The effects of conflict resolution styles on perceived relational self-concept change

ABSTRACT

No romantic relationship, no matter how fulfilling and satisfying it may be, is without conflict. We sought to understand how the conflict resolution styles a person uses in their relationship are associated with their perception of their self. Specifically, we examined negative conflict resolution styles and their influence on perceived relational self-concept degradation. In two studies (Ns = 272 and 184), we predicted and found that the more frequently people use negative conflict resolution styles in their relationship, the more strongly they perceive that their relationship has contributed to self-concept degradation. In Study 3 (N = 122), we conducted a longitudinal study and found greater destructive conflict styles at Time 1 predicted greater perceived relational self-concept degradation at Time 2.

KEYWORDS:

• Conflict resolution styles
• self-concept change
• self-contraction

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. All studies in this manuscript received IRB approval from [Blind for Review]. Participants consented to partake in these studies.

2. The current data was collected as part of a different study (Blinded for Review) not related to the current hypotheses of interest. The data were collected in October of 2020.

3. In the study, we identified 312 eligible participants for our study. Participants were removed from the study’s final sample if they indicated they did not want their data to be used (n = 14) or if they reported they took the survey before (n = 25).

4. Additionally, we opted to analyze the self-change processes as composites given how highly correlated the processes are with one another. Specifically, self-contraction and self-adulteration were highly correlated (Study 1: (r(270) = .81, p < .01; 95% CI [.76, .85]); Study 2: (r(182) = .67, p < .01; 95% CI [.58, .74]) with each other. Additionally, self-expansion and self-pruning were highly correlated with each other (Study 1: (r(270) = .62, p < .01; 95% CI [.54, .69]); Study 2: (r(182) = .48, p < .01; 95% CI [.36, .59]). Thus, we opted to create the two composite scores as opposed to testing the four processes individually.

5. The current data was collected as part of a different study not related to the current hypotheses of interest (Caselli & Machia, 2022). The data were collected in March of 2020.

6. In the study, we identified 200 eligible participants for our study. Participants were removed from the study’s final sample if they indicated they did not want their data to be used (n = 9), if they reported they took the survey before (n = 1) or if they were not born in the United States (n = 6; a requirement of the original sampling plan). In Caselli and Machia (2022), we removed additional participants who did provide high-quality data, but unfortunately did not follow the instructions of the manipulation so were not usable for the purpose of that study. However, the manipulation was unrelated to our variables of interest, so we opted to keep these participants in our final sample.

7. In this dataset, the instructions of the scale were altered to read, “Using the scale, rate how frequently you use each of the following styles to deal with arguments or disagreements about politics or the political climate in the United States with your partner.” However, we do not have any theoretical predictions for why the topic of conflict would matter in our proposed process (e.g., Cramer, 2002).

8. The current data were collected as part of a different study (Blinded for Review) not related to the current hypotheses of interest. For a different purpose, there was a manipulation with four conditions early in the survey. All items we report on here were collected later in the survey, well after we expect the effect of the manipulation to have dissipated. Nevertheless, we conducted several analyses of variance (ANOVA) in order to ensure there were no differences on our variables of interest among the conditions. There were no differences across condition for any of our variables, thus we collapsed across condition for our main analyses.

9. The data were collected in February and March of 2021.

10. We aimed to recruit 400 participants at the Time 1 with the expectation of having at least 250 of those participants complete the follow-up at Time 2. This sampling plan follows previous attrition rates when conducting longitudinal studies on Amazon’s Mechanical Turk (Boynton & Richman, 2014; Kim & Hodgins, 2017).

11. We opted to include participants who completed Time 1 and Time 2 measures. At Time 1 there were 168 participants, of those, 161 took day one of the daily survey. After attrition, 122 participants completed the follow-up survey at Time 2. Those who completed Time 2 were included in our analyses.

12. The data for this study are available on OSF: https://osf.io/5rzas/?view_only=7e522d30b4e943c9a4870966802e2864.