ABSTRACT
People's relationship between positive and negative affect varies on a continuum from relatively independent to bipolar opposites, with stronger bipolar opposition being termed affective bipolarity. Experiencing more depressive symptoms is associated with increased bipolarity, but the processes underlying this relation are not yet understood. Here, we sought to replicate this link, and to examine the role of two potential mediating mechanisms: emotion regulation ability, and trait brooding. Drawing from the Dynamic Model of Affect, we hypothesised that (a) a poor ability to regulate negative emotion, and (b) the tendency to brood over one's depressed feelings would predict stronger affective bipolarity, and mediate the relationship between depressive symptoms and affective bipolarity. To measure affective bipolarity, we calculated within-person affect correlations using two weeks of experience sampling data from a community sample (n = 100). Mediation analyses indicated that baseline assessments of an inability to regulate negative emotions in general, but not brooding specifically, mediated the relation between depressive symptoms and affective bipolarity. These findings highlight an initial mechanism through which depressive symptoms are associated with lower emotional complexity and flexibility.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Egon Dejonckheere http://orcid.org/0000-0002-7219-7506
Elise K. Kalokerinos http://orcid.org/0000-0002-3282-6806
Notes
1 For a full protocol, the data, and a reproducible R-code, see: http://osf.io/mqwyv.
2 Due to ethical constraints, suicidal ideation was not assessed.
3 Item-level internal consistency reflects to what extent the individual items that make up the global composite co-vary in a similar way within the average person across time. The lower internal consistency for NA may be a consequence of people's ability to differentiate between various negative emotional states (e.g. At this moment, I feel stressed, but not sad.; Erbas, Ceulemans, Koval, & Kuppens, Citation2015).
4 Although the nested data structure of our ESM protocol suggests the use of multilevel regression, we use a correlational approach to model the relation between PA and NA within individuals. Our rationale, as well as the equivalent multilevel models (which fully replicate our correlation results) can be found in Supplemental Materials 1.
5 This relation holds after controlling for mean levels of, and variability in, PA and NA, and is robust across various PA and NA item operationalizations (see Supplemental Materials 2 and 3).
6 A model in which we entered both mediators simultaneously replicates the results of the separate mediation models that we describe here (see Supplemental Materials 4 for the full results of this model).
7 Interestingly, the mediating effect of emotion regulation ability is slightly stronger for the CES-D than the BDI-II. The CES-D's greater emphasis on emotional disturbances (versus the BDI-II's primary focus on cognitive distortions; e.g. Fried, Citation2017) may explain this finding.