Search for a command to run...
Abstract Based on traditional theories of interpersonal relationship development and on the hyperpersonal communication theory, this study examined predictors of relationship satisfaction for individuals involved in online romantic relationships. One hundred-fourteen individuals (N = 114) involved in online romantic relationships, and who had only engaged in computer-mediated communication (CMC) with their partners, completed an online questionnaire about their relationships. Intimacy, trust, and communication satisfaction were found to be the strongest predictors of relationship satisfaction for individuals involved in online romances. Additionally, perceptions of relationship variables differed depending on relationship length and time spent communicating. Implications for interpersonal and hyperpersonal communication theories, and future investigation of online relationships, are discussed. Keywords: Computer-Mediated CommunicationRelationshipOnlineSatisfactionUncertaintyHypersonalInterpersonal Manuscript accepted for publication with minor revisions in Communication Studies, June 2005. This manuscript represents a portion of the first author's dissertation that was directed by the second author. An earlier version of this paper was presented at International Network on Personal Relationships conference in 2001. The authors would like to thank two anonymous reviewers and Jim Query for their helpful comments. Notes ∗n sizes for each respective relationship group are as follows: short = 34, average = 40, long = 40. ∗∗each group significantly different from the other at p < .05. ∗∗∗groups significantly different from one another at p < .05. ∗∗∗∗group significantly different from other groups at p < .05. ∗n sizes for each respective amount of communication group are as follows: low = 33, moderate = 38, high = 43. ∗∗each group significantly different from the other at p < .05. ∗∗∗groups significantly different from one another at p < .05. ∗∗∗∗group significantly different from other groups at p < .05. Using a list of all chat rooms available on AOL that dealt with online friendship and romance, and long-distance relationships (e.g., social support and sexual chat rooms, for example, were not used), every 30th chat room was visited by the researcher for a total of 15 chat rooms. Country of origin break down for participants was as follows: United States (n = 77, 67.5%), Canada (n = 14, 12.3%), Australia (n = 8, 7%), France (n = 3, 2.6%), Germany (n = 2, 1.8%), Italy (n = 1, .9%), the Netherlands (n = 3, 2.6%), New Zealand (n = 1, .9%), and the United Kingdom (n = 3, 2.6%). Initially, we examined frequency of interaction as a predictor variable as well, but the variable suffered from lack of variability. Specifically, the mean, median, and mode were the same (7 on a scale of 1 to 7, with 7 = days a week). As a result, we removed frequency of interaction from the current analyses. However, consistent with Walther's work, we recognize frequency of interaction to be an important variable and one worthy of inclusion in future studies. We used criteria established by Stevens (Citation1996) to test for high multicollinearity. These criteria included the examination of a correlation matrix for any bivariate correlation over .80 and the examination of the predictors' variance inflation factors for any variance inflation factor (VIF) over 10.00, which identified one correlation higher than .80; trust and intimacy were correlated at .842 (p < .001). However, Meyers (Citation1990) argues there is need for concern (and subsequent variable deletion) if a VIF exceeds 10 and, because neither VIF was above ten (trust VIF = 5.89; intimacy VIF = 5.86), both distinct variables were retained. Although beta coefficients indicate a positive relationship among trust, intimacy, and communication satisfaction, due to the prior decision rules established for dealing with multicollinearity, variables were kept as distinct entities and not combined into any composite variables. It should be taken into consideration that Emmers-Sommer (Citation2004) did not collect data in an online forum, whereas the current study involves an online collection (i.e., participants might self-select medium based on preferences). Additional informationNotes on contributorsTraci L. Anderson Traci Anderson (Ph.D., University of Oklahoma) is Assistant Professor at Bryant University. Tara M. Emmers-Sommer Tara Emmers-Sommer (Ph.D., Ohio University) is Associate Professor at the University of Nevada-Las Vegas.