Abstract
This study investigates relational maintenance in parasocial relationships (PSRs) by applying an interpersonal model of friendships (the investment model) and an exploratory mediated model. Undergraduate students (N = 490) reported on their relationships with a close friend and a favorite mediated personality. Despite differences in the strengths of associations, the investment model largely predicted commitment in PSRs through similar processes as it did in friendships. Specifically, greater relational investment and satisfaction predicted relational commitment. Unlike in interpersonal relationships, though, attractiveness of alternatives was unrelated to commitment in PSRs. The study further found that parasocial strength was predicted by identification with and commitment to the character and by the character's integration within a larger social network. The findings extend past applications of interpersonal theories to the media context and support the importance of assessing relational commitment, investment, and network status in PSRs.
Notes
1Comparison of the survey's two versions revealed some differences. For example, participants who answered about a mediated character first rated the friend variables higher. Perhaps participants rated their relationships with mediated characters as fairly positive but felt they needed to rate their friendships as even more positive when completing the section second. Yet the effect sizes of these differences were rather low, with most partial eta-squares at .03 or less, suggesting the significant differences may be at least partially a factor of the sample size. Counterbalancing was used to negate these effects and rerunning the ANOVA analyses while controlling for the survey order did not alter the results.
2Of the designated favorite mediated personalities, 193 were classified as fictional and 270 were classified as “real” personalities. In comparing the two groups on all study variables, only two revealed significant differences. Identification was greater for fictional characters (M = 3.88, SD = 0.89) than for personalities (M = 3.71, SD = 1.01), t = 1.99, p = .047. Investment was greater for fictional characters (M = 2.40, SD = 0.85) than for personalities (M = 2.19, SD = 0.91), t = 2.53, p = .012. Therefore, the two types of mediated personae were collapsed for the analyses.
Note. Correlations for friendships are below the diagonal; correlations for parasocial relationships are italicized above the diagonal.
*p < .05. **p < .001.
Note. All F tests are significant at p < .001. The degrees of freedom are 1 and 489 for all analyses.
3To assess whether models differ, a model in which the parameters (structural paths, covariances, variances) are constrained to be equal across groups (i.e., fully constrained) is compared to a model in which the parameters are allowed to vary for each group (i.e., fully unconstrained). The chi-square values of the constrained and unconstrained models are compared to determine which model yields a better fit. A significant chi-square reduction (relative to the change in degrees of freedom) when parameters are unconstrained suggests model differs for the two groups. If the groups significantly vary, additional models with increasing constraints are conducted to determine where the specific differences exist.
4Full tables with test for all variances, covariances, and structural paths can be obtained from the authors.