What is the difference between a hyperpersonal and face to face relationship?
|
This study investigated the impact of (the lack of) audiovisual cues during conversations preceding a first face-to-face meeting among prospective daters on daters' perceptions of partners' social and romantic attraction. Additionally, the study examined the effect of modality switching, from online to offline interaction. Thirty-nine individuals participated in a round-robin speed dating event, resulting in 95 unique conversation pairings. For their first conversations they were randomly assigned to meet via text-based CMC or videoconferencing. The dyads then had a second encounter, which was face-to-face. Results showed more social attraction between interactants who used text-based CMC than videoconferencing, supporting the hyperpersonal model of CMC. Furthermore, after a modality switch to a face-to-face encounter the hyperpersonal effect persisted for social attraction, while romantic attraction declined. Show
Journal Article Erin K. Ruppel, 1Department of Communication, University of Wisconsin-Milwaukee, Milwaukee, WI 53201 Search for other works by this author on: Clare Gross,2Department of Communication Arts and Sciences, Baldwin Wallace University, Berea, OH 44017 Search for other works by this author on: Arrington Stoll,3Department of Communication, University of Wisconsin-Milwaukee, Milwaukee, WI 53201 Search for other works by this author on: Brittnie S. Peck,3Department of Communication, University of Wisconsin-Milwaukee, Milwaukee, WI 53201 Search for other works by this author on: Mike Allen,3Department of Communication, University of Wisconsin-Milwaukee, Milwaukee, WI 53201 Search for other works by this author on: Sang-Yeon Kim3Department of Communication, University of Wisconsin-Milwaukee, Milwaukee, WI 53201 Search for other works by this author on: Revision received: 02 August 2016 Accepted: 24 October 2016 Revision received: 26 October 2016 Published: 30 November 2016
Close Navbar Search Filter Microsite Search Term Search AbstractSelf-disclosure is a key concept in computer-mediated communication (CMC) theory and research, but disagreement exists about the impact of CMC, relative to face-to-face (FtF) communication, on self-disclosure. We conducted a meta-analysis of studies comparing self-disclosure in CMC and FtF communication to summarize and clarify existing research. We also examined potential moderators of this difference—measure of self-disclosure, study design (survey or experiment), interaction context (task or social), type of CMC (text-based or video-based), and interaction length. Overall, self-disclosure was higher in FtF communication than in CMC. Measure of self-disclosure, study design, and type of CMC moderated this difference. Findings suggest mixed support for predictions derived from key CMC theories and a need for CMC theory to more explicitly address self-disclosure. Self-disclosure—the verbal revealing of personal information, thoughts, or feelings about oneself (Derlega, Metts, Petronio, & Margulis, 1993)—“has emerged as one of the most salient and critical behaviors in CMC” (Jiang, Bazarova, & Hancock, 2011, p. 59). Self-disclosure is a crucial component of relationship development and maintenance in both computer-mediated (CM) and face-to-face (FtF) relationships (Altman & Taylor, 1973; Yum & Hara, 2005). It is also relevant to several theories of interpersonal computer-mediated communication (CMC), particularly social presence theory (Hooi & Cho, 2014; Short, Williams, & Christie, 1976), social information processing (SIP) theory (Tidwell & Walther, 2002; Walther, 1992), and hyperpersonal theory (Jiang, Bazarova, & Hancock, 2013; Walther, 1996). A clear understanding of whether and how CM and FtF self-disclosure differ is crucial to theory and research on CM relational processes. Disagreement exists, however, about the consistency, size, and direction of differences. Some scholars argue for a “robust” (Jiang et al., 2013, p. 128) and “pervasive” (Suler, 2004, p. 321) effect that favors CM self-disclosure. Others have noted inconsistent findings regarding differences in CM and FtF self-disclosure (Kim & Dindia, 2011; Nguyen, Bin, & Campbell, 2012; Ruppel, 2015). Research directly comparing CMC and FtF communication is necessary not only to understanding how these modalities differ but also to contextualizing research that only examines CMC (Walther, 2011). The current meta-analysis aims to synthesize and clarify existing findings regarding differences in CM and FtF self-disclosure and, as a result, provide an integrated view of research on these differences, suggest implications for relevant CMC theories, and propose directions for future research in this area. Although reviews of this topic have been conducted (Kim & Dindia, 2011; Nguyen et al., 2012), this meta-analysis advances work in this area because it uses a larger sample of studies and quantitatively tests potential moderators of the observed effects. We focus on breadth and depth of self-disclosure in text-based (e.g., instant messaging) and video-based (e.g., video chat) CMC because these forms have been most frequently compared to FtF communication and represent interpersonal, as opposed to non-directed (e.g., on social networking sites), self-disclosure (Nguyen et al., 2012).1 Existing Views of Differences in CM and FtF Self-DisclosureInconsistent findings have generated divergent conclusions regarding differences in CM and FtF self-disclosure. Some researchers have concluded that self-disclosure is reliably higher in CMC than FtF. One review of research in this area observed that, “Discussions of online communication often assume greater self-disclosure in computer-mediated communication (CMC) than face-to-face (FTF)” (Nguyen et al., 2012, p. 103). Another review of CM disclosure observed that, “A rapidly increasing body of experimental and anecdotal evidence suggests that CMC and general Internet-based behavior can be characterized as containing high levels of self-disclosure” (Joinson & Paine, 2007, p. 239). Similarly, Schouten, Valkenburg, and Peter wrote that both survey and experimental studies “have repeatedly demonstrated that CMC stimulates self-disclosure” (2007, p. 293). Jiang and colleagues also stated that research has “found a robust tendency for people to self-disclose more frequently in text-based computer-mediated compared to FtF interactions” (2013, p. 128). This idea has been echoed by other researchers (e.g., Joinson, 2003; Suler, 2004). Others have argued, however, that research findings regarding CM and FtF self-disclosure are inconsistent and inconclusive. Ruppel (2015a) observed that experimental research often finds more self-disclosure in CMC than in FtF communication but that findings from survey research are mixed. In reference to existing survey research, she concluded, “It is, consequently, unclear precisely when and how self-disclosure is related to [communication technology] use in ongoing relationships” (Ruppel, 2015a, p. 668). Similarly, a qualitative review of previous research on CM and FtF self-disclosure concluded that differences in self-disclosure are inconsistent and do not provide clear support for any of the theories to which the authors applied the findings (e.g., hyperpersonal theory, SIP theory; Nguyen et al., 2012). A small meta-analysis of studies comparing CMC and FtF self-disclosure found no overall difference between them; the researchers concluded that there was mixed support for SIP theory and no support for hyperpersonal theory (Kim & Dindia, 2011). Kim and Dindia also noted that the overall effect was heterogeneous, indicating the presence of potential moderating variables, but that their small sample precluded testing for moderation (Kim & Dindia, 2011). Because disagreement exists about the consistency and direction of findings regarding differences in CM and FtF self-disclosure, the current meta-analysis aims to quantify and clarify these differences. In contrast to qualitative reviews of the literature (e.g., Nguyen et al., 2012), meta-analysis has the advantage of quantitatively summarizing existing research and testing potential moderating variables to draw conclusions across existing studies. The current meta-analysis is based on a larger sample of studies than was used in Kim and Dindia's (2011) meta-analysis, allowing more robust estimation of overall differences in CM and FtF self-disclosure and tests of the potential moderating variables described below. In addition to providing a comprehensive view of previous research in this area, the results of the meta-analysis can help determine which aspects of key CMC theories are consistent with existing research and what potential boundary conditions might be applied to these theories. The results can also suggest future directions for research and theory testing in this area. Self-disclosure is crucial to developing and maintaining both CM and FtF personal relationships (e.g., Altman & Taylor, 1973; Valkenburg & Peter, 2009; Yum & Hara, 2005). Clear knowledge about differences between CM and FtF self-disclosure is therefore key to understanding these processes. Self-Disclosure in CMC TheoryDifferences in CM and FtF self-disclosure are relevant to several key theories of interpersonal CMC, particularly social presence theory (Short et al., 1976), SIP theory (Walther, 1992), and hyperpersonal theory (Walther, 1996). Although these theories do not make explicit predictions regarding CM and FtF self-disclosure, each theory proposes different communication mechanisms that generate predictions regarding self-disclosure (Kim & Dindia, 2011; Nguyen et al., 2012; Walther, 2011). A clearer understanding of whether and when differences between CM and FtF self-disclosure are observed could not only provide insight into the mechanisms proposed by each of these theories but also suggest boundary conditions for them. Articulating such boundary conditions is essential to the refinement of CMC theories and the progress of the field of CMC (Walther, 2009). Social Presence TheorySocial presence theory belongs to a group of theories often referred to as cues-filtered-out theories and has experienced renewed interest from researchers (Walther, 2011). Social presence refers to the perceived degree to which a communication medium conveys interactants' presence. Presence involves perceptions of both intimacy and immediacy and operates as a function of the availability of verbal, nonverbal, and contextual cues in that medium (Short et al., 1976). Although social presence is based on interactants' perceptions of a given medium, and is therefore subjective, FtF communication is typically considered the medium with the most social presence, and text-based CMC such as e-mail typically rates lowest in social presence (Rice, 1993). Some research suggests that low social presence can encourage self-disclosure (Bailenson, Yee, Merget, & Schroeder, 2006; Sproull, Subramani, Kiesler, Walker, & Waters, 1996). Lower feelings of identifiability in media with lower social presence (such as text-based CMC) can reduce worry about negative evaluations or repercussions for one's self-disclosure. Media with low social presence might, as a result, encourage self-disclosure (Hooi & Cho, 2014; Nguyen et al., 2012). Nguyen and colleagues (2012) found some support for this contention, with self-disclosure frequency being higher in CMC than in FtF communication in most of the studies they examined, and self-disclosure depth being higher in some studies. In contrast, Kim and Dindia (2011) found an overall null effect of CMC on self-disclosure. In the current meta-analysis, significantly higher CM disclosure than FtF disclosure could suggest support for the idea that lower social presence encourages self-disclosure. SIP TheoryAccording to SIP theory, the lack of nonverbal cues in many forms of CMC causes relational information to be exchanged more slowly. As a result, relationships develop more slowly via CMC than FtF but eventually reach equivalent levels of development. The time-restricted experiments sometimes used in research might not capture this process (Walther, 1992). SIP theory also posits that interactants adapt to the lack of nonverbal cues in CMC by verbally exchanging more relational information than they would FtF (Walther, 1992). For example, one study employing SIP theory found that self-disclosure as a proportion of total utterances was higher when participants interacted via text-based CMC (instant messenger) than when they interacted FtF (Tidwell & Walther, 2002). Within SIP theory, self-disclosure is expected to be higher in CMC than FtF because it functions to facilitate relationship development and decrease uncertainty in the absence of nonverbal cues (Tidwell & Walther, 2002; Walther, 2011). In their literature review, Nguyen and colleagues (2012) concluded that findings regarding self-disclosure frequency, and potentially self-disclosure breadth, supported SIP theory. Kim and Dindia (2011) also concluded that there was tentative support for SIP theory. Hyperpersonal TheoryRelationships that develop via CMC can sometimes become more positive than comparable FtF relationships, a phenomenon known as hyperpersonal communication (Walther, 1996). Hyperpersonal communication involves four characteristics of CMC. First, the lack of nonverbal cues in many forms of CMC allows senders to present themselves more deliberately and strategically. Second, receivers possess limited information with which to develop impressions of senders. Any information received becomes especially important to their impressions, resulting in overattribution based on that information. Third, the channel used to convey the message facilitates this process because CMC often allows interactants to construct and edit messages before transmission. Fourth, the feedback process in CMC encourages behavioral confirmation: Receivers' messages reflect positive impressions of senders, and the positive impressions reinforce senders' positive self-presentation (Walther, 1996; 2011). Self-disclosure is related to strategic self-presentation, overattribution, and behavioral confirmation in this process. Heightened perceptions of intimacy of CM self-disclosure (relative to FtF self-disclosure) and reciprocation of that self-disclosure are expected to encourage CM self-disclosure (Jiang et al., 2011; Jiang et al., 2013). CMC, by encouraging greater intimacy, should increase self-disclosure frequency, breadth, and depth (Nguyen et al., 2012; Walther, 2011). In their literature review, Nguyen and colleagues (2012) found mixed support for hyperpersonal theory, with findings regarding self-disclosure frequency and depth generally providing support for the theory's predictions and findings regarding self-disclosure breadth providing inconsistent support for these predictions. In contrast, Kim and Dindia (2011) did not find support for hyperpersonal theory, with findings indicating no overall difference between CM and FtF self-disclosure. Moderating VariablesDifferences between CM and FtF self-disclosure appear to be at least partly attributable to certain moderator variables (Kim & Dindia, 2011). As Joinson observes, “Although the ‘natural’ state of the Internet may well encourage self-disclosure through a reduction in accountability concerns and an increase in private self-awareness, these effects are neither universal, nor should they be taken for granted” (2003, p. 133). Drawing on previous research and theory regarding interpersonal CMC, we examined the following potential moderating variables: 1) measure of self-disclosure (breadth, depth, or other); 2) study design (experimental or survey); 3) interaction context (social or task); 4) type of CMC examined (text-based or video-based); and 5) amount of time allowed for interaction in experimental studies. Measure of self-disclosureFindings regarding breadth, depth, and other measures of self-disclosure (e.g., frequency) appear to differ (Nguyen et al., 2012). Self-disclosure breadth refers to the number of topics or range of information that is shared, whereas depth refers to the personalness or intimacy of information that is shared (Altman & Taylor, 1973). Although all three of the theories reviewed above would generally predict differences in CM and FtF communication self-disclosure, they also suggest some potential subtleties in these predictions. SIP theory proposes that CM self-disclosure serves to replace nonverbal and contextual information that is available in FtF communication but not in CMC (Walther, 1992; Walther, 2011). Thus, differences in self-disclosure breadth might be particularly pronounced. In contrast, hyperpersonal theory proposes heightened intimacy as the driver of CM self-disclosure (Jiang et al., 2011; Walther, 1996). From this perspective, differences in CM and FtF self-disclosure depth might be most apparent (with higher self-disclosure depth via CMC than FtF). Study designExperimental studies appear to find generally more consistent support for higher self-disclosure in CMC than do survey studies (Nguyen et al., 2012; Ruppel, 2015a). Experimental studies usually involve zero-history dyads with no anticipation of future interaction and use observed, as opposed to self-reported, measures of self-disclosure. In contrast, survey studies typically ask participants about self-nominated friends and use self-report measures of self-disclosure (Nguyen et al., 2012; Ruppel, 2015a). Because of the way survey and experimental studies of CM and FtF self-disclosure have been conducted, study design, relationship, and method of measuring self-disclosure are confounded. We have mirrored previous discussions of these differences and use study design as the moderating variable, though we acknowledge that there are multiple issues at play in this variable (and elaborate on them in the discussion section). Interaction contextThe context of the interaction, and the norms that accompany it, is often considered to be more salient in FtF communication than in CMC (Nguyen et al., 2012). In socially oriented interactions (e.g., a getting-to-know-you task), social presence might be a more salient factor than in task-oriented interactions (e.g., a decision-making task). As a result, we might expect to see stronger effects of CMC versus FtF communication on self-disclosure in socially oriented interactions than in task-oriented interactions. Task-oriented interactions are likely to call for low self-disclosure in both CMC and FtF communication. Thus, the more impersonal nature of task-oriented interactions might lead to smaller or nonexistent differences between CM and FtF self-disclosure. In contrast, social presence is likely to be more relevant in socially oriented interactions, leading to larger differences between CM and FtF self-disclosure. Type of CMCThe type of CMC used could also affect findings (Nguyen et al., 2012). Forms of CMC with relatively more cues (e.g., video chat) might differ from text-based forms of communication (Walther, 2011). All three theories described above point to reduced cues as driving differences in CM and FtF self-disclosure. In social presence theory, nonverbal cues are central to perceptions of intimacy (Short et al., 1976). In SIP theory, self-disclosure functions as a replacement for nonverbal cues that are not available via CMC (Walther, 2011). Walther (2011) also reviews recent research regarding SIP theory that suggests that adding visual or vocal cues to CMC increases the rate of information sharing. Finally, hyperpersonal theory posits that the lack of cues in text-based CMC encourages more (and more intimate) self-disclosure (Walther, 2011). Based on the theories, we would expect differences in CM and FtF self-disclosure to be greater for text-based CMC than for CMC that offers visual and vocal cues. Interaction lengthTime limits in experimental studies could create time pressure that makes interactions more task-focused and impersonal (Walther, 1992). Time as a moderating variable is particularly relevant to SIP theory, which posits that relational information accumulates more slowly in CMC than FtF, but can reach similar levels with enough time (Walther, 1992). The theory proposes that “given sufficient time and message exchanges for interpersonal impression formation and relational development to accrue, and all other things being equal, relational valences in later periods of CMC and face-to-face communication will be the same” (Walther, 1992, p. 69). When participants are given more time to interact, relational communication should be more similar in CMC and FtF conditions. Thus, we might expect differences in CM and FtF self-disclosure to be smaller in experiments that allow participants more time to interact. Alternatively, research using hyperpersonal theory has found a stronger reciprocal relationship between self-disclosure and intimacy in CMC relative to FtF communication (Jiang et al., 2011). This feedback loop might amplify differences between CM and FtF self-disclosure over time, with CM self-disclosure becoming increasingly higher relative to FtF self-disclosure. MethodLiterature Search ProcedureThe electronic databases Psychlit, Medline, Academic Search Complete, ERIC, Communication & Mass Media Complete, PsycARTICLES, PsycCRITIQUES, PsycINFO, Social Sciences Full Text, ProQuest Dissertations and Theses, and Google Scholar were searched to find studies for inclusion. The search used was: self-disclosure AND (“face-to-face” OR online OR “mediated communication” OR internet OR offline). Reviews of the literature related to the topic (Kim & Dindia, 2011; Nguyen et al., 2012) were also examined for references, and the reference page of every included article was examined for studies for inclusion. To be included, a manuscript had to be available and meet the following conditions: a) be available in English; b) incorporate some measure of self-disclosure; c) provide a comparison of FtF communication to CMC; and d) contain sufficient information to estimate the difference between the CM and FtF self-disclosure. A total of 31 estimates from 25 studies met these criteria and were included in the analyses. Included studies and the ways they were coded for moderator variables are listed in Table 1. Table 1 Coding of studies included in meta-analysis.
Note: For measure, B = breadth, D = depth, and O = other. For design, E refers to an experiment, and S refers to a survey. For context, S = social, and T = task. For type, T = text-based and V = video-based. Time is the number of minutes allotted to the CMC experimental condition. Table 1 Coding of studies included in meta-analysis.
Note: For measure, B = breadth, D = depth, and O = other. For design, E refers to an experiment, and S refers to a survey. For context, S = social, and T = task. For type, T = text-based and V = video-based. Time is the number of minutes allotted to the CMC experimental condition. Coding of ModeratorsType of self-disclosure measuredSelf-disclosure was measured using an instrument examining either a) breadth, b) depth, or c) other. Breadth of self-disclosure considers the scope or number of different topics provided to the other person (usually the measure involved an inventory and the number of topics for which self-disclosure occurred). Depth is the level of intimacy or personalness of the information provided during the disclosure and was typically measured using the extent to which participants disclosed on intimate or personal topics. Study designThis variable represents whether the study used either an experimental design or a survey design. An experimental design could be either an independent groups design where each participant interacted with another person in either a CM or FtF setting or a repeated measures design where each participant interacted with another person in both CM and FtF settings. A survey is a questionnaire that asks people about some previous disclosure and compares people based on whether the recall targets either a CM or FtF relationship. Interaction contextExperiments were coded for whether they used the context of a social or task interaction. A social interaction is where interaction partners interacted with no instructions to accomplish a specific task objective (e.g., a getting-to-know-you interaction). A task orientation requires that partners work at achieving an objective (e.g., making a decision). Type of CMCForms of CMC were coded as text-based or video-based. Length of interactionThis was coded for experimental interactions and refers to how many minutes the interaction was allowed to continue in the CMC condition. In all cases, the amount of time used involves a set amount of time for the face-to-face conversation that was doubled for the computer-mediated conversation. Statistical AnalysisThe analysis used a random-effects form of psychometric meta-analysis developed by Schmidt and Hunter (2015). The technique has statistical information converted to a common metric, in this case the correlation coefficient. The effects were coded so that a positive effect indicates greater disclosure in the FtF conditions. Each effect is corrected for various statistical artifacts to permit comparison of studies that are differentially impacted by the artifact. The correction creates a common metric for the analysis and inclusion of each study. The corrections most often used in this investigation were for attenuated measurement and restrictions in range, and the procedures are outlined in Schmidt and Hunter (2015). The generation of an average effect is sample weighted. Along with statistical power estimates for nonsignificant effect sizes (Valentine, Pigott, & Rothstein, 2010), accuracy or confidence interval for each estimate provides a means of assessing the power associated with the particular correlation. This is combined with Rosenthal's (1984) “fail safe N” estimate, which provides a means of consideration for the “file drawer effect” and provides an estimate of how many unpublished or nonincluded studies that are nonsignficant would be needed to make the average correlation reported nonsignficant. The estimate is only provided for average effects that are statistically significant (indicated by a noninclusion of zero in the confidence interval). An analysis of the heterogeneity of the effects was conducted using a chi-square statistic recommended by Hedges and Olkin (1984). A significant chi-square indicates a greater level of variability than expected due to random chance and the probable existence of a moderator variable. A nonsignficant chi-square indicates the existence of homogeneity and the assumption that any variability in observed effects can be explained on the basis of sampling error. ResultsTable 2 summarizes the results, including fail safe N values. The overall analysis demonstrates an effect that FtF self-disclosure was greater than CM self-disclosure, average r = .211, 95% CI [.129, .293], k = 31, based on a heterogeneous set of estimates, χ2 (30, N = 6,216) = 262.62, p < .001. Table 2 Summary of meta-analytic findings.
Note: For the continuous moderator of interaction length, the effect size (r) represents the correlation between studies' effect sizes and the amount of time participants received in minutes. Table 2 Summary of meta-analytic findings.
Note: For the continuous moderator of interaction length, the effect size (r) represents the correlation between studies' effect sizes and the amount of time participants received in minutes. Measure of Self-DisclosureThe average effect for depth indicated greater self-disclosure in FtF communication than in CMC, average r = .276, 95% CI [.224, .328], k = 17, based on a heterogeneous set of estimates, χ2 (16, N = 3,054) = 82.49, p < .001. Breadth of self-disclosure also favored FtF settings, average r = .170, 95% CI [.076, .264], k = 7, based on a heterogeneous set of estimates, χ2 (6, N = 2,154) = 104.29, p < .001. Measures that assessed other forms of self-disclosure or a combination of types of self-disclosure also favored FtF, average r = .102, k = 7, 95% CI [.041, .161], based on a heterogeneous set of estimates, χ2 (6, N = 1,017) = 47.86, p < .001. Study DesignExperimental conditions demonstrated no significant effect, average r = .037, 95% CI [−.108, .180], k = 12, based on a heterogeneous set of estimates, χ2 (11, N = 1,101) = 80.34, p < .001. Given the observed average sample size and number of studies, power to detect medium and small effect sizes (Cohen, 1992) was approximately 1 and .91, respectively. Survey designs demonstrated higher self-disclosure in FtF than CM settings, average r = .249, 95% CI [.199, .303], k = 19, based on a heterogeneous set of estimates, χ2 (18, N = 5.115) = 142.58, p < .001. Interaction ContextInteraction context reflects whether the self-disclosure was within a task or social context. Survey research all employed social or nontask situations, so the overall design issues covered above represent those findings. Experiments that used a social context demonstrated no difference in self-disclosure between FtF and CM conditions, average r = .060, k = 7, 95% CI [−.099, .219], based on a heterogeneous set of estimates, χ2 (6, N = 534) = 43.95, p < .001. Given the observed average sample size and number of studies, power to detect medium and small effect sizes was approximately 1 and .64, respectively. Experiments that used a task demonstrate no difference in self-disclosure between FtF and CM conditions, average r = .015, k = 5, 95% CI [−.111, .141], based on a heterogeneous set of estimates, χ2 (4, N = 567) = 36.51, p < .001. Given the observed average sample size and number of studies, power to detect medium and small effect sizes was approximately 1 and .67, respectively. Type of CMCThe use of text-based forms of CMC generated an average effect favoring the FtF setting for self-disclosure, average r = .219, k = 27, 95% CI [.130, .296], based on a heterogeneous set of estimates, χ2 (26, N = 5,994) = 259.39, p < .001. For investigations using video-based forms of CMC, the average effect also favors the FtF setting for self-disclosure, average r = .132, k = 3, 95% CI [.007, .157], based on a homogeneous set of estimates, χ2 (2, N = 182) = 3.08, p = .21. Length of InteractionThe correlation between studies' effect sizes and their interaction length was negative, r = −.255 (indicating smaller differences between CM and FtF self-disclosure when interactions were longer), but not significant, p = .42. Given the observed average sample size and study size, power to detect medium and small effect sizes was approximately 1 and .91, respectively. DiscussionThe current meta-analysis was conducted to summarize and clarify previous findings regarding CM and FtF self-disclosure. Overall, self-disclosure was higher in FtF communication than in CMC. However, this effect was heterogeneous, suggesting the existence of moderating variables. Of the moderating variables examined, measure of self-disclosure, study design, and type of CMC revealed differences in effects. The difference between FtF and CM self-disclosure was greater for self-disclosure depth than for self-disclosure breadth. On average, survey studies exhibited higher self-disclosure via FtF communication than via CMC, but there was no significant effect for experimental studies. Finally, the difference between FtF and CM self-disclosure was greater when text-based CMC was used than when video-based forms of CMC were used. These findings have implications for CMC theory and research and suggest several directions for future research in this area. Theoretical ImplicationsWe used previous research and mechanisms identified by social presence theory, SIP theory, and hyperpersonal theory (e.g., Bailenson et al., 2006; Hooi & Choi, 2014; Kim & Dindia, 2011; Nguyen et al., 2012; Sproull et al., 1996) to generate predictions regarding differences in CM and FtF self-disclosure and identify potential moderating variables of this difference. The results of the meta-analysis indicate that existing research does not unequivocally align with any one particular theory. However, the findings suggest potentially more support for some aspects of these theories than for others. The findings are also useful because they can highlight areas where research is lacking or where boundary conditions of the theories should be further examined. As such, the findings of this meta-analysis can provide stimulus and direction for future research on these theories. Below, we offer some potential implications of the findings for each of the theories previously discussed, as well as related theories and frameworks. Social presence theorySocial presence theory suggests that low presence likely encourages self-disclosure (Bailenson et al., 2006; Hooi & Choi, 2014; Sproull et al., 1996). Our findings generally contradict this prediction, with self-disclosure generally being higher in FtF communication than in CMC. The finding that differences in CM and FtF self-disclosure are larger for text-based CMC than for video-based CMC is also inconsistent with this proposed effect of social presence, because the lower-cue nature of text-based CMC (relative to video-based CMC) further suppressed self-disclosure. Future research is needed to explicate the role of social presence in self-disclosure and the potential mechanisms by which presence is related to self-disclosure. SIP theorySIP theory predicts that interactants self-disclose more in CMC to compensate for the relative lack of nonverbal cues and that differences between FtF communication and CMC should become smaller over time (Tidwell & Walther, 2002; Walther, 1992; Walther, 2011). The findings of this meta-analysis do not support either of these predictions. However, we did find a moderate but nonsignificant negative correlation between the amount of time allotted in experiments and differences between CM and FtF self-disclosure, suggesting the possibility that these differences are smaller in experiments that allow more interaction time. Such an effect would be consistent with SIP theory. Unfortunately, the small number of experiments that reported interaction time and the wide variation in their findings (reflected in the effect's confidence interval) makes this finding inconclusive. As Walther (2011) argues, more specific tests of SIP theory's propositions regarding time are needed. It is also possible that the temporal mechanisms proposed by SIP theory would be more evident when examining other outcome variables (e.g., attributional confidence; Tidwell & Walther, 2011) or when examining trajectories of self-disclosure rates over time. Hyperpersonal theorySimilar to Kim and Dindia's (2011) meta-analysis, our findings do not support hyperpersonal theory's proposal that greater intimacy in CMC leads to greater self-disclosure (Nguyen et al., 2012; Walther, 2011). The findings that differences between FtF and CM self-disclosure (favoring higher self-disclosure FtF) are greatest for self-disclosure depth (relative to breadth) and when text-based (as opposed to video-based) CMC are used contradict hyperpersonal theory's prediction that CMC relationships can become highly intimate and that text-based interaction facilitates this process. A potential explanation for this lack of support relates to characteristics of the relationships examined in some studies. The available survey studies largely addressed pre-existing relationships, some of which were multimodal (i.e., not exclusively online). Most of the survey studies did not assess the nature of the relationship (e.g., type of relationship, intimacy, or relationship development). The selective self-presentation characteristic of hyperpersonal communication would probably be less likely to occur when partners communicate via both CMC and FtF communication (Walther & Ramirez, 2010). Further, the affordance utilization model hypothesizes that higher levels of relationship development between interactants will reduce the extent to which people capitalize on the self-presentational and conversational affordances described by hyperpersonal theory (Ruppel, 2015b). However, these boundary conditions are currently underexplored (Walther, 2011). Overall, the current findings suggest that greater theoretical and methodological precision are needed when testing the tenets of hyperpersonal theory. Future DirectionsIn addition to the suggestions for theoretical refinement and application outlined above, the findings suggest three primary directions for future research. The first is that, as a whole, more research needs to be conducted on FtF and CM self-disclosure. The relatively small number of studies that exist on this topic and the inconsistency of their findings make it difficult to suggest whether, in what direction, and in what situations FtF and CM self-disclosure will differ. Although statistical power was high at medium effect sizes, it was relatively low at small effect sizes for some tests, and confidence intervals for the nonsignificant analyses were wide. Further, the heterogeneity observed in most of the analyses indicates a need for further research to allow consistent patterns to emerge. Overall, this meta-analysis highlights the need for more research to draw conclusive findings on this topic. Second, as noted above, research comparing CM and FtF self-disclosure has underexplored the extent to which relationships are multimodal and the strength of the ties between relational partners. One likely source of differences is relationship development or intimacy (Ruppel, 2015b; Utz, 2007). Self-disclosure might also serve different functions (e.g., relationship development versus maintenance) in different relationships, leading to variations in self-disclosure. Future research should examine differences in CM and FtF self-disclosure between zero-history and pre-existing relationships and between pre-existing relationships with different tie strength or goals. Examining the frequency and content of FtF and CM self-disclosure in multimodal relationships, and assessing characteristics of those relationships, will be important to clarifying research in this area. Third, although the theories discussed above are relevant to self-disclosure (e.g., Walther, 2011) and have been included in previous reviews comparing CM and FtF self-disclosure (Kim & Dindia, 2011; Nguyen et al., 2012), none of them was explicitly designed to address self-disclosure. The functional model of self-disclosure on social network sites was designed to predict the occurrence and intimacy of social media self-disclosure, which is primarily public as opposed to dyadic (Bazarova & Choi, 2014). A theory of interpersonal CM self-disclosure could incorporate the factors identified in this study to better explain when, how, and why people self-disclose in CMC. This suggestion is similar to Nguyen and colleagues' (2012) suggestion for a comprehensive theory of online communication. However, narrowing the scope to focus only on interpersonal self-disclosure and related processes could yield a more useful framework for understanding this phenomenon. This meta-analysis has identified several factors that should be included in such a theory, such as type of self-disclosure (breadth or depth) and type of CMC (text-based or video-based). As described above, time, multimodality, and relationship type are also potentially influential factors in the extent to which FtF and CM self-disclosure differ. A theory of CM interpersonal self-disclosure would identify these factors and the mechanisms by which they cause differences in FtF and CM self-disclosure. LimitationsAlthough meta-analysis provides the benefit of quantitatively summarizing a body of research, there are three key limitations to the current study. First, studies often omitted relevant information such as the number of utterances in experimental interaction or the strength of the tie between participants and the targets of their self-disclosure. Thus, we were unable to address some potential moderator variables. Second, the design of the study (survey or experimental) was confounded with both the type of relationship examined (pre-existing or zero-history) and whether self-disclosure was measured using self-reported perceptions of behaviors or observation and coding of self-disclosure behaviors. These confounds mean that we could not separate the moderating effects of study design, relationship type, and method of measuring self-disclosure. We chose to follow previous research (e.g., Nguyen et al., 2012) and use study design as the focal variable, but clearly the other variables are also important and should be examined in ways that address these confounds. Further, survey research usually asked participants about their perceptions of self-disclosure in a relationship or overall, as opposed to examining particular interactions. Existing research does not allow us to assess the extent to which this method would constitute a valid test of the theories we used, relative to experimental designs. Third, although the results provide insight into some aspects of key CMC theories and the theories allowed us to identify potentially important moderating variables, the findings are limited in their ability to provide explicit support for particular theories. A primary reason for this limitation is that many of the studies did not test particular theories, and therefore did not include variables (e.g., time, multimodality) that those theories would predict to be important. Support for specific theories can therefore be inferred from our findings, but unequivocal support for particular mechanisms proposed by the theories cannot. For example, the finding that differences in FtF and CM self-disclosure are greater in text-based than in video-based CMC contradicts predictions derived from social presence theory because social presence is lower in text-based CMC. However, because the studies we included did not directly assess social presence, this evidence is indirect. Future research regarding FtF and CM self-disclosure should take a more theory-based approach to continue identifying and validating important factors (e.g., time, social presence, relationship between interactants) in FtF and CM self-disclosure. ConclusionSelf-disclosure is a key construct in CMC theory and research. A full understanding of how and when self-disclosure differs in CMC and FtF communication is therefore crucial. With the exception of experimental settings, the current meta-analysis found higher self-disclosure in FtF communication than in CMC. The results suggest that some previous characterizations of existing research as providing clear support for higher self-disclosure in CMC than in FtF communication are no longer correct. The inconsistencies identified by this meta-analysis point to the need for continued research on CM self-disclosure and the explication of key CMC theories as they relate to self-disclosure. Note1 We are not aware of any studies that compare directed self-disclosure on social networking sites to FtF self-disclosure. References(References marked with an asterisk indicate studies included in the meta-analysis.) Altman ,I. , &Taylor ,D. A. (1973 ). Social penetration: The development of interpersonal relationships . Oxford, England : Holt, Rinehart & Winston . Bailenson ,J. N. ,Yee ,N. ,Merget ,D. , &Schroeder ,R. (2006 ). The effect of behavioral realism and form realism of real-time avatar faces on verbal disclosure, nonverbal disclosure, emotion recognition, and copresence in dyadic interaction . Presence: Teleoperators & Virtual Environments , 15 , 359 – 372 . http://doi.org/10.1162/pres.15.4.359 Bazarova ,N. N. , &Choi ,Y. H. (2014 ). Self-disclosure in social media: Extending the functional approach to disclosure motivations and characteristics on social network sites . Journal of Communication , 64 , 635 – 657 . https://doi.org/10.1111/jcom.12106 * Bruss ,O. E. , &Hill ,J. M. (2010 ). Tell me more: Online versus face-to-face communication and self-disclosure . Psi Chi Journal of Undergraduate Research , 15 , 3 – 7 . * Buote ,V. M. ,Wood ,E. , &Pratt ,M. (2009 ). Exploring similarities and differences between online and offline friendships: The role of attachment style . Computers in Human Behavior , 25 , 560 – 567 . http://doi.org/10.1016/j.chb.2008.12.022 * Chan ,D. K.-S. , &Cheng ,G. H.-L. (2004 ). A comparison of offline and online friendship qualities at different stages of relationship development . Journal of Social and Personal Relationships , 21 , 305 – 320 . http://doi.org/10.1177/0265407504042834 * Chiou ,W.-B. , &Wan ,C-S . (2006 ). Sexual self-disclosure in cyberspace among Taiwanese adolescents: Gender differences and the interplay of cyberspace and real life . CyberPsychology & Behavior , 9 , 46 – 53 . http://doi.org/10.1089/cpb.2006.9.46 * Coleman ,L. H. ,Paternite ,C. E. , &Sherman ,R. C. (1999 ). A reexamination of deindividuation in synchronous computer-mediated communication . Computers in Human Behavior , 15 , 51 – 65 . http://doi.org/10.1016/S0747-5632(98)00032-6 Derlega ,V. J. ,Metts ,S. ,Petronio ,S. , &Margulis ,S. T. (1993 ). Self-disclosure . Newbury Park, CA : SAGE . * Emanuel ,L. ,Neil ,G. J. ,Bevan ,C. ,Fraser ,D. S. ,Stevenage ,S. V. ,Whitty ,M. T. , &Jamison-Powell ,S. (2014 ). Who am I? Representing the self offline and in different online contexts . Computers in Human Behavior , 41 , 146 – 152 . http://doi.org/10.1016/j.chb.2014.09.018 * Embry ,K. E. (2009 ). Friendship and components of intimacy: Role of relationship modality and gender (Doctoral dissertation). Retrieved from ProQuest Dissertations and Theses (304925562). * Green ,S. C. (2006 ). Computer-mediated communication in society: A study on self-disclosure (Doctoral dissertation). Retrieved from ProQuest Dissertations and Theses (305337062). Hedges ,L , &Olkin ,I . (1985 ). Statistical methods for meta-analysis . Orlando, FL : Academic Press . Hiemstra ,G. E. (1983 ). Teleconferencing, concern for face, and organizational culture . Communication Yearbook , 6 , 874 – 904 . Jiang ,L. C. ,Bazarova ,N. N. , &Hancock ,J. T. (2011 ). The disclosure–intimacy link in computer-mediated communication: An attributional extension of the hyperpersonal model . Human Communication Research , 37 , 58 – 77 . http://doi.org/10.1111/j.1468-2958.2010.01393.x Jiang ,L. C. ,Bazarova ,N. N. , &Hancock ,J. T. (2013 ). From perception to behavior: Disclosure reciprocity and the intensification of intimacy in computer-mediated communication . Communication Research , 40 , 125 – 143 . http://doi.org/10.1177/0093650211405313 * Joinson ,A. N. (2001 ). Self-disclosure in computer-mediated communication: The role of self-awareness and visual anonymity . European Journal of Social Psychology , 31 , 177 – 192 . http://doi.org/10.1002/ejsp.36 Joinson ,A. N. (2003 ). Understanding the psychology of internet behaviour: Virtual worlds, real lives . New York, NY : Palgrave Macmillan . Joinson ,A. N. , &Paine ,C. B. (2007 ). Self-disclosure, privacy and the Internet . In A. N. Joinson ,K. Y. A. McKenna ,T. Postmes , &U.-D. Reips (Eds.),Oxford handbook of Internet psychology (pp. 237 – 252 ). Oxford, England : Oxford University Press . Kim ,J. , &Dindia ,K. (2011 ). Online self-disclosure: A review of research . In K. B. Wright &L. M. Webb (Eds.),Computer-mediated communication in personal relationships (pp. 156 – 180 ). New York, NY : Peter Lang Publishing . * Mallen ,M. J. ,Day ,S. X. , &Green ,M. A. (2003 ). Online versus face-to-face conversation: An examination of relational and discourse variables . Psychotherapy: Theory, Research, Practice, Training , 40 , 155 – 163 . http://doi.org/10.1037/0033-3204.40.1-2.155 Nguyen ,M. ,Bin ,Y. S. , &Campbell ,A. (2012 ). Comparing online and offline self-disclosure: A systematic review . CyberPsychology, Behavior & Social Networking , 15 , 103 – 111 . http://doi.org/10.1089/cyber.2011.0277 * Parks ,M. R. , &Roberts ,L. D. (1998 ). “Making MOOsic”: The development of personal relationships on line and a comparison to their off-line counterparts . Journal of Social and Personal Relationships , 15 , 517 – 537 . http://doi.org/10.1177/0265407598154005 * Ponder ,E. (2009 ). Attachment and internet relationships (Doctoral dissertation). Retrieved from ProQuest Dissertations and Theses (305135914). * Ranney ,J. D. , &Troop-Gordon ,W. (2015 ). Problem-focused discussions in digital contexts: The impact of information and communication technologies on conversational processes and experiences . Computers in Human Behavior , 51 , 64 – 74 . http://doi.org/10.1016/j.chb.2015.04.038 * Rogers ,V. L. ,Griffin ,M. Q. ,Wykle ,M. L. , &Fitzpatrick ,J. J. (2009 ). Internet versus face-to-face therapy: Emotional self-disclosure issues for young adults . Issues in Mental Health Nursing , 30 , 596 – 602 . http://doi.org/10.1080/01612840903003520 Rosenthal ,R. (1984 ). Meta-analytic procedures for social research . Beverly Hills, CA : SAGE . * Ruppel ,E. K. (2015a ). Use of communication technologies in romantic relationships: Self-disclosure and the role of relationship development . Journal of Social and Personal Relationships , 32 , 667 – 686 . http://doi.org/10.1177/0265407514541075 * Schiffrin ,H. ,Edelman ,A. ,Falkenstern ,M. , &Stewart ,C. (2010 ). The associations among computer-mediated communication, relationships, and well-being . Cyberpsychology, Behavior, and Social Networking , 13 , 299 – 306 . http://doi.org/10.1089/cyber.2009.0173 Schmidt ,F. L. , &Hunter ,J. E. (2015 ). Methods of meta-analysis: Correcting error and bias in research findings . Thousand Oaks, CA : SAGE . * Schouten ,A. P. ,Valkenburg ,P. M. , &Peter ,J. (2007 ). Precursors and underlying processes of adolescents' online self-disclosure: Developing and testing an “Internet-attribute-perception” model . Media Psychology , 10 , 292 – 315 . http://doi.org/10.1080/15213260701375686 *Sheldon ,P. (2013 ). Voices that cannot be heard: Can shyness explain how we communicate on Facebook versus face-to-face? Computers in Human Behavior , 29 , 1402 – 1407 . http://doi.org/10.1016/j.chb.2013.01.016 Short ,J. ,Williams ,E. , &Christie ,B. (1976 ). The social psychology of telecommunications . London : John Wiley & Sons . Sproull ,L. ,Subramani ,M. ,Kiesler ,S. ,Walker ,J. H. , &Waters ,K. (1996 ). When the interface is a face . Human-Computer Interaction , 11 , 97 – 124 . http://doi.org/10.1207/s15327051hci1102_1 * Stritzke ,W. G. K. ,Nguyen ,A. , &Durkin ,K. (2004 ). Shyness and computer-mediated communication: A self-presentational theory perspective . Media Psychology , 6 , 1 – 22 . http://doi.org/10.1207/s1532785xmep0601_1 *Taddei ,S. ,Contena ,B. , &Grana ,A. (2010 ). Does web communication warm-up relationships? Self-disclosure in computer mediated communication (CMC) . BPA - Applied Psychology Bulletin (Bollettino di Psicologia Applicata) , 260 , 13 – 22 . *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 . http://doi.org/10.1111/j.1468-2958.2002.tb00811.x Valentine ,J. C. ,Pigott ,T. D. , &Rothstein ,H. R. (2010 ). How many studies do you need? A primer on statistical power for meta-analysis . Journal of Educational and Behavioral Statistics , 35 , 215 – 247 . http://doi.org/10.3102/1076998609346961 * Valkenburg ,P. M. ,Sumter ,S. R. , &Peter ,J. (2011 ). Gender differences in online and offline self-disclosure in pre-adolescence and adolescence . British Journal of Developmental Psychology , 29 , 253 – 269 . http://doi.org/10.1348/2044-835X.002001 Walther ,J. B. (2011 ). Theories of computer-mediated communication and interpersonal relations . In M. L. Knapp &J. A. Daly (Eds.),The SAGE handbook of interpersonal communication (4th ed., pp. 443 – 479 ). Thousand Oaks, CA : SAGE . Walther ,J. B. , &Ramirez ,A. (2010 ). New technologies and new directions in online relating . In S. W. Smith &S. R. Wilson (Eds.),New directions in interpersonal communication research . (pp. 264 – 284 ). Thousand Oaks, CA : SAGE . * Wang ,H. , &Andersen ,P. (2007 , May). Computer-mediated communication in relationship maintenance: An examination of self-disclosure in long-distance friendships . Paper presented at the 57th Annual Conference of the International Communication Association , San Francisco, CA . * Weidman ,A. C. ,Fernandez ,K. C. ,Levinson ,C. A. ,Augustine ,A. A. ,Larsen ,R. J. , &Rodebaugh ,T. L. (2012 ). Compensatory internet use among individuals higher in social anxiety and its implications for well-being . Personality and Individual Differences , 53 , 191 – 195 . http://doi.org/10.1016/j.paid.2012.03.003 About the AuthorsErin K. Ruppel (Ph.D., University of Arizona) is an Assistant Professor at the University of Wisconsin-Milwaukee. Her research focuses on how people integrate different modes of communication in interpersonal relationships. Address: Erin K. Ruppel, Department of Communication, University of Wisconsin-Milwaukee, Milwaukee, WI 53201, USA. (e-mail: ). Clare Gross (Ph.D., University of Wisconsin-Milwaukee) is an Assistant Professor at Baldwin Wallace University. Her research examines conflict and relational communication, and the role of communication technologies in these contexts. Arrington Stoll (M.A., University of Wisconsin-Milwaukee) is a Ph.D. student at the University of Wisconsin-Milwaukee and an Adjunct Professor at Carthage College. Her research explores the dark side of communication within close interpersonal relationships. Brittnie S. Peck (M.A., Northern Illinois University) is a Ph.D. student at the University of Wisconsin-Milwaukee. Her research focuses on focuses on social influence in family and interpersonal relationship. Mike Allen (Ph.D, Michigan State University) is a Professor at the University of Wisconsin-Milwaukee. His research focuses on issues of social influence across various contexts (media, interpersonal, organizational, and intercultural). Sang-Yeon Kim (Ph.D., Michigan State University) is an Associate Professor at the University of Wisconsin-Milwaukee. His research agenda includes cross/intercultural communication, persuasion, diversity, and quantitative methods. © 2017 International Communication Association © 2017 International Communication Association What is the hyperpersonal perspective?The hyperpersonal perspective, a special case of social information processing theory, extends the dynamics to explain the circumstances under which communicators may achieve outcomes that exceed those of their face-to-face counterparts.
What are the four elements of hyperpersonal model?The hyperpersonal model addresses how the four components of Berlo's (1960) model of communication—senders, receivers, channel, and feedback—are affected by aspects of communication technology.
How are online relationships different from in person ones?The major difference here is that an internet relationship is sustained via computer or online service, and the individuals in the relationship may or may not ever meet each other in person. Otherwise, the term is quite broad and can include relationships based upon text, video, audio, or even virtual character.
Why you don't engage in an online virtual romantic dating or relationship?Addictive: Virtual relationships can be addictive. People get so used to it that it becomes difficult for them to meet & communicate with people in real life, and they tend to get uncomfortable with physical contact and lose confidence. Sometimes it also becomes impossible for them to handle a real life relationship.
|
