Emotion Regulation Diversity in Current and Remitted Depression

Abstract

Depression is associated with reduced flexibility in emotion regulation (ER). Diversity in the use of ER strategies is crucial for ER flexibility. In this study, we examined associations between depression and ER diversity and proposed a novel measure: the ER diversity index. Currently depressed (n = 58), remitted depressed (n = 65), and healthy control participants (n = 55) rated their use of nine ER strategies. Four ER measures were computed (diversity index, sum score, flexibility score, intraindividual standard deviation), and their association with diagnostic group was compared. The ER diversity index was associated with depression status more strongly than all other ER measures. Currently and remitted depressed individuals exhibited greater diversity in ER strategies overall and maladaptive ER strategies but less diversity in adaptive ER strategies compared with healthy individuals. Thus, the ER diversity index may be a valid measure of ER diversity, and ER diversity may have an important role in depression.

Keywords

depression emotion regulation emotion-regulation diversity emotion-regulation flexibility emotion regulation diversity index

Major depressive disorder (MDD) has been consistently linked with impaired emotion regulation (ER; for review, see Rottenberg, 2017). Theoretical models posit that maladaptive ER that results in dysregulated negative affect and low levels of positive affect is at the core of mood disorders (e.g., Hofmann et al., 2012). A substantial amount of theoretical and empirical work has led to the functional differentiation of adaptive ER strategies as opposed to maladaptive ER strategies in individuals with MDD (for meta-analytic review, see Aldao et al., 2010). Strategies such as cognitive reappraisal and problem-solving are associated with adaptive outcomes, whereas strategies such as rumination and suppression are linked to maladaptive outcomes (e.g., Campbell-Sills et al., 2006; Goldin et al., 2008). More recently, the importance of contextual factors in determining the efficacy and adaptiveness of particular ER strategies has also been emphasized (Aldao, 2013; Hofmann, 2014; Sheppes & Gross, 2011, 2012). Researchers have suggested that the effectiveness of ER depends less on which particular strategy is used and more on the ability to flexibly and effectively use ER strategies appropriate to the context (Aldao et al., 2015; Bonanno & Burton, 2013).

Flexibility in the use of ER strategies has been conceptualized as the synchronization between changes in environmental demands and changes in the use of ER strategies (Aldao et al., 2015). Diversity in the use of ER strategies encompasses the variety, prevalence, and relative abundance of ER strategies and is an essential condition for the flexible regulation of emotions. Low ER diversity might indicate a limited repertoire (e.g., an inability to use a diverse range of ER strategies) or the inflexible use of regulatory strategies (e.g., the tendency to use the same ER strategies across situations). Limited ER repertoire and inflexible use of ER strategies can lead to prolonged negative mood, which may exacerbate depressive symptoms. Thus, low ER diversity may be a vulnerability factor for MDD onset and recurrence.

Several studies have examined the association between depressive symptoms and self-reported diversity in the use of coping strategies. Whereas ER refers to processes by which individuals influence their experience and expression of emotions (Gross, 1998), coping is defined as cognitive and behavioral efforts to manage internal and/or external demands that are appraised as taxing (Lazarus & Folkman, 1984). Thus, ER differs from coping in that the former can involve the regulation of both negative and positive emotions, whereas the latter centers on negative emotions. However, there is substantial overlap between the two constructs: Both are conceptualized as processes of regulation, and both include controlled, purposeful efforts to regulate emotions (for review, see Compas et al., 2014). Thus, although there exist no measures of perceived ER diversity specifically, inferences can be drawn from studies of coping diversity.

Several self-reported measures that assess the perceived diversity in the use of coping strategies have been developed. These measures contain items such as “When stressed, I use several ways to cope and make the situation better” (from the Coping Flexibility Scale; Kato, 2012) and “When confronted with an important problem, I have enough different options to quickly solve a problem” (from the Coping Flexibility Questionnaire; Vriezekolk et al., 2012b). Such items therefore ask participants to appraise their ability to use a variety of strategies to cope effectively. Perceived coping diversity has been linked to lower depressive symptom levels both cross-sectionally and longitudinally (Kato, 2012, 2015; Vriezekolk et al., 2012a). Perceived coping diversity has also been associated with better psychological adjustment (for meta-analytic review, see Cheng et al., 2014), a construct that is conceptually linked to depression.

Although these studies provide initial evidence for a link between coping diversity and depressive symptoms as well as well-being, measures of perceived diversity may be particularly confounded by cognitive biases. Individuals with high levels of depressive symptoms appraise themselves and their abilities in a negative manner, whereas healthy individuals appraise themselves more positively, which may lead to an overestimation of their abilities (for review, see LeMoult & Gotlib, 2019). These differences in self-appraisal may extend to ER and coping ability, which can artificially inflate differences in ER and coping diversity between individuals with elevated depressive symptom levels and healthy individuals. As a result, measures of perceived ER diversity that ask individuals to report their use of various strategies to effectively regulate emotions are affected by self-appraisal biases in addition to recall biases. In contrast, diversity indices computed from measures that assess the frequency of use of a variety of ER strategies, independent of their perceived effectiveness, are less subject to self-appraisal biases. Thus, diversity indices based on the frequency with which different ER strategies are used likely provide more accurate assessments of group differences in ER diversity than the indices based on perceived ER diversity.

Despite the theorized advantage of ER and/or coping diversity indices over measures of perceived diversity, studies that have examined such indices in the context of depression have produced mixed findings. One measure used in prior research to assess diversity is the variance within individuals across self-reported ER or coping strategies, specifically, the intraindividual standard deviation (iSD). Lower levels of depressive symptoms were associated with more diverse ER use, which the authors operationalized as higher iSDs on four coping styles (action-oriented coping, positive reappraisal, avoidance, and social support) across various hypothetical situations (Fresco et al., 2006). A study that used hierarchical cluster analysis to categorize participants’ coping strategies reported that participants who used more emotion- and problem-focused coping strategies across situations exhibited fewer depressive symptoms than participants who used fewer coping strategies (Cheng, 2001). Likewise, in a study that created a summary score by adding the number of ER strategies in which each participant scored above the sample mean, researchers found that higher scores were linked to lower psychological distress (Lam & McBride-Chang, 2007). Other studies, however, failed to report a link between lower depressive symptom levels and more diverse use of ER or coping strategies. For example, participants who were categorized by a hierarchical cluster analysis into more and less diverse groups according to self-reported use of monitoring and blunting across various situations did not differ in depressive symptom levels (Cheng & Cheung, 2005). Likewise, variance in the use of emotion- and problem-focused coping strategies was not associated with the severity of depressive symptoms (Zong et al., 2010).

The equivocal findings on the link between depressive symptoms and ER or coping diversity may be due to the limitations of existing measures. For instance, the iSD (e.g., Fresco et al., 2006) captures variability or evenness in ER or coping strategy use (i.e., differences in the usage across strategies) but does not account for overall frequency of strategy use. Similar frequency of use across strategies will produce a low iSD, and uneven use of strategies will produce a high iSD regardless of the absolute level of frequency. For example, an individual who frequently uses a few strategies and infrequently uses others will have a high iSD. In contrast, an individual who frequently uses, and thus scores high on, various strategies (i.e., high diversity) will have a low iSD, reflecting small differences in the usage of different strategies. However, an individual who infrequently uses, and thus scores low on, these strategies (i.e., low diversity) will also have a low iSD, highlighting the problem with directly interpreting the iSD as a measure of diversity. Thus, iSD is a measure of variability or evenness but cannot be viewed as an index of diversity (i.e., the extent to which different strategies are used frequently). Categorizing individuals using hierarchical cluster analysis could also be problematic. Because this analysis is data driven, different samples may produce clusters with different ER profiles. That is, differences in sample characteristics may render cross-study comparisons invalid, which could account for the mixed findings using this method.

Another index that has been used to assess ER or coping diversity is the “coping flexibility score” (Bonanno et al., 2004, 2011). To compute the coping flexibility score, a sum score is first created by adding scores on subscales that assess the use of two ER strategies. Next, a polarity score is calculated as the absolute value of the discrepancy between two subscale scores. Finally, a coping flexibility score is created as the sum score minus the polarity score. These steps are as follows (using scores on reappraisal and suppression subscales as examples because researchers using the flexibility score have predominantly examined these strategies):

Sum: (Reappraisal + Suppression)

Polarity: | Reappraisal − Suppression |

Flexibility: Sum − Polarity

Using this method, an individual who uses both reappraisal and suppression frequently will have a higher score than an individual who rarely uses either of the strategies, thus overcoming the aforementioned problem with the iSD. Higher flexibility scores, which reflect greater ER diversity, were associated with better long-term psychological adjustment and lower scores on trauma (Bonanno et al., 2004, 2011). However, one drawback of this index is its limited sensitivity. That is, it can differentiate ER profiles that moderately to strongly differ from each other but will produce identical scores for ER profiles that mildly differ from each other. For instance, on subscales that range from 0 to 20, a participant who scored 10 on reappraisal and 5 on rumination and another participant who scored 5 on both reappraisal and rumination will both have a flexibility score of 10. In contrast, a participant who scored 15 on reappraisal and 10 on rumination will have a flexibility score of 20.

To address the limitations of existing measures, we propose a novel method to examine ER diversity: the ER diversity index. The ER diversity index is adapted from the Shannon biodiversity index (H; Shannon, 1948; Washington, 1984), which traditionally is used in ecology as a measure of both the abundance and evenness in the distribution of species in a given area. The Shannon index has previously been used to operationalize diversity in emotional experiences, termed emodiversity and defined as the variety and relative abundance of the emotions experienced (Quoidbach et al., 2014). The emodiversity index has been associated with higher depressive symptom levels and lower inflammatory marker levels more strongly than the mean levels of emotions (Ong et al., 2018; Quoidbach et al., 2014). We propose to quantify ER diversity by both richness (how many different ER strategies are used and how frequently they are used) and evenness (the extent to which specific ER strategies are used in a similar proportion) using the following formula derived from the Shannon index:

ER diversity = - \sum_{i = 1}^{s} p_{i} \times l n (p_{i}),

where S = total number of ER strategies and p_i = proportion of the maximum possible score across all ER strategies made up of the ith ER strategy score.

We present an illustration of the ER diversity index for three individuals whose use of three ER strategies were assessed: reappraisal, rumination, and acceptance. Each strategy is rated on a Likert scale from 1 (infrequent use) to 5 (frequent use), resulting in a maximum total sum score of 15 for the three strategies. Each individual demonstrates a different level of richness and evenness:

\begin{array}{l} Person A : reappraisal = 4, rumination = 4, acceptance = 4 \\ Person B : reappraisal = 1, rumination = 3, acceptance = 5 \\ Person C : reappraisal = 1, rumination = 1, acceptance = 2 \end{array}

Their respective diversity scores are as follows:

\begin{array}{l} H (person A) = - (\begin{array}{l} (4 / 15) \times \ln (4 / 15) + (4 / 15) \times \\ \ln (4 / 15) + (4 / 15) \times \ln (4 / 15) \end{array}) = 1.06 \\ H (person B) = - (\begin{array}{l} (1 / 15) \times \ln (1 / 15) + (3 / 15) \times \\ \ln (3 / 15) + (5 / 15) \times \ln (5 / 15) \end{array}) = 0.87 \\ H (person C) = - (\begin{array}{l} (1 / 15) \times \ln (1 / 15) + (1 / 15) \times \\ \ln (1 / 15) + (2 / 15) \times \ln (2 / 15) \end{array}) = 0.63 \end{array}

As illustrated, an individual who uses all strategies frequently (Person A) has a higher ER diversity index than an individual who uses some strategies frequently and others infrequently (Person B), who in turn has a higher ER diversity index than an individual who uses all strategies infrequently (Person C). Higher ER diversity index scores reflect more frequent and even use of several ER strategies. The ER diversity index differentiates individuals according to both the absolute and relative levels of ER strategy use and thus overcomes the limitations of existing measures used to assess ER diversity.

In sum, there is a growing emphasis on the flexible use of ER strategies as a protective factor against depression. ER diversity is an essential component of flexibility and is thus a relevant and useful construct of study. However, the results from empirical studies of ER diversity and depression have been mixed, which may be due in part to limitations of existing measures. Furthermore, no study to our knowledge has reported the link between ER diversity and depression using clinical samples, which limits inferences that can be made to current and remitted MDD. If ER diversity is reduced in both currently and remitted depressed individuals, low ER diversity may be a stable vulnerability factor for MDD that increases the likelihood of both onset and recurrence. In the present study, we address these limitations by using the ER diversity index to investigate the relations between ER diversity and depression status. Specifically, we evaluated the ER diversity index, existing measures used to assess the diversity in ER strategy use, and a measure of the frequency in ER strategy use in currently depressed, remitted depressed, and healthy control participants. We made the following hypotheses:

Hypothesis 1: The ER diversity index would be more strongly associated with depression status than the iSD and the flexibility score.

Hypothesis 2: ER diversity assessed by the ER diversity index would be more strongly associated with depression status than frequency of ER strategy use alone.

Hypothesis 3: Individuals with current and remitted depression would exhibit less ER diversity than healthy individuals.

Method

Participants

The sample consisted of three participant groups: (a) a currently depressed group (CD), (b) a remitted depressed group (RD), and (c) a healthy control group (HC). Inclusion criteria for all groups were an age between 18 and 65 years inclusive and meeting specific eligibility criteria for one of the three participant groups. Exclusion criteria for all groups were a current or past diagnosis of bipolar disorder or psychotic disorder or current alcohol or substance dependence. Participants were recruited from the community primarily through social media, online classified ads, and recruitment posters placed in public locations.

Diagnosis and eligibility were determined through the administration of the Mini International Neuropsychiatric Interview, Version 6.0.0 (MINI; Sheehan et al., 1998). Participant group definitions were consistent with the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 2000). CD participants met diagnostic criteria for a current major depressive episode (MDE) over the past 2 weeks. RD participants met diagnostic criteria for a past but not current MDE and had a score of less than 20 on the Beck Depression Inventory–II (BDI-II; Beck et al., 1996). HC participants had no current psychiatric diagnoses and no lifetime history of MDD.

Participants were administered the MINI to assess for inclusion and exclusion criteria and establish diagnoses. A total of 227 individuals participated. On the basis of the MINI, 49 participants were excluded for the following reasons: categorized as RD according to the MINI but had a score higher than 20 on the BDI-II (n = 27); did not meet full diagnostic criteria for either a current or past MDE on the MINI (n = 4); met exclusionary diagnoses of psychotic disorder (n = 6), bipolar disorder (n = 4), or alcohol or substance dependence (n = 4); exceeded age limit of the study (n = 1); and displayed cognitive or interpersonal deficits that interfered with completion of the MINI interview or the computer tasks (n = 3). The final sample (N = 178) consisted of 58 CD, 65 RD, and 55 HC participants. In the current study, we used data from a larger study on the cognitive control of emotional information in depression (Quigley et al., 2020). The sample size from the larger study was obtained from power analysis using G*Power3 (Faul et al., 2007) to detect a medium effect size for group differences (Cohen’s d = 0.5).

Measures

Mini International Neuropsychiatric Interview

The MINI Version 6.0 (Sheehan et al., 1998) was administered to participants to determine eligibility and group assignment. The MINI was developed as a brief alternative to more lengthy structured diagnostic interviews, such as the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I; First et al., 2002) and the Composite International Diagnostic Interview (CIDI; World Health Organization, 1990), and demonstrates good correspondence with the SCID-I and CIDI for the majority of diagnoses (Lecrubier et al., 1997; Sheehan et al., 1998).

All MINI interviews were conducted by a senior doctoral student in clinical psychology (L. Quigley). The reliability of the diagnoses based on these interviews was evaluated by having an experienced diagnostician (K. S. Dobson) review and score a sample of 30 audio-recorded interviews (approximately equal numbers across participant groups and selected randomly within each group) without knowledge of the diagnoses provided by the interviewer. There was perfect correspondence for the diagnoses of current and past MDD (K = 1.00).

Beck Depression Inventory-II

The BDI-II (Beck et al., 1996) is a 21-item self-report inventory that assesses the presence and severity of depressive symptoms over the past 2 weeks. Items are rated on a 4-point scale; total scores ranged from 0 to 63. Total scores are categorized as minimal depression (0–13), mild depression (14–19), moderate depression (20–28), and severe depression (29–63; Beck et al., 1996). The internal consistency for the BDI-II in the current sample was α = .97.

Cognitive Emotion Regulation Questionnaire

The CERQ (Garnefski et al., 2001) is a 36-item questionnaire that measures the use of cognitive ER strategies in response to stressful events. The CERQ consists of nine four-item subscales, five of which are ostensibly adaptive (acceptance, putting into perspective, positive refocusing, positive reappraisal, refocus on planning) and four of which are ostensibly maladaptive (rumination, catastrophizing, self-blame, other-blame). Items are rated on a 5-point scale. Higher scores reflect more frequent use of a specific strategy. Adequate psychometric properties have been demonstrated for the CERQ subscales (Garnefski et al., 2002, 2009; Garnefski & Kraaij, 2006). In the current sample, the CERQ subscales demonstrated good internal consistency (see Table 2).

Procedure

Participants were first screened by phone for inclusion and exclusion criteria. Eligible participants provided written informed consent, completed the MINI diagnostic interview in person, and answered questions about therapy and medication history. Participants then completed a battery of self-report questionnaires in random order, including the BDI-II and the CERQ, and a demographic questionnaire created for the purpose of this study, which included information on age, gender, ethnic background, and level of education. All procedures were approved by the local Institutional Review Board.

Analytic approach

To compute the iSD score, we calculated the within-persons standard deviation across the nine CERQ subscale scores for each participant. High scores on the iSD reflect uneven use of ER strategies and/or infrequent use of some ER strategies. Because in the current study we focused on ER rather than coping, we refer to the coping flexibility score as the ER flexibility score. Traditionally, the flexibility score has been computed using only two strategies (Bonanno et al., 2004, 2011). To compute the ER flexibility score using all nine CERQ subscales, we computed separate scores for all possible pairs of the nine strategies by taking the sum of the scores of the two subscales and then subtracting the absolute value of the difference between the two subscales scores (i.e., sum score − polarity score). The flexibility scores for each subscale pair were then averaged to produce an overall ER flexibility score. High ER flexibility scores reflect more frequent and/or greater evenness in the use of ER strategies. The ER diversity index was computed by dividing each subscale score by the maximum total sum score across the nine CERQ subscales, multiplying each proportion by its natural log, taking the sum of the products for each subscale, and multiplying by −1. High scores on the ER diversity index reflect higher frequency and evenness (i.e., greater diversity) in the use of ER strategies.

Finally, sum scores were computed from CERQ subscales, which reflect frequency of strategy use. In accordance with the multidimensional nature of the CERQ (Garnefski et al., 2001) and with previous studies that reported ER using the CERQ (e.g., Krkovic et al., 2018), sum scores were computed separately for adaptive and maladaptive strategies. Higher sum scores indicate more frequent use of adaptive and maladaptive ER strategies, respectively. For the purpose of conducting parallel model comparisons with the adaptive and maladaptive sum scores, separate ER diversity indices were also computed from adaptive and maladaptive ER strategies. Thus, to test Hypothesis 1, a single score of the ER diversity index, iSD, and ER flexibility score was used; to test Hypotheses 2 and 3, adaptive and maladaptive composite scores of the ER diversity index and the sum scores were used. All scores (i.e., ER diversity index, iSD, ER flexibility score, and ER sum scores) were then transformed into z scores to obtain standardized regression coefficients for ease of interpretation.

Three separate multinomial logistic regression analyses were conducted with the ER diversity index, the iSD, and the ER flexibility score as the predictor and diagnostic group (CD, RD, and HC) as the outcome. Parallel analyses were conducted using the traditional ER frequency measures (adaptive and maladaptive ER sum scores) for comparison. The HC group was the reference group in the analyses because hypotheses focused on comparing ER diversity in the CD and RD groups to the HC group. Fit indices (Akaike information criterion [AIC] and Bayesian information criterion [BIC]) for the regression models were computed. AIC and BIC were chosen because they provide model fit for nonnested models with categorical outcome variables (e.g., Lumley & Scott, 2015; Vrieze, 2012). Lower AIC and BIC values indicate superior model fit, and a ΔAIC of 4 or more between models is interpreted as a significant difference in fit to the data (Burnham & Anderson, 2002). Note that larger values of Nalgelkerke pseudo-R²s in multinomial logistic regression models also reflect better model fit, and these values provide broad indicators of model fit; however, they are not recommended for model comparisons because they are not directly comparable with ordinary least squared R²s (Long & Freese, 2006). To test Hypothesis 1, we compared fit indices of the ER diversity index, iSD, and ER flexibility score models. To test Hypothesis 2, we compared fit indices of the adaptive and maladaptive ER diversity index models with the fit indices of the adaptive and maladaptive ER sum score models, respectively. Finally, to test Hypothesis 3, the parameter estimates for the association of the ER diversity index, the iSD, and the ER flexibility score with depression status from the separate models were interpreted.

Results

Sample characteristics

Demographic characteristics across groups are presented in Table 1. One-way analyses of variance or χ² tests, as appropriate, indicated that the participant groups did not differ significantly on age, gender, ethnicity, or education, all ps > .10. As expected, BDI scores differed significantly across groups, F(2, 175) = 294.10, p < .001, η_p² = .771; the CD group scored the highest, followed by the RD group, followed by the HC group.

Table 1.

Demographic Information Across Participant Groups and of the Full Sample

Variable	Currently depressed participants	Remitted depressed participants	Healthy control participants	Full sample
Age^a	M = 37.50 (SD = 13.86)	M = 34.17 (SD = 14.53)	M = 32.93 (SD = 14.47)	M = 34.88 (SD = 14.34)
Gender^b
Female	50 (86.2)	58 (89.2)	44 (80.0)	152 (85.4)
Male	8 (13.8)	7 (10.8)	11 (20.0)	26 (14.6)
Ethnicity^c
White	41 (70.7)	49 (75.4)	33 (60.0)	123 (69.1)
Asian	3 (5.2)	4 (6.2)	9 (16.4)	16 (9.0)
South Asian	1 (1.7)	7 (10.8)	6 (10.9)	14 (7.9)
Latin American	3 (5.2)	0	3 (5.5)	6 (3.4)
West Indian	3 (5.2)	2 (3.1)	0	5 (2.8)
Filipino	3 (5.2)	1 (1.5)	0	4 (2.2)
First Nations	1 (1.7)	0	2 (3.6)	3 (1.7)
Black	1 (1.7)	0	1 (1.8)	2 (1.1)
Other	2 (3.4)	2 (3.1)	1 (1.8)	5 (2.8)
Education^d
High school or less	9 (15.5)	9 (13.8)	11 (20.0)	29 (16.3)
Some community/technical college or university	20 (34.5)	24 (36.9)	14 (25.5)	58 (32.6)
Completed community/technical college or bachelor’s degree	25 (43.1)	26 (40)	20 (36.4)	71 (39.8)
Master’s or professional degree	4 (6.9)	6 (9.2)	10 (18.2)	20 (11.2)
Symptom measures
BDI-II score^e	M = 32.04_a (SD = 9.04)	M = 8.94_b (SD = 6.69)	M = 2.56_c (SD = 3.62)	M = 14.49 (SD = 14.24)

Note: Values are ns with percentages unless otherwise specified. Values denoted by different subscript letters within the same row are significantly different at p < .05. BDI-II = Beck Depression Inventory–II (Beck et al., 1996).

F(2, 174) = 1.56. ^bχ²(2, N = 178) = 2.08. ^cχ²(16, N = 178) = 23.57. ^dχ²(12, N = 178) = 12.71. ^eF(2, 175) = 294.10 (p < .001).

Correlational analyses

To examine the relations of the ER diversity index, the iSD, and the ER flexibility score with the CERQ subscale scores, Pearson correlation coefficients were calculated with Bonferroni corrections (see Table 2). As expected, most of the adaptive ER subscale scores were positively correlated with each other, as were most of the four maladaptive ER subscale scores. In addition, most of the adaptive subscale scores were negatively correlated with most of the maladaptive subscale scores. The only exceptions to these correlational patterns were acceptance, which was uncorrelated with other adaptive strategies and positively correlated with maladaptive subscale scores, and blaming others, which was uncorrelated with all of the adaptive subscale scores and two of four maladaptive subscale scores. The ER diversity index was positively correlated with the ER flexibility score and negatively correlated with the iSD. The ER flexibility score and the iSD were not significantly correlated. In general, scores on these three measures were moderately correlated with the adaptive and maladaptive sum scores and positively correlated with the CERQ subscale scores.

Table 2.

Bivariate Correlations Between Depressive Symptoms, CERQ Subscale Scores, and ER Measures Computed From the CERQ Subscales

Scale	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. BDI-II	.97
2. Acceptance	.260**	.72
3. Positive refocusing	−.195*	.131	.81
4. Refocusing on planning	−.426**	.133	.485**	.82
5. Positive reappraisal	−.511**	.080	.418**	.776**	.88
6. Putting into perspective	−.355**	.119	.345**	.600**	.636**	.82
7. Self-blame	.580**	.357**	−.154	−.251**	−.372**	−.253**	.88
8. Rumination	.547**	.380**	−.066	−.164*	−.235**	−.189*	.635**	.78
9. Catastrophizing	.650**	.259**	−.214**	−.401**	−.454**	−.403**	.561**	.615**	.82
10. Blaming others	.186*	.135	.067	.013	−.084	.003	.057	.167*	.377**	.80
11. ER flexibility score	.188*	.545**	.447**	.375**	.295**	.350**	.371**	.510**	.389**	.483**	—
12. iSD	.019	.255**	.114	.445**	.411**	.313**	.160*	.158*	−.070	−.191*	−.028	—
13. ER DI	.121	.192*	.243**	−.011	−.051	.059	.163*	.308**	.286**	.449**	.697**	−.605**	—
14. ER DI adaptive	−.633**	−.042	.548**	.632**	.693**	.621**	−.630**	−.567**	−.739**	−.217**	.081	−.02	.071	—
15. ER DI maladaptive	.614**	.147	−.332**	−.550**	−.625**	−.500**	.635**	.664**	.800**	.445**	.331**	−.331**	.513**	−.819**	—
16. Sum adaptive	−.377**	.372**	.644**	.864**	.855**	.784**	−.219**	−.103	−.370**	.025	.544**	.443**	.101	.710**	−.552**	—
17. Sum maladaptive	.678**	.389**	−.139	−.287**	−.398**	−.300**	.804**	.843**	.861**	.445**	.567**	.047	.374**	−.741**	.853**	−.240**

Note: Values in boldface type in the diagonal are Cronbach’s α coefficients. BDI-II = Beck Depression Inventory–II (Beck et al., 1996); CERQ = Cognitive Emotion Regulation Questionnaire (Garnefski et al., 2001); ER = emotion regulation; iSD = intraindividual standard deviation; DI = diversity index.

p < .05. **p < .01.

Multinomial regression models with depression status as the outcome variable

Mean ER diversity index, iSD, and ER flexibility scores across the participant groups are presented in Table 3. All three measures were significant predictors of depression status in their respective models: the ER diversity index (p < .001), the iSD (p = .002), and the ER flexibility score (p = .034) models (see Table 4).

Table 3.

Descriptive Statistics of the CERQ Subscales and ER Measures Across Participant Groups

Statistic	CD(n = 58)	RD(n = 65)	HC(n = 55)
Acceptance	8.51 (3.45)	7.45 (3.18)	6.65 (2.90)
Positive refocusing	3.81 (3.06)	4.20 (2.90)	5.82 (3.23)
Refocusing on planning	6.11 (3.07)	7.29 (3.74)	10.51 (3.46)
Positive reappraisal	5.91 (3.53)	7.47 (3.80)	12.08 (3.41)
Putting into perspective	5.77 (3.05)	7.22 (3.48)	9.43 (4.24)
Self-blame	9.23 (4.18)	6.84 (3.85)	3.51 (2.28)
Rumination	10.57 (3.31)	8.09 (3.75)	4.88 (2.94)
Catastrophizing	7.36 (3.90)	4.91 (3.30)	1.41 (1.68)
Blaming others	3.91 (3.03)	3.82 (2.76)	2.90 (2.06)
ER flexibility score	9.29 (3.05)	8.80 (3.00)	7.79 (3.02)
iSD	3.71 (1.27)	3.35 (1.06)	4.23 (1.45)
ER diversity index	2.03 (0.13)	2.04 (0.10)	1.94 (0.16)
ER diversity index—adaptive	1.06 (0.21)	1.18 (0.17)	1.37 (0.10)
ER diversity index—maladaptive	0.96 (0.16)	0.86 (0.16)	0.56 (0.19)
Sum score—adaptive	30.11 (11.62)	33.64 (12.00)	44.49 (12.69)
Sum score—maladaptive	31.06 (9.62)	23.65 (9.25)	12.71 (6.72)

Note: Values in parentheses are standard deviations. CERQ = Cognitive Emotion Regulation Questionnaire; ER = emotion regulation; CD = currently depressed participants; RD = remitted depressed participants; HC = healthy control participants; iSD = intraindividual standard deviation.

Table 4.

Multinomial Logistic Regression Models With Depression Status as the Outcome Variable

Variable	Model information					Parameter estimates
Variable	χ²	p	R ²	AIC	BIC	β	SE	p	Exp(β)
ER flexibility score	6.75	.034	.047	350.43	362.74
CD						0.52	0.21	.013	1.69
RD						0.36	0.21	.083	1.43
iSD	12.45	.002	.085	344.75	357.03
CD						−0.42	0.21	.043	0.66
RD						−0.72	0.22	.001	0.49
ER DI, total	16.76	< .001	.11	340.45	352.73
CD						0.66	0.22	.003	1.93
RD						0.77	0.23	.001	2.16
ER DI, adaptive	91.65	< .001	.49	265.56	277.83
CD						−3.49	0.56	< .001	0.030
RD						−2.79	0.53	< .001	0.061
ER DI, maladaptive	100.78	< .001	.53	256.43	268.71
CD						3.13	0.47	< .001	22.86
RD						2.12	0.40	< .001	8.31
Adaptive sum	34.91	< .001	.22	322.29	334.57
CD						−1.29	0.26	< .001	0.28
RD						−0.95	0.24	< .001	0.39
Maladaptive sum	90.77	< .001	.49	266.44	278.72
CD						2.97	0.46	< .001	19.54
RD						2.02	0.40	< .001	7.56

Note: Reference group is healthy control. For all χ² values, df = 2; for all parameter estimates, df = 1. R² = Nalgelkerke pseudo-R²; β = standardized regression coefficient; iSD = intraindividual standard deviation; CD = currently depressed; RD = remitted depressed; ER = emotion regulation; DI = diversity index; AIC = Akaike information criterion; BIC = Bayesian information criterion.

Hypothesis 1: comparison of the ER diversity index with the iSD and the ER flexibility score

The ER diversity index model had lower AIC and BIC values than both the iSD and the ER flexibility score models (Table 4). The difference in AIC (ΔAIC) of the ER diversity index model compared with the iSD model and the ER flexibility model was 4.3 and 10, respectively. Thus, as hypothesized, the ER diversity index was more strongly associated with depression status than the iSD and the ER flexibility score.

Hypothesis 2: comparison of the ER diversity indices and ER sum scores

The regression models were significant for both the adaptive and maladaptive ER diversity indices, ps < .001, as well as for both the adaptive and maladaptive sum scores, ps < .001 (see Table 4). For adaptive strategies, the ER diversity index model had considerably smaller AIC and BIC values (ΔAIC = 56.7) than the sum score model.¹ The AIC and BIC values for the maladaptive ER diversity index model were also smaller than those of the maladaptive sum score model (ΔAIC = 10); this difference was smaller than that for the adaptive models but was still sizeable. Thus, the ER diversity indices were more strongly associated with depression status than the ER sum scores, particularly for adaptive strategies.

Hypothesis 3: ER diversity index, iSD, and ER flexibility scores across diagnostic groups

Higher ER diversity index scores were significantly associated with membership in the CD group, β = 0.66, SE = 0.22, p = .003, and the RD group, β = 0.77, SE = 0.23, p = .001, compared with the HC group.² Higher iSD scores were significantly associated with membership in the HC group compared with both the RD group, β = 0.72, SE = 0.22, p = .001, and the CD group, β = 0.42, SE = 0.21, p = .043. Finally, higher ER flexibility scores were significantly associated with membership in the CD group, β = 0.52, SE = 0.21, p = .013, but were not significantly associated with membership in the RD group, β = 0.36, SE = 0.21, p = .083, relative to the HC group. In general, the CD and RD groups were associated with greater diversity in their use of ER strategies than the HC group, as reflected by higher ER diversity index and ER flexibility scores and lower iSDs, although we note that the latter indicates primarily greater evenness in strategy use.

Analyses were also conducted with separate ER diversity indices for the adaptive and maladaptive subscales (see Table 4). Higher adaptive ER diversity index scores were significantly associated with membership in the HC group relative to the CD and RD groups, both ps < .001.³ In contrast, higher maladaptive ER diversity index scores were significantly associated with membership in the CD and RD groups compared with the HC group, both ps < .001. Thus, CD and RD participants were more diverse in their use of maladaptive strategies than HC participants but less diverse in their use of adaptive strategies (see Note 2).

Discussion

In the present study, we investigated the diversity in the use of ER strategies in individuals with current and remitted MDD compared with healthy individuals. Consistent with our hypotheses, the ER diversity index was more strongly associated with depression status than the iSD and the ER flexibility score. Furthermore, ER diversity as measured by the ER diversity index was more strongly associated with depression status than the traditional sum scores reflecting the frequency of ER strategy use, particularly for adaptive ER strategies. Finally, compared with the healthy control group, the current and remitted MDD groups exhibited greater diversity in their use of overall and maladaptive strategies but were less diverse in their use of adaptive ER strategies.

The model with the ER diversity index demonstrated a better fit to the data than the iSD and the ER flexibility score models. These findings were as expected because the ER diversity index addresses limitations inherent to the other two measures. Specifically, the ER diversity index considers both richness (the number of ER strategies used and the extent to which each ER strategy is used) and evenness in ER strategy use (whether ER strategies are used in a similar proportion) such that more frequent use of different ER strategies and greater evenness in the use of the strategies result in higher scores. In contrast, the iSD measures the evenness of ER strategy use but provides limited information on the absolute frequency of ER strategy use. The ER diversity index is also able to discriminate small differences in the frequency among different ER strategies, unlike the ER flexibility score for which different ER profiles can result in the same score. Thus, the ER diversity index may be preferred to existing measures in assessing ER diversity. Given that this is the first study to use the ER diversity index, additional research is needed to corroborate the current findings.

In general, the ER diversity index, the iSD, and the ER flexibility score were moderately correlated, which suggests that these indices reflect related but not identical constructs. As expected, the ER flexibility score and the ER diversity index were positively correlated because consistent and frequent use of ER strategies produce higher scores on both measures. The iSD was negatively correlated with the ER diversity index, which is also expected given that frequent use of some strategies and infrequent use of the other strategies (i.e., unevenness) produces higher iSD scores but lower ER diversity index scores. The lack of correlation between the iSD and the ER flexibility score might be due to the noted limitations of both measures that confound their interpretation, including the low sensitivity of the ER flexibility score and the lack of information on the frequency of ER strategy use in the iSD. For instance, low iSD scores can reflect frequent use of all ER strategies or infrequent use of all strategies; whereas the former is associated with higher ER flexibility scores, the latter is associated with lower ER flexibility scores.

Our results also suggest that the diversity in ER strategy use is more strongly associated with both current and remitted MDD than the frequency of strategy use alone, particularly for adaptive strategies. The model with the ER diversity index computed from adaptive strategies showed considerably better fit to the data than the model using the adaptive sum score. Diversity in the use of adaptive strategies may reflect a larger repertoire of effective ER strategies that is used more flexibly. Specifically, individuals with higher adaptive ER diversity index scores exhibit more frequent and equitable use of many adaptive strategies. Both the diversity index and the sum score encompass the frequency of strategy use, and indeed, the ER diversity indices were correlated with their respective sum scores. However, the ER diversity index provides a more comprehensive profile of ER that considers both richness and evenness of ER strategy use, in contrast to traditional measures that assess only the frequency of ER strategy use. The fact that the ER diversity index contains more information about the patterns of ER likely contributed to its stronger association with depression status than the frequency of ER strategy use, particularly for adaptive strategies.

The model with the maladaptive ER diversity index also demonstrated better fit to the data than the maladaptive sum score model, but this advantage in model fit was less pronounced compared with adaptive ER. Although diversity in the use of ER strategies can provide substantially more information about depression-related ER patterns than the frequency of strategy use, diversity in maladaptive strategies may be less relevant to outcomes than diversity in adaptive strategies. For instance, individuals with high adaptive ER diversity index scores may differentially use positive reappraisal and refocusing on planning in situations in which each strategy is the most effective. In this regard, individuals’ ability to flexibly use different adaptive ER strategies in varying contexts may be a protective factor against MDD onset and recurrence. In contrast, an individual who uses self-blame in one situation and rumination in another may have comparable outcomes with an individual who uses self-blame in both situations in terms of MDD vulnerability because self-blame and rumination may be ineffective in both scenarios. Thus, although ER diversity may be more strongly associated with depression vulnerability than the frequency of ER strategy use, this difference may be more important for adaptive than maladaptive strategies.

Individuals with current and remitted MDD had greater diversity in ER overall, which was consistent across the ER diversity index, the ER flexibility score, and the iSD, although the latter predominantly assesses the evenness component of ER diversity. These results indicate that currently and remitted depressed individuals may have more diverse but less effective use of ER strategies. Diverse but ineffective ER may result from inadequate consideration of contextual information in ER strategy selection, lack of knowledge about which ER strategy may be appropriate for a particular situation, and/or poor implementation of the strategy in regulating negative emotions. Ineffective ER attempts can result in prolonged negative mood, which may prompt the individuals to try different ER strategies, resulting in high ER diversity. Because ER diversity alone cannot indicate whether an individual is using ER strategies effectively, information on context sensitivity and outcome adaptiveness (e.g., Bonanno & Burton, 2013) is needed to further delineate the link between ER diversity and depression.

The association between greater overall ER diversity and current and remitted MDD in the current study may also be explained by the increased experience of stressors in depression. The stress-generation hypothesis posits that depressive symptoms actively contribute to frequent stressful experiences (Hammen, 1991, 2006). Cognitive theories also propose that depression is associated with a negative bias and the absence of a positive bias in information processing (Gotlib & Joormann, 2010). It is thus possible that individuals with current and remitted MDD have greater diversity in overall ER because of their increased likelihood of experiencing stressful life events and/or of perceiving life events as stressful. For instance, when faced with three different life events, healthy individuals may find one of these events stressful, whereas those with current and remitted MDD may find all three to be stressful. Thus, healthy individuals may regulate their emotions for only one event using one strategy, whereas individuals with current and remitted MDD may regulate their emotions for all three events using different strategies. Thus, both reduced effectiveness of ER attempts and greater actual and/or perceived need to regulate negative emotion may contribute to higher ER diversity among individuals with current and remitted MDD relative to healthy individuals.

Our findings on the association between depression status and ER diversity are consistent with a previous study that examined daily ER strategy use and well-being (Grommisch et al., 2020). Individuals were categorized into various classes of ER profiles on the basis of their use of ER strategies over time. Well-being was highest for individuals classified as diverse users of adaptive ER strategies (e.g., acceptance) over time and lowest for individuals classified as diverse users of maladaptive strategies (e.g., suppression) over time. Furthermore, the diverse adaptive ER class had higher pleasant affect, lower unpleasant affect, lower anxiety, and lower stress than the diverse maladaptive class. Although this study was elegantly designed and analyzed via multilevel latent profiles of experiential sampling data, the analytic procedures were complex and involved numerous steps. In contrast, the ER diversity index captures individual differences in ER repertoire using one index that is readily computed using the formula presented and thus may have broader accessibility and utility for researchers. Our findings also parallel the findings of a study that examined the diversity in emotional experiences using the emodiversity index (Quoidbach et al., 2014), in which higher positive emodiversity and lower negative emodiversity were associated with fewer depressive symptoms (Urban-Wojcik et al., 2020).

On the other hand, our findings differ from some other studies that reported lower ER and coping diversity to be associated with greater symptoms and worse well-being (Bonanno et al., 2004, 2011; Cheng, 2001; Lam & McBride-Chang, 2007). Previous research on diversity in ER or coping strategy use assessed diversity across only a few strategies (e.g., Bonanno et al., 2004, 2011; Cheng & Cheung, 2005; Zong et al., 2010), whereas the ER diversity in the current study was based on nine ER strategies. Thus, the current study may have more comprehensively captured the attempts to regulate emotions using different strategies, which could account for the discrepant findings on the diversity-depression relation in comparison with that in previous studies. Finally, existing studies computed measures used to assess ER diversity on the basis of mostly adaptive strategies (e.g., Fresco et al., 2006; Lam & McBride-Chang, 2007), which may account for the negative associations between depressive symptoms and scores on these measures. The proportion of putatively adaptive and maladaptive strategies involved in the computation of ER diversity measures likely influences the patterns of associations with psychological outcomes, including depression. Future research is needed to investigate these possible explanations.

Finally, we observed the novel finding that when computed separately for adaptive and maladaptive strategies, ER diversity demonstrates opposite associations with depression status. Specifically, current and remitted MDD participants had more diverse maladaptive ER but less diverse adaptive ER than healthy control participants. These findings parallel findings in the literature on ER frequency in MDD given that MDD is characterized by more frequent use of maladaptive ER strategies and less frequent use of adaptive ER strategies (Aldao et al., 2010) and suggest that diversity in ER may reflect different phenomena for adaptive and maladaptive strategies. High diversity in the use of putatively adaptive strategies may reflect frequent use of various ER strategies that are consistently effective across stressful situations. Use of a large repertoire of effective ER strategies would facilitate mood regulation and lower risk of MDD. In contrast, diversity in the use of maladaptive ER strategies may reflect frequent use of various ER strategies that are consistently ineffective in response to stressors and thus maintain and exacerbate negative mood, thereby increasing the risk of MDD. The differential association of ER diversity with depression status depending on the putative adaptiveness of the ER strategies suggests that it is important to examine the adaptive and maladaptive ER diversity indices separately in future research on ER diversity and depression or other psychological conditions.

Study strengths and limitations

Several methodological features strengthen the conclusions drawn from the current study. First, in this study, we used clinical and control samples recruited from the community and diagnosed via the administration of a structured clinical interview, which enhances the generalizability and clinical relevance of the findings. Another strength of the study was the use of a well-validated measure of ER strategies and the assessment of substantially more ER strategies than previous research on ER diversity. Finally, the study had adequate sample sizes in the diagnostic groups for statistical analyses.

Several limitations should be noted. First, even with the inclusion of individuals with remitted MDD, the cross-sectional design of the study precludes conclusions about directionality in the link between MDD and ER diversity. Future research with prospective designs is needed to establish the temporal relations and test whether ER diversity predicts the onset and/or relapse of MDD. Second, the current study assessed for ER use tendency in general rather than across specific scenarios. Future studies that assess situational demand (e.g., various scenarios for ER) and regulatory effectiveness (e.g., hedonic changes, goal achievement) would allow for a more comprehensive investigation of ER flexibility to corroborate the present findings (Aldao, 2013). Third, to increase the generalizability and ecological validity of the present findings, as well as to better assess and differentiate between profiles of ER, future studies should consider using experiential sampling methods to assess daily variations in ER strategy use.

Given that this study was the first study to conceptualize and use the ER diversity index as a measure of ER diversity, we note several considerations regarding its use. First, the number of species in the biodiversity index is intended to have no upper bound, whereas the number of ER strategies in the ER diversity index is limited by the number of strategies assessed. Thus, the ER diversity index may not completely capture the richness of one’s ER strategy use. We used the CERQ to compute the ER diversity index, which assesses a relatively large number of ER strategies (i.e., nine strategies). Nonetheless, the validity of the ER diversity index will be limited by the number of ER strategies assessed and may underestimate ER diversity, especially if only a few strategies are assessed.

Second, whereas the input values in the biodiversity index formula reflect discrete counts of members of a species, the input values in the ER diversity index formula reflect the frequency in the use of an ER strategy measured on a Likert-type scale, with a more limited range. This may reduce the ability of the ER diversity index to capture individual differences. Still, the ER diversity index had a stronger association with depression status and was better able to capture individual differences in ER than existing ER measures in the current study. The emodiversity index (Quoidbach et al., 2018) is similarly limited in range given that frequency of emotions as measured by Likert-type scales are used in its computation. Quoidbach et al. (2018) conducted simulations using emotion frequency ratings with different response option ranges and found that responses with larger and smaller ranges gave similar weight to richness and evenness, indicating that the range of responses had minimal impact on the emodiversity index scores.

Finally, we note that in contrast to the emodiversity index, the denominator in the ER diversity index reflects the maximum score possible across all ER strategies rather than the individual’s total sum score. Consequently, the ER diversity index can distinguish individuals who use the same set of ER strategies frequently compared with infrequently (unlike the iSD), thereby better capturing individual differences. Future research is needed to replicate the findings in the current investigation and to further evaluate the ER diversity index as a measure of ER diversity.

Future directions and clinical implications

Apart from the research directions already discussed, we propose three important avenues for future studies on ER diversity and depression. First, additional research is needed to examine ER diversity in the context of MDD remission and recurrence. Research on MDD recurrence has focused on studying individuals with remitted MDD, and cognitive and emotional patterns identified are viewed as underlying risk factors for MDD (Just et al., 2001). However, recent theories of depression posit that people remitted from MDD may belong to either a high-recurrence group or a low-recurrence group (Monroe et al., 2019). Therefore, having experienced a prior MDE does not necessarily indicate that one is at risk of recurrence, and cognitive or emotional characteristics associated with remitted MDD may not reflect underlying vulnerability for MDD. Future research on ER and ER diversity should consider the heterogeneous nature of remitted MDD. In addition, potential recurrence risk factors identified using remission designs should be corroborated using prospective designs.

Second, there is a need to examine whether ER diversity predicts relevant clinical features or comorbidities of MDD. MDD has high rates of comorbidity with both internalizing and externalizing disorders as well as suicidality and nonsuicidal self-injury (Kessler et al., 2014), and ER plays a key role in such high co-occurrences (e.g., Aldao et al., 2010; McLaughlin et al., 2011). For instance, emotion dysregulation in MDD may predict comorbid substance use (e.g., Conway et al., 2006; Swendsen et al., 2010). Thus, low diversity in the use of adaptive ER strategies and/or high diversity in maladaptive ER strategy use may increase the risk of comorbid disorders that are especially characterized by emotion dysregulation. The stronger association of the ER diversity index with depression status compared with existing ER measures in the current study suggests that the ER diversity index may be a useful metric in understanding the complex relations between depression, emotion dysregulation, and associated outcomes and comorbidities. Future research should examine whether ER diversity predicts disorders and symptoms associated with ER, such as suicidality and substance use, better than other ER measures.

Third, future research may incorporate other emotions and types of ER in examining the link between ER diversity and depression. The current study focused on intrapersonal ER strategies used to regulate negative emotions (e.g., Garnefski et al., 2001; Gross, 2002). However, it would be valuable to examine the link between depression and diversity in the use of interpersonal ER strategies and in the regulation of positive emotions. Interpersonal models of emotion regulation (e.g., Hoffmann, 2014; Zaki & Williams, 2013) highlight the social contexts in which emotions and regulatory processes occur. Interpersonal ER strategies (e.g., initiating social contact to regulate one’s emotional experience) may be particularly relevant to MDD, which is characterized by deficits in interpersonal functioning, including heightened vigilance for interpersonal rejection, reduced prosocial behaviors, and excessive reassurance seeking (e.g., Joiner et al., 2004).

Research on both intrapersonal and interpersonal ER strategies have focused on the regulation of negative emotions (e.g., Hofmann, 2014; Rottenberg, 2017). However, MDD is characterized by low positive affect in addition to high negative affect (for review, see Khazanov & Ruscio, 2016). Furthermore, maladaptive responses to positive affect have been linked to higher depressive symptom levels (for review, see Silton et al., 2020). For instance, suppression and dampening of positive emotion has been associated with MDD (e.g., Beblo et al., 2012; Werner-Seidler et al., 2013). The examination of diversity in the regulation of both positive and negative emotions would provide a more comprehensive understanding of ER diversity and depression.

Finally, the current findings on ER diversity may have important clinical implications. The ER diversity index is more strongly associated with depression status and provides additional information about ER patterns beyond the frequency of ER strategy use. Thus, assessments of the diversity in ER strategy use in clinical settings using the ER diversity index can comprehensively assess for ER patterns and help track changes in ER over time and/or the course of therapeutic treatment. Furthermore, given the importance of ER diversity in MDD, therapeutic treatments may benefit from focusing not only on learning and applying adaptive ER skills but also on practicing to apply a diverse set of ER strategies in a context-dependent and effective manner. However, additional studies on ER diversity in MDD and related conditions are needed before furthering the discussion on clinical implications.

Conclusion

In the present study, we proposed and evaluated the ER diversity index as a novel measure of ER diversity in current and remitted MDD. The ER diversity index was more strongly associated with depression status than existing measures of ER. The stronger association of current and remitted MDD status with the ER diversity index relative to the ER sum score was particularly pronounced for adaptive ER strategies. Individuals with current and remitted MDD displayed more diversity in the use of ER strategies overall and of maladaptive strategies than healthy control participants but less diversity in the use of adaptive ER strategies. These findings emphasize the importance of considering the diversity in ER strategy use in addition to the frequency in ER strategy use in depression. Future studies integrating ER diversity with other aspects of flexible ER strategy use (e.g., contextual demand, regulatory effectiveness) using more ecologically valid methods will further advance research in this area.

Footnotes

Transparency

Action Editor: Stefan G. Hoffman

Editor: Kenneth J. Sher

Author Contributions

The study concept was developed by A. Wen and was refined by L. Quigley and K. L. Yoon. The study design was developed by L. Quigley under the supervision of K. S. Dobson. Testing and data collection were performed by L. Quigley and A. Wen under the supervision of K. S. Dobson. A. Wen. performed the data analysis and interpretation under the supervision of L. Quigley and K. L. Yoon. A. Wen drafted the manuscript, and L. Quigley, K. L. Yoon, and K. S. Dobson provided revisions. All of the authors approved the final manuscript for submission.

ORCID iDs

Alainna Wen

K. Lira Yoon

Supplemental Material

Additional supporting information can be found at

Notes

References

Aldao

(2013). The future of emotion regulation research: Capturing context. Perspectives on Psychological Science, 8(2), 155–172. https://doi.org/10.1177/1745691612459518

Aldao

Nolen-Hoeksema

Schweizer

(2010). Emotion-regulation strategies across psychopathology: A meta-analytic review. Clinical Psychology Review, 30(2), 217–237. https://doi.org/10.1016/j.cpr.2009.11.004

Aldao

Sheppes

Gross

J. J.

(2015). Emotion regulation flexibility. Cognitive Therapy and Research, 39(3), 263–278. https://doi.org/10.1007/s10608-014-9662-4

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.).

Beblo

Fernando

Klocke

Griepenstroh

Aschenbrenner

Driessen

(2012). Increased suppression of negative and positive emotions in major depression. Journal of Affective Disorders, 141(2–3), 474–479. https://doi.org/10.1016/j.jad.2012.03.019

Beck

A. T.

Steer

R. A.

Brown

G. K.

(1996). Manual for the Beck Depression Inventory-II. Psychological Corporation.

Bonanno

G. A.

Burton

C. L.

(2013). Regulatory flexibility: An individual differences perspective on coping and emotion regulation. Perspectives on Psychological Science, 8(6), 591–612. https://doi.org/10.1177/1745691613504116

Bonanno

G. A.

Papa

Lalande

Westphal

Coifman

(2004). The importance of being flexible: The ability to both enhance and suppress emotional expression predicts long-term adjustment. Psychological Science, 5(7), 482–487. https://doi.org/10.1111/j.0956-7976.2004.00705.x

Bonanno

G. A.

Pat-Horenczyk

Noll

(2011). Coping flexibility and trauma: The Perceived Ability to Cope with Trauma (PACT) scale. Psychological Trauma: Theory, Research, Practice, and Policy, 3(2), 117–129. https://doi.org/10.1037/a0020921

10.

Burnham

K. P.

Anderson

D. R.

(2002). A practical information-theoretic approach: Model selection and multimodel inference (2nd ed.). Springer.

11.

Campbell-Sills

Barlow

D. H.

Brown

Hofmann

S. G.

(2006). Acceptability and suppression of negative emotion in anxiety and mood disorders. Emotion, 6(4), 587–595. https://doi.org/10.1037/1528-3542.6.4.587

12.

Cheng

(2001). Assessing coping flexibility in real-life and laboratory settings: A multimethod approach. Journal of Personality and Social Psychology, 80(5), 814–833. https://doi.org/10.1037/0022-3514.80.5.814

13.

Cheng

Cheung

M. W. L.

(2005). Cognitive processes underlying coping flexibility: Differentiation and integration. Journal of Personality, 73(4), 859–886. https://doi.org/10.1111/j.1467-6494.2005.00331.x

14.

Cheng

Lau

H. P. B.

Chan

M. P. S.

(2014). Coping flexibility and psychological adjustment to stressful life changes: A meta-analytic review. Psychological Bulletin, 140(6), 1582–1607. https://doi.org/10.1037/a0037913

15.

Compas

B. E.

Jaser

S. S.

Dunbar

J. P.

Watson

K. H.

Bettis

A. H.

Gruhn

M. A.

Williams

E. K.

(2014). Coping and emotion regulation from childhood to early adulthood: Points of convergence and divergence. Australian Journal of Psychology, 66(2), 71–81. https://doi.org/10.1111/ajpy.12043

16.

Conway

K. P.

Compton

Stinson

F. S.

Grant

B. F.

(2006). Lifetime comorbidity of DSM-IV mood and anxiety disorders and specific drug use disorders: Results from the National Epidemiologic Survey on alcohol and related conditions. The Journal of Clinical Psychiatry, 67, 247–257. https://doi.org/10.4088/JCP.v67n0211

17.

Faul

Erdfelder

Lang

A.-G.

Buchner

(2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175–191. https://doi-org.proxy.library.nd.edu/10.3758/BF03193146

18.

First

M. B.

Spitzer

R. L.

Gibbon

Williams

J. B. W.

(2002). Structured Clinical Interview for DSM-IV-TR Axis I Disorders, research version, non-patient edition for DSM IV. New York State Psychiatric Institute.

19.

Fresco

D. M.

Williams

N. L.

Nugent

N. R.

(2006). Flexibility and negative affect: Examining the associations of explanatory flexibility and coping flexibility to each other and to depression and anxiety. Cognitive Therapy and Research, 30(2), 201–210. https://doi.org/10.1007/s10608-006-9019-8

20.

Garnefski

Koopman

Kraaij

ten Cate

(2009). Brief report: Cognitive emotion regulation strategies and psychological adjustment in adolescents with a chronic disease. Journal of Adolescence, 32(2), 449–454. https://doi.org/10.1016/j.adolescence.2008.01.003

21.

Garnefski

Kraaij

(2006). Relations between cognitive emotion regulation strategies and depressive symptoms: A comparative study of five specific samples. Personality and Individual Differences, 40(8), 1659–1669. https://doi.org/10.1016/j.paid.2005.12.009

22.

Garnefski

Kraaij

Spinhoven

(2001). Negative life events, cognitive emotion regulation and emotional problems. Personality and Individual Differences, 30(8), 1311–1327. https://doi.org/10.1016/S0191-8869(00)00113-6

23.

Garnefski

Legerstee

Kraaij

van Den Kommer

Teerds

J. A. N.

(2002). Cognitive coping strategies and symptoms of depression and anxiety: A comparison between adolescents and adults. Journal of Adolescence, 25(6), 603–611. https://doi.org/10.1006/jado.2002.0507

24.

Goldin

P. R.

McRae

Ramel

Gross

J. J.

(2008). The neural bases of emotion regulation: Reappraisal and suppression of negative emotion. Biological Psychiatry, 63(6), 577–586. https://doi.org/10.1016/j.biopsych.2007.05.031

25.

Gotlib

I. H.

Joormann

(2010). Cognition and depression: Current status and future directions. Annual Review of Clinical Psychology, 6, 285–312. https://doi.org/10.1146/annurev.clinpsy.121208.131305

26.

Grommisch

Koval

Hinton

J. D.

Gleeson

Hollenstein

Kuppens

Lischetzke

(2020). Modeling individual differences in emotion regulation repertoire in daily life with multilevel latent profile analysis. Emotion, 20(8), 1462–1474. https://doi.org/10.1037/emo0000669

27.

Gross

J. J.

(1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, 2(3), 271–299. https://doi.org/10.1037/1089-2680.2.3.271

28.

Gross

J. J.

(2002). Emotion regulation: Affective, cognitive, and social consequences. Psychophysiology, 39(3), 281–291. https://doi.org/10.1017.S0048577201393198

29.

Hammen

(1991). Generation of stress in the course of unipolar depression. Journal of Abnormal Psychology, 100(4), 555–561. https://doi.org/10.1037/0021-843X.100.4.555

30.

Hammen

(2006). Stress generation in depression: Reflections on origins, research, and future directions. Journal of Clinical Psychology, 62(9), 1065–1082. https://doi.org/10.1002/jclp.20293

31.

Hofmann

S. G.

(2014). Interpersonal emotion regulation model of mood and anxiety disorders. Cognitive Therapy and Research, 38(5), 483–492. http://doi.org/10.1007/s10608-014-9620-1

32.

Hofmann

S. G.

Sawyer

A. T.

Fang

Asnaani

(2012). Emotion dysregulation model of mood and anxiety disorders. Depression and Anxiety, 29(5), 409–416. http://doi.org/10.1002/da.21888

33.

Joiner

Coyne

J. C.

Blalock

(1999). On the interpersonal nature of depression: Overview and synthesis. In Joiner

Coyne

J. C.

(Eds.), The interactional nature of depression: Advances in interpersonal approaches (p. 3–19). American Psychological Association. https://doi.org/10.1037/10311-013

34.

Just

Abramson

L. Y.

Alloy

L. B.

(2001). Remitted depression studies as tests of the cognitive vulnerability hypotheses of depression onset: A critique and conceptual analysis. Clinical Psychology Review, 21(1), 63–83. https://doi.org/10.1016/S0272-7358(99)00035-5

35.

Kato

(2012). Development of the Coping Flexibility Scale: Evidence for the coping flexibility hypothesis. Journal of Counseling Psychology, 59(2), 262–273. https://doi.org/10.1037/a0027770

36.

Kato

(2015). Testing of the coping flexibility hypothesis based on the dual-process theory: Relations between coping flexibility and depressive Symptoms. Psychiatry Research, 230(2), 137–142. https://doi.org/10.1016/j.psychres.2015.07.030

37.

Kessler

R. C.

de Jonge

Shahly

van Loo

H. M.

Wang

P. S.-E.

Wilcox

M. A.

(2014). Epidemiology of depression. In Gotlib

I. H.

Hammen

C. L.

(Eds.), Handbook of depression (p. 7–24). The Guilford Press.

38.

Khazanov

G. K.

Ruscio

A. M.

(2016). Is low positive emotionality a specific risk factor for depression? A meta-analysis of longitudinal studies. Psychological Bulletin, 142(9), 991. https://doi.org/10.1037/bul0000059

39.

Krkovic

Krink

Lincoln

T. M.

(2018). Emotion regulation as a moderator of the interplay between self-reported and physiological stress and paranoia. European Psychiatry, 49, 43–49. https://doi.org/10.1016/j.eurpsy.2017.12.002

40.

Lam

C. B.

McBride-Chang

C. A.

(2007). Resilience in young adulthood: The moderating influences of gender-related personality traits and coping flexibility. Sex Roles, 56(3–4), 159–172. https://doi.org/10.1007/s11199-006-9159-z

41.

Lazarus

R. S.

Folkman

(1984). Stress, appraisal, and coping. Springer.

42.

Lecrubier

Sheehan

D. V.

Weiller

Amorim

Bonora

Sheehan

K. H.

Janvas

Dunbar

G. C.

(1997). The Mini International Neuropsychiatric Interview (MINI). A short diagnostic structured interview: Reliability and validity according to the CIDI. European Psychiatry, 12(5), 224–231. https://doi.org/10.1016/S0924-9338(97)83296-8

43.

LeMoult

Gotlib

I. H.

(2019). Depression: A cognitive perspective. Clinical Psychology Review, 69, 51–66. https://doi.org/10.1016/j.cpr.2018.06.008

44.

Long

J. S.

Freese

(2006). Regression models for categorical dependent variables using Stata (Vol. 7). Stata Press.

45.

Lumley

Scott

(2015). AIC and BIC for modeling with complex survey data. Journal of Survey Statistics and Methodology, 3(1), 1–18.

46.

McLaughlin

K. A.

Hatzenbuehler

M. L.

Mennin

D. S.

Nolen-Hoeksema

(2011). Emotion dysregulation and adolescent psychopathology: A prospective study. Behaviour Research and Therapy, 49(9), 544–554. https://doi.org/10.1016/j.brat.2011.06.003

47.

Monroe

S. M.

Anderson

S. F.

Harkness

K. L.

(2019). Life stress and major depression: The mysteries of recurrences. Psychological Review, 126(6), 791–816. https://doi.org/10.1037/rev0000157

48.

Ong

A. D.

Benson

Zautra

A. J.

Ram

(2018). Emodiversity and biomarkers of inflammation. Emotion, 18(1), 3–14. https://doi.org/10.1037/emo0000343

49.

Quigley

Wen

Dobson

K. S.

(2020). Cognitive control over emotional information in current and remitted depression. Behaviour Research and Therapy, 132, Article 103658. https://doi.org/10.1016/j.brat.2020.103658

50.

Quoidbach

Gruber

Mikolajczak

Kogan

Kotsou

Norton

M. I.

(2014). Emodiversity and the emotional ecosystem. Journal of Experimental Psychology: General, 143(6), 2057–2066. https://doi.org/10.1037/a0038025

51.

Quoidbach

Mikolajczak

Gruber

Kotsou

Kogan

Norton

M. I.

(2018). Robust, replicable, and theoretically-grounded: A response to Brown and Coyne’s (2017) commentary on the relationship between emodiversity and health. Journal of Experimental Psychology: General, 147(3), 451–458. https://doi.org/10.1037/xge0000400

52.

Rottenberg

(2017). Emotions in depression: What do we really know? Annual Review of Clinical Psychology, 13(1), 241–263. https://doi.org/10.1146/annurev-clinpsy-032816-045252

53.

Shannon

C. E.

(1948). A mathematical theory of communication. Bell System Technical Journal, 27(3), 379–423. https://doi.org/10.1002/j.1538-7305.1948.tb00917.x

54.

Sheehan

D. V.

Lecrubier

Sheehan

K. H.

Amorim

Janavs

Weiller

Hergueta

Baker

Dunbar

G. C.

(1998). The Mini-International Neuropsychiatric Interview (M.I.N.I.): The development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. In. Journal of Clinical Psychiatry, 59, 22–33. https://doi.org/10.1016/S0924-9338(99)80239-9

55.

Sheppes

Gross

J. J.

(2011). Is timing everything? Temporal considerations in emotion regulation. Personality and Social Psychology Review, 15(4), 319–331. https://doi.org/10.1177/1088868310395778

56.

Sheppes

Gross

J. J.

(2012). Emotion regulation effectiveness: What works when. In Tennen

Suls

Weiner

I. B.

(Eds.), Handbook of psychology. Vol. 5: Personality and social psychology (2nd ed., pp. 391–405). John Wiley & Sons, Inc. https://doi.org/10.1002/9781118133880.hop205018

57.

Silton

R. L.

Kahrilas

I. J.

Skymba

H. V.

Smith

Bryant

F. B.

Heller

(2020). Regulating positive emotions: Implications for promoting well-being in individuals with depression. Emotion, 20(1), 93–97. http://doi.org/10.1037/emo0000675

58.

Swendsen

Conway

K. P.

Degenhardt

Glantz

Jin

Merikangas

K. R.

Sampson

Kessler

R. C.

(2010). Mental disorders as risk factors for substance use, abuse and dependence: Results from the 10-year follow-up of the National Comorbidity Survey. Addiction, 105, 1117–1128. https://doi.org/10.1111/j.1360-0443.2010.02902.x

59.

Urban-Wojcik

E. J.

Mumford

J. A.

Almeida

D. M.

Lachman

M. E.

Ryff

C. D.

Davidson

R. J.

Schaefer

S. M.

(2020). Emodiversity, health, and well-being in the Midlife in the United States (MIDUS) daily diary study. Emotion. Advance online publication. https://doi.org/10.1037/emo0000753

60.

Vrieze

S. I.

(2012). Model selection and psychological theory: A discussion of the differences between the Akaike information criterion (AIC) and the Bayesian information criterion (BIC). Psychological Methods, 17(2), 228–243. https://doi.org/10.1037/a0027127

61.

Vriezekolk

J. E.

Geenen

van den Ende

C. H. M.

Slot

van Lankveld

W. G. J. M.

van Helmond

(2012a). Behavior change, acceptance, and coping flexibility in highly distressed patients with rheumatic diseases: Feasibility of a cognitive-behavioral therapy in multimodal rehabilitation. Patient Education and Counseling, 87(2), 171–177. https://doi.org/10.1016/j.pec.2011.09.001

62.

Vriezekolk

J. E.

van Lankveld

W. G.

Eijsbouts

A. M.

van Helmond

Geenen

van den Ende

C. H.

(2012b). The coping flexibility questionnaire: development and initial validation in patients with chronic rheumatic diseases. Rheumatology International, 32(8), 2383–2391. https://doi.org/10.1007/s00296-011-1975-y

63.

Washington

H. G.

(1984). Diversity, biotic and similarity indices. A review with special relevance to aquatic ecosystems. Water Research, 18(6), 653–694. https://doi.org/10.1016/0043-1354(84)90164-7

64.

Werner-Seidler

Banks

Dunn

B. D.

Moulds

M. L.

(2013). An investigation of the relationship between positive affect regulation and depression. Behaviour Research and Therapy, 51(1), 46–56. http://doi.org/10.1016/j.brat.2012.11.001

65.

World Health Organization. (1990). Composite international diagnostic interview.

66.

Zaki

Williams

W. C.

(2013). Interpersonal emotion regulation. Emotion, 13(5), 803–810. http://doi.org/10.1037/a0033839

67.

Zong

J. G.

Cao

X. Y.

Cao

Shi

Y. F.

Wang

Y. N.

Yan

Abela

J. R. Z.

Gan

Y-Q.

Gong

Q.-Y.

Chan

R. C. K.

(2010). Coping flexibility in college students with depressive symptoms. Health and Quality of Life Outcomes, 8, Article 66. https://doi.org/10.1186/1477-7525-8-66