence Scale Item M SD 1 2 3 4 5 6 7 First-Wave (α = .94) CL7 3.82 1.46 1 4.12 1.70 -- 2 4.62 1.63 .64** -- 3 3.90 1.71 .74** .62** -- 4 3.67 1.73 .75** .61** .75** -- 5 3.30 1.69 .70** .60** .76** .80** -- 6 3.94 1.71 .65** .66** .69** .66** .69** -- 7 3.15 1.88 .63** .67** .66** .70** .75** .66** -- Item M SD 1 2 3 4 5 6 7 Second-Wave (α = .91) CL7 4.81 1.10 1 5.02 1.22 -- 2 5.32 1.21 .51** -- 3 4.89 1.29 .59** .57** -- 4 4.76 1.29 .65** .58** .64** -- 5 4.53 1.34 .64** .59** .65** .78** -- 6 4.88 1.26 .55** .59** .51** .58** .59** -- 7 4.50 1.60 .57** .66** .59** .62** .69** .60** 139 -- Table 26 (cont’d) Item M SD 1 2 3 4 5 6 7 Third-Wave (α = .92) CL7 4.98 1.27 1 5.14 1.48 -- 2 5.32 1.51 .50** -- 3 5.07 1.50 .62** .47** -- 4 5.02 1.47 .73** .48** .70** -- 5 4.85 1.53 .76** .54** .68** .82** -- 6 4.77 1.64 .57** .68** .56** .60** .61** -- 7 4.72 1.73 .53** .68** .57** .60** .62** .65** **. Correlation is significant at the 0.01 level (2-tailed). Table 27 Impression for Social Attractiveness Items Social Attractiveness Items*: 1 I think he/she could be a friend of mine. 2 It would be difficult to meet and talk with him/her.** 3 He/She just wouldn’t fit into my circle of friends. ** 4 We could never establish a personal friendship with each other. ** 5 I would like to have a friendly chat with him/her. *. Response categories range from 1 = strongly disagree to 7 = strongly agree **. Items were reversed before performing statistical analyses. 140 -- Table 28 Summary of Means, Standard Deviations, and Intercorrelations of Items for Social Attractiveness Item M SD 1 2 3 4 5 Frist-Wave (α = .79) Social 5.54 .98 1 5.70 1.19 -- 2 5.55 1.38 .51** -- 3 5.10 1.57 .58** .44** -- 4 5.73 1.40 .50** .46** .39** -- 5 5.64 1.07 .56** .36** .29** .30** -- Item M SD 1 2 3 4 5 Second-Wave (α = .87) Social 5.34 1.16 1 5.55 1.26 -- 2 5.50 1.36 .66** -- 3 4.79 1.68 .64** .56** -- 4 5.50 1.46 .64** .69** .54** -- 5 5.55 1.11 .65** .52** .45** .45** 141 -- Table 28 (cont’d) Item M SD 1 2 3 4 5 Third-Wave (α = .89) Social 5.15 1.45 1 5.17 1.73 -- 2 5.33 1.68 .69** -- 3 4.53 2.02 .61** .65** -- 4 5.37 1.74 .68** .75** .66** -- 5 5.33 1.51 .72** .54** .40** .54** **. Correlation is significant at the 0.01 level (2-tailed). Table 29 Impression for Appearance Attractiveness Items Appearance Attractiveness Items*: 1 I think he/she is quite handsome/pretty. 2 He/She is very sexy looking. 3 I find him/her very attractive physically. 4 I don’t like the way he/she looks.** 5 He/She is somewhat ugly.** *. Response categories range from 1 = strongly disagree to 7 = strongly agree **. Items were reversed before performing statistical analyses. 142 -- Table 30 Summary of Means, Standard Deviations, and Intercorrelations of Items for Appearance Attractiveness Item M SD 1 2 3 4 5 First-Wave (α = .79) Appearance 4.35 1.10 1 4.52 1.50 -- 2 3.30 1.49 .55** -- 3 3.38 1.63 .66** .68** -- 4 5.40 1.46 .40** .18** .21** -- 5 5.39 1.45 .52** .26** .28** .65** 1 2 3 4 Item M SD -5 Second-Wave (α = .82) Appearance 4.30 1.10 1 4.38 1.48 -- 2 3.59 1.35 .70** -- 3 3.55 1.52 .63** .72** -- 4 5.15 1.34 .43** .22** .15** -- 5 5.05 1.40 .56** .34** .27** .68** 143 -- Table 30 (cont’d) Item M SD 1 2 3 4 5 Third-Wave (α = .83) Appearance 4.23 1.15 1 4.33 1.58 -- 2 3.44 1.42 .73** -- 3 3.44 1.51 .69** .77 -- 4 5.00 1.44 .47** .29** .26** -- 5 5.03 1.46 .51** .25** .27** .74** **. Correlation is significant at the 0.01 level (2-tailed). Table 31 Impressions for Task Attractiveness Items Task Attractiveness Items*: 1 He/She is a typical goof-off when assigned a job to do.** 2 I have confident in his/her ability to get the job done. 3 If I wanted to get things done, I could probably depend on him/her. 4 I couldn’t get anything accomplished with him/her.** 5 He/She would be a poor problem solver.** *. Response categories range from 1 = strongly disagree to 7 = strongly agree **. Items were reversed before performing statistical analyses. 144 -- Table 32 Summary of Means, Standard Deviations, and Intercorrelations of Items for Task Attractiveness Item M SD 1 2 3 4 5 First-Wave (α = .82) Task 5.13 .94 1 4.97 1.26 -- 2 5.05 1.19 .40** -- 3 5.06 1.14 .45** .69** -- 4 5.42 1.26 .44** .43** .47** -- 5 5.12 1.28 .48** .49** .49** .47** -- Item M SD 1 2 3 4 5 Second-Wave (α = .83) Task 5.19 1.03 1 5.21 1.32 -- 2 5.29 1.20 .38** -- 3 5.11 1.33 .35** .75** -- 4 5.40 1.28 .47** .49** .50** -- 5 5.12 1.39 .47** .46** .45** .56* 145 -- Table 32 (cont’d) Item M SD 1 2 3 4 5 Third-Wave (α = .89) Task 5.02 1.29 1 5.12 1.51 -- 2 4.94 1.58 .56** -- 3 4.85 1.58 .52** .83** -- 4 5.19 1.53 .60** .63** .62** -- 5 4.99 1.54 .59** .57** .55** .64** -- **. Correlation is significant at the 0.01 level (2-tailed). Table 33 RCI Frequency Items 1 2 3 DURING THE PAST WEEK, what is the average amount of time, per day, that you spent alone with your roommate in the MORNING (e.g., between the time you wake and 12 noon)? ___ hour(s) ___minutes DURING THE PAST WEEK, what is the average amount of time, per day, that you spent alone with your roommate in the AFTERNOON (e.g., between 12 noon and 6 pm)? ___ hour(s) ___minutes DURING THE PAST WEEK, what is the average amount of time, per day, that you spent alone with your roommate in the EVENING (e.g., between 6 pm and bedtime)? ___ hour(s) ___minutes 146 Table 34 RCI Diversity Items 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 In the past week, I did the following activities alone with my roommate: (Check all that apply) did laundry prepared a meal watched TV attended a non-class lecture or presentation went to a restaurant went grocery shopping went for a walk/drive discussed things of a personal nature went to a museum/art show planned a party/social event attended class went on a trip (e.g., vacation or weekend) cleaned our dorm room went to church/religious function worked on homework discussed things of a non-personal nature went to a clothing store talked on the phone went to a movie ate a meal participated in a sporting activity outdoor recreation (e.g., sailing) went to a play went to a bar visited family visited friends went to a department, book, hardware store, etc. played cards/board game attended a sporting event exercised (e.g., jogging, aerobics) went on an outing (e.g., picnic, beach, zoo, winter carnival) wilderness activity (e.g., hunting, hiking, fishing) went to a concert went dancing went to a par played music/sang 147 Table 35 RCI Strength Items 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 My roommate will influence my future financial security. My roommate does not influence everyday things in my life. My roommate influences important things in my life. My roommate influences which parties and other social events I attend. My roommate influences the extent to which I accept responsibilities in our relationship. My roommate does not influence how much time I spend doing household work. My roommate does not influence how I choose to spend my money. My roommate influences the way I feel about myself. My roommate does not influence my moods. My roommate influences the basic values that I hold. My roommate does not influence the opinions that I have of other important people in my life. My roommate does not influence when I see, and the amount of time I spend with my family. My roommate influences when I see, and the amount of time I spend with, my friends. My roommate does not influence which of my friends I see. My roommate does not influence the type of career I have. My roommate influences or will influence how much time I devote to my career. My roommate does not influence my chances of getting a good job in the future. My roommate influences the way I feel about the future. My roommate does not have the capacity to influence how I act in various situations. My roommate influences and contributes to my overall happiness. My roommate does not influence my present financial security. My roommate influences how I spend my free time. My roommate influences when I see him/her and the amount of time the two of us spend together. My roommate does not influence how I dress. My roommate influences how I decorate my dorm room. My roommate does not influence where I live. My roommate influences what I watch on TV My vacation plans My marriage plans My plans to have children My plans to make major investments (house, car, etc.) My plans to join a club, social organization, church, etc. My school-related plans My plans for achieving a particular financial standard of living. 148 Table 36 RCI Conversion Table Scale Score 1 2 3 4 5 6 7 8 9 10 Frequency (minutes) 0-12 13-48 49-108 109-192 193-300 301-432 433-588 589-768 769-972 973-1200 Diversity (number of activities) 0 1 2-3 4-6 7-9 10-13 14-18 19-24 25-30 31-38 149 Strength (strength total) 34-53 54-73 74-93 94-113 114-133 134-153 154-173 174-193 194-213 214-238 REFERENCES 150 REFERENCES Afifi, W., & Burgoon, J. K. (2000). The impact of violations on uncertainty and the consequences for attractiveness. Human Communication Research, 26, 203-233. Antheunis, M. L., Valkenburg, P. M., & Peter, J. (2010). Getting acquainted through social network sites: Testing a model of online uncertainty reduction and social attraction. Computers in Human Behavior, 26, 100-109. Baym, N. K. (2001). Interpersonal life online. In L. Lievrouw & S. Livingstone (Eds.), Handbook of new media (pp. 62-76). Thousand Oaks, CA: Sage Publications. Berger, C. R. (1979). Beyond initial interaction: Uncertainty, understanding, and the development of interpersonal relationships. In H. Giles & R. N. S. Clair (Eds.), Language and social psychology (pp. 122-144). Baltimore: University Park Press. Berger, C. R., & Calabrese, R. J. (1975). Some explorations in initial interaction and beyond: Toward a developmental theory of interpersonal communication. Human Communication Research, 1, 99-112. Berscheid, E. (1985). Interpersonal attraction. In G. Lindzy & E. Aronson (Eds.), Handbook of social psychology (3rd ed., pp. 413-484). New York: Random House. Berscheid, E., & Reis, H. T. (1998). Attraction and close relationships. In D. Gilbert, S. T. Fiske & G. Lindzey (Eds.), The handbook of social psychology (4th ed., Vol. 2, pp. 193-281). Boston: McGraw-Hill. Berscheid, E., Snyder, M., & Omoto, A. M. (1989). The relationship closeness inventory: Assessing the closeness of interpersonal relationship. Journal of Personality and Social Psychology, 57, 792-807. boyd, d. m., & Ellison, N. B. (2007). Social network sites: Definition, history, and scholarship. Journal of Computer-Mediated Communication, 13(1). Retrieved from http://jcmc.indiana.edu/vol13/issue1/boyd.ellison.html Brandenburg, C. (2008). The newest way to screen job applicants: A social networker's nightmare. Federal Communications Law Journal, 60, 597-626. Burgoon, J. K., & Hale, J. L. (1988). Nonverbal expectancy violations: Model elaboration and application to immediacy behaviors. Communication Monographs, 55, 58-79. Cass, A. (2011). Defining stalking: The influence of legal factors, extralegal factors, and paricular actions on judgments of college students. Western Criminology Review, 12(1), 1-14. 151 Clatterbuck, G. W. (1979). Attributional confidence and uncertainty in initial interaction. Human Communication Research, 5, 147-157. Collura, H. (2007, August 7). Facebook pages concern parents of college freshmen. USA Today. from http://www.usatoday.com/news/nation/2007-08-07-facebook-housing_N.htm#. Corwin, J. R., & Cintron, R. (2011). Social networking phenomena in the first-year experience. Journal of College Teaching & Learning, 8(1), 25-37. Dion, K. K., Berscheid, E., & Walster, E. (1972). What is beautiful is good. Journal of Personality and Social Psychology, 24, 285-290. Donath, J., & boyd, d. (2004). Public displays of connection. BT Technology Journal, 22(4), 7182. Douglas, W. (1990). Uncertainty, information-seeking, and liking during initial interaction. Western Journal of Speech Communication, 54, 66-81. Du, W. (2007). Job candidates getting tripped up by Facebook: Many students learn the hard way that online image can limit opportunity. msnbc.com. from http://www.msnbc.msn.com/id/20202935/ns/business-personal_finance/t/job-candidatesgetting-tripped-facebook/. Eberhardt, D. M. (2007). Facing up to facebook. About Campus, 12(4), 18-26. Ellison, N., Heino, R., & Gibbs, J. (2006). Managing impressions online: Self-presentation processes in the online dating environment. Journal of Computer-Mediated Communication, 11(2). Retrieved from http://jcmc.indiana.edu/vol11/issue2/ellison.html Ellison, N., Steinfield, C., & Lampe, C. (2007). The benefits of Facebook "friends:" Social capital and college students' use of online social network sites. Journal of ComputerMediated Communication, 12(4). Retrieved from http://jcmc.indiana.edu/vol12/issue4/ellison.html Facebook (2011). Press room: Statistics. Retrieved 5/1/2011, from http://www.facebook.com/press/info.php?statistics Fisher, B. S., Cullen, F. T., & Turner, M. G. (2002). Being pursued: Stalking victimization in a national study of college women Criminology & Public Policy, 1, 257-308. Gibbs, J., Ellison, N., & Heino, R. (2006). Self-presentation in online personals: The role of anticipated future interaction, self-disclosure, and perceived success in internet dating. Communication Research, 33, 152-177. Gosling, S., Gaddis, S., & Vazire, S. (2007, March). Personality impressions based on Facebook profiles. Paper presented at the International Conference on Weblogs and Social Media, Boulder, Colorado. 152 Gosling, S., Gaddis, S., & Vazire, S. (2008). First impressions based on the environments we create and inhibit. In J. Skowrouski & N. Ambady (Eds.), First impressions (pp. 334356). New York: Guilford. Hancock, J. T., & Dunham, P. J. (2001). Impression formation in computer-mediated communication revisited: An analysis of the breadth and intensity of impressions. Communication Research, 28, 325-347. Haythornthwaite, C. (2002). Strong, weak, and latent ties and the impact of new media. The Information Society, 18, 385-401. Haythornthwaite, C., & Wellman, B. (1998). Work, friendship, and media use for information exchange in a networked organization. Journal of The American Society for Information Science, 49, 1101-1114. Jacobson, D. (1999). Impression formation in cyberspace: Online expectations and offline experiences in text-based virtual communities. Journal of Computer-Mediated Communication, 5(1). Retrieved from http://jcmc.indiana.edu/vol5/issue1/jacobson.html Joinson, A. N. (2008). 'Looking at', 'looking up', or 'keeping up with' people? Motives and Uses of Facebook. Paper presented at the the 26th Annual SIGCHI Conference on Human Factors in Computing Systems, Florence, Italy. Kellermann, K. (1984). The negativity effect and its implication for initial interaction. Communication Monographs, 51, 37-55. Kellermann, K. (1986). Anticipation of future interaction and information exchange in initial interaction. Human Communication Research, 13, 41-75. Kelly, H. H., Berscheid, E., Christensen, A., Harvey, J. H., Huston, T. L., Levinger, G., et al. (1983). Close relationships. New York: Freeman. Knobloch, L., Satterlee, K., & DiDomenico, S. (2010). Relational uncertainty predicting appraisals of face threat in courtship: Integrating uncertainty reduction theory and politeness theory. Communication Research, 37, 303-334. Knobloch, L., & Solomon, D. (2002). Information seeking beyond initial interaction: Negotiating relationship uncertainty within close relationships. Human Communication Research, 28, 243-257. Lampe, C., Ellison, N., & Steinfield, C. (2006). A Face(book) in the crowd: Social searching vs. social browsing. Paper presented at the ACM Special Interest Group on ComputuerSupported Cooperative Work, Banff, Canada. Lampe, C., Ellison, N., & Steinfield, C. (2007, April). A familiar face(book): Profile elements as signals in an online social network. Paper presented at the Proceedings of the SIGCHI conference on Human factors in computing systems, San Jose, California. 153 Lampe, C., Ellison, N. B., & Steinfield, C. (2008, November). Changes in use and perception of Facebook. Paper presented at the ACM Conference on Computer-Supported Cooperative Work, San Diego, CA. Leafgren, F. A. (1989). Health and wellness programs. In M. L. Upcraft & J. N. Gardner (Eds.), The freshmen year experience: Helping students survive and succeed in college (pp. 156167). San Francisco: Jossey-Bass. Leary, M. R. (1996). Self-presentation: Impression management and interpersonal behavior. Boulder, CO: Westview Press, Inc. Lenhart, A., & Madden, M. (2007). Social networking websites and teens: An overview. The Pew Internet & American Life Project. Retrieved February 26, 2007, from http://www.pewinternet.org/pdfs/PIP_SNS_Data_Memo_Jan_2007.pdf Lenhart, A., Purcell, K., Smith, A., & Zickuhr, K. (2010). Social media & mobile Internet use among teens and young adults. The Pew Internet & American Life Project, from http://www.pewinternet.org/~/media//Files/Reports/2010/PIP_Social_Media_and_Young _Adults_Report_Final_with_toplines.pdf Lou, L. L. (2009). Loneliness, friendship, and self-esteem: First-year college students' experience of using Facebook. Unpublished Doctoral Dissertation, State University of New York, Albany. McCroskey, J. C., & McCain, T. A. (1974). The measurement of interpersonal attraction. Speech Monographs, 41, 261-266. McCroskey, L. L., McCroskey, J. C., & Richmond, V. P. (2006). Analysis and improvement of the measurement of interpersonal attraction and homophily. Communication Quarterly, 54, 1-31. Pempek, T. A., Yermolayeva, Y. A., & Calvert, S. L. (2009). College students' social networking experiences on Facebook. Journal of Applied Developmental Psychology, 30, 227-238. Planalp, S., & Honeycutt, J. M. (1985). Events that increase uncertainty in personal relationships. Human Communication Research, 11, 593-604. Ramirez, A., Jr. (2007). The effect of anticipated future interaction and initial impression valence on relational communication in computer-mediated interaction. Communication Studies, 58, 53-70. Ramirez, A., Jr., & Burgoon, J. K. (2004). The effect of interactivity on initial interactions: the influence of information valence and modality and information richness on computermediated interaction. Communication Monographs, 71, 422-447. Ramirez, A., Jr., Walther, J. B., Burgoon, J. K., & Sunnafrank, M. (2002). Information-seeking strategies, uncertainty, and computer-mediated communication toward a conceptual model. Human Communication Research, 28, 213. 154 Ramirez, A., Jr., & Wang, Z. (2008). When online meets offline: An expectancy violations theory perspective on modality switching. Journal of Communication, 58, 20-39. Scissors, L. (2007). Facebook your future roommate. The Pew Internet & American Life Project, from http://www.pewinternet.org/Commentary/2007/August/Facebook-Your-FutureRoommate.aspx Shaw Taylor, L., Fiore, A. T., Mendelsohn, G. A., & Cheshire, C. (2010, May). A second chance to make a first impression: Factors affecting the longevity of online dating relationships. Paper presented at the The 4th International AAAI Conference on Weblogs and Social Media, Washington, DC. Smith, M. L., & Blanchard, C. (2008, January 9). Facebook photos land Eden Prairie kids in trouble. StarTribune. from http://www.startribune.com/local/west/13549646.html. Steinfield, C., Ellison, N. B., & Lampe, C. (2008). Social capital, self-esteem, and use of online social network sites: A longitudinal analysis. Journal of Applied Developmental Psychology, 29, 434-445. Sunnafrank, M. (1986). Predicted outcome value during initial interactions: A reformulation of uncertainty reduction theory. Human Communication Research, 13, 3-33. Sunnafrank, M. (1990). Predicted outcome value and uncertainty reduction theories: A test of competing perspectives. Human Communication Research, 17, 76-103. Tidwell, L. C., & Walther, J. B. (2002). Computer-mediated communication effects on disclosure, impressions, and interpersonal evaluations: Getting to know one another a bit at a time. Human Communication Research, 28, 317-348. Tinto, V. (1997). Classrooms as communities: Exploring the education character of studnet persistence. The Journal of Higher Education, 68, 599-623. Toma, C. L., Hancock, J. T., & Ellison, N. B. (2008). Separating fact from fiction: An examination of deceptive self-presentation in online dating profiles. Personality and Social Psychology Bulletin, 34, 1023-1036. Tong, S. T., Van Der Heide, B., Langwell, L., & Walther, J. B. (2008). Too much of a good thing? The relationship between number of friends and interpersonal impressions on Facebook. Journal of Computer-Mediated Communication, 13, 531-549. Van Laar, C., Levin, S., Sinclair, S., & Sidanius, J. (2005). The effect of university roommate contact on ethnic attitudes and behavior. Journal of Experimental Social Psychology, 41, 329-345. Walsh-Sarnecki, P. (2007, August 27). Facebook unites or divides roommates. Detroit Free Press. 155 Walther, J. B. (1992). Interpersonal effects in computer-mediated interaction: A relationship perspective. Communication Research, 19, 52-90. Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal, and hyperpersonal interaction. Communication Research, 23, 1-43. Walther, J. B. (1997). Group and interpersonal effects in international computer-mediated collaboration. Human Communication Research, 23, 342-369. Walther, J. B., & Parks, M. R. (2002). Cues filtered out, cues filtered in: Computer-mediated communication and rleationships. In M. L. Knapp & J. L. Daly (Eds.), Handbook of interpersonal communcation (3rd ed., pp. 529-563). Thousand Oaks, CA: Sage. Walther, J. B., Slovacek, C. L., & Tidwell, L. C. (2001). Is a picture worth a thousand words? Photographic images in long-term and short-term computer-mediated communication. Communication Research, 28, 105-134. Walther, J. B., Van Der Heide, B., Hamel, L. M., & Shulman, H. C. (2009). Self-generated versus other-generated statements and impressions in computer-mediated communication: A test of warranting theory using Facebook. Communication Research, 36, 229-253. Walther, J. B., Van Der Heide, B., Kim, S.-Y., Westerman, D., & Tong, S. T. (2008). The role of friends' appearance and behavior evaluations of individuals on Facebook: Are we know by the company we keep? Human Communication Research, 34, 28-49. Yoo, J. H. (2003, November). Uncertainty reduction and information valence: Tests of uncertainty reduction theory, predicted outcome value, and an alternative explanation? Paper presented at the the annual conference of the National Communication Association, Miami Beach, FL. 156