Rumination in Interpersonal Relationships: Does Co-rumination Explain Gender Differences in Emotional Distress and Relationship Satisfaction Among College Students?

All analyses were conducted with each of the two subscales separately and yielded the same pattern of results. Therefore, for simplicity’s sake, only the combined results are presented in the paper.

All analyses were conducted with each of the two subscales separately and yielded the same pattern of results. Therefore, for simplicity’s sake, only the combined results are presented in the paper.

Given that the majority of participants reported that they had the closest relationship with their mother, we explored whether there were gender differences in co-rumination with mothers. We did not find evidence of gender differences in co-rumination with mothers specifically t(266) = 1.26, p = .21).

Similar results were found when substituting co-rumination with parents, romantic partners, and roommates in place of co-rumination with friends in the confirmatory factor analyses. Specifically, the model with a freely estimated co-variance between co-rumination and rumination fit the data well, whereas the model constraining co-rumination and rumination to equality did not provide a good fit to the data. Furthermore, the model in which the covariance between co-rumination and rumination was constrained to equality led to a statistically significant decrement in model fit relative to the model with the freely estimated covariance between co-rumination and rumination. Results were similar in all 4 models examining the discriminant validity between rumination and co-rumination with friends, parents, romantic partners, and roommates, respectively.

In light of the recent findings by Treynor and colleagues (2003), suggesting that two distinct dimensions of rumination, brooding and reflection, hold different relationships to depression, we re-ran this mediation analysis twice controlling for each subscale of rumination separately in each mediation analysis. First, we constructed each subscale based on the items used by Treynor and colleagues (2003). Then, one of the rumination subscales was controlled for in the mediation analysis involving gender, co-rumination, and friendship satisfaction by adding paths from gender to the rumination subscale, and from the rumination subscale to both co-rumination and friendship satisfaction. Co-rumination remained a significant mediator of the relationship between gender and friendship satisfaction after controlling for the reflection subscale (beta = .005, 99% CI = .004–1.40) and after controlling for the brooding subscale (beta = .005, 99% CI = .004–1.32).

In light of the recent findings by Treynor and colleagues (2003), suggesting that two distinct dimensions of rumination, brooding and reflection, hold different relationships to depression, we re-ran this mediation analysis twice controlling for each subscale of rumination separately in each mediation analysis. First, we constructed each subscale based on the items used by Treynor and colleagues (2003). Then, one of the rumination subscales was controlled for in the mediation analysis involving gender, co-rumination, and depression by adding paths from gender to the rumination subscale, and from the rumination subscale to both co-rumination and friendship satisfaction. Based on findings by Treynor and colleagues (2003) suggesting that the brooding dimension of rumination is more strongly associated with depression, we hypothesized that co-rumination would remain a significant mediator of the relationship between gender and depression after controlling for the reflection subscale of rumination, but not after controlling for the brooding subscale. Our findings were consistent with these hypotheses in that co-rumination remained a significant mediator of the relationship between gender and depression after controlling for the reflection subscale (beta = .03, 99% CI = .01–1.23), but was not a significant mediator of the relationship between gender and depression after controlling for the brooding subscale (beta = −.002, 99% CI = −.41–.94).