Coping Strategies and Subjective Well-being: Context Matters

Poor subjective well-being (SWB)—defined as “the extent to which a person believes or feels his or her life is going well” (Diener et al. 2018, p. 1)—has been observed in various psychiatric disorders, such as depression (Baselmans et al. 2018), psychosis (Broyd et al. 2016; van Dijk et al. 2019), and gambling disorder (Farrell 2018). Seligman (2018), the founder of positive psychology, has emphasized that reducing or eliminating negative impacts of psychopathology, for example, does not inevitably result in elevated well-being. Similarly, Keyes (2005) has proposed a complete state model of health. According to this model, health and illness are not extreme poles on one bipolar dimension but rather two unipolar dimensions that correlate. Thus, individuals can be classified as completely mentally ill (presence of mental illness and absence of mental health), languishing (absence of mental health), flourishing (presence of mental health), or completely mentally healthy (absence of mental illness and presence of mental health). Complete mental health is associated with a variety of favorable outcomes such as low helplessness and high goals, resilience, and intimacy. However, it is also rare and has been found in only 17% of a representative U.S. population sample (Keyes 2005). Consequently, the absence of (dysfunctional) coping strategies that coincide with psychopathology does not automatically predict increased SWB. We also sought to explore whether a high use of presumably adaptive strategies directly translates into high SWB.

Thus far, few studies have assessed the relationship between coping strategies and SWB in the general population. One recent study investigated the correlations of SWB with the frequency and perceived efficacy of coping strategies for general affect regulation and for sadness regulation specifically (Korpela et al. 2018). The authors found varying but generally positive associations with SWB for positive activities, positive thinking, problem-directed action, and cognitive reappraisal. Similar associations were found in an Asian population in which positive reinterpretation, active coping, and planning were among the coping strategies positively correlated with life satisfaction and positive affect, whereas mental disengagement, denial, and behavioral disengagement were negatively associated with SWB measures (Chang et al. 2020). Nonetheless, many other strategies’ associations with SWB remain unexplored.

The present study pursued two goals. First, we aimed to examine a revised version of the Maladaptive and Adaptive Coping Scale (MAX), which we extended to incorporate contextual factors. This allowed us to test whether, for example, expressive suppression, a strategy considered to be maladaptive, may under some circumstances manifest as adaptive (e.g., hiding negative emotions in front of strangers). Second, we examined which coping styles contribute to SWB. As mentioned, research has established that certain maladaptive styles contribute to the formation and maintenance of psychological disorders. What remains to be investigated is whether high levels of SWB and high levels of depressiveness are associated with opposite engagement patterns of presumably adaptive and presumably maladaptive coping.

Participants were recruited using Amazon’s Mechanical Turk (MTurk) in February and March 2019. Inclusion criteria were age 18 years or older and an IP address from the United States. A total of 1000 participants started the survey; 986 completed it. To address concerns of inattention, participants were required to have approval rates on previous studies of 95 percent or higher (Buhrmester et al. 2018). Poor results on an explicit attention check item (“I was focused while taking part in this study,” answered on a 7-point scale: “strongly disagree” through “strongly agree”; a score of < 5 led to exclusion) or on an implicit attention check item adopted from Oppenheimer et al. (2009) led to the exclusion of presumably inattentive participants. Finally, participants who sped through the survey (taking less than 60% of the estimated completion time) were also excluded (Arditte et al. 2016). All exclusions were made blind to results. After removal of inattentive subjects, n = 836 participants were eligible for the final analyses.

Data collection took place within a larger study that included several other questionnaires and experimental paradigms unrelated to the present study. The estimated time to complete the entire study was 30 min; participants received $2.50 upon completion. All questionnaires were implemented using the online survey tool Qualtrics.

All analyses were performed in SPSS (IBM Corp., Version 25). First, we submitted the MAX to an exploratory factor analysis with orthogonal (varimax) rotation. Second, Pearson correlation coefficients were calculated for study variables. Finally, we performed hierarchical multiple linear regression using forced entry to predict (a) affect balance, (b) life satisfaction, and (c) depressiveness (as measured by the PHQ-9). In the first step of the hierarchical model, we entered sociodemographic data (age, gender, education, ethnicity), psychiatric diagnoses, and current mood. When predicting (a) affect balance and (b) life satisfaction, we also entered depressiveness as a predictor in the first step of the model; when predicting (c) depressiveness, we also included affect balance and life satisfaction as predictors in the first step of the model. In the second step (final model) we added the four coping subscales as well as depressiveness for the outcomes (a) affect balance and (b) life satisfaction, and SWB measures for (c) depressiveness.

As previous research has focused on the genders of male and female, we only investigated gender differences between males and females (“binary gender”) and excluded four participants from the analysis who did not identify as either male or female. Education was operationalized as three dummy variables comparing (1) started but incomplete college degree, (2) completed college degree, and (3) professional degree to high school or less. As the present sample consisted mostly of individuals identifying as White/European American, we entered ethnicity as a binary variable (White vs. non-White). Psychiatric diagnoses were entered as two dichotomous variables. The first variable reflected whether a participant had screened positive for at least one psychiatric diagnosis on the WSQ; the second indicated whether a participant had previously been diagnosed with a psychiatric illness, according to self-report.

Mean age of participants was 40.3 (SD = 12.6) years with a range of 18 to 75 years. Almost all participants were native English speakers (96.9%) and U.S. citizens (98.2%) according to self-report. Slightly more than half of the sample (56.2%) identified as female; 0.4% identified as neither male nor female. Overall, 0.6% of participants had completed less than a high school degree, 11.1% had graduated high school, 24.4% had started college, 48.6% had a college degree, and 15.3% had a professional degree or doctorate. More than three-quarters of the participants (78.0%) identified as White/European American (Black/African American: 11%, Latino/Hispanic: 6.3%, Asian/Asian American: 5.7%, Native American/Hawaiian: 1.9%, other: 0.6%; multiple answers possible). In the sociodemographic questionnaire, 54.8% of all participants stated that they had not previously received a psychiatric diagnosis. The numbers of participants who had received psychiatric diagnoses is displayed in Table 4 in the Appendix. Table 4 also shows the numbers of participants who screened positive for psychiatric disorders using the WSQ. According to the PHQ-9, 24.2% of participants were classified as currently depressed when using a cut-off value of 10 as recommended by Manea et al. (2012).

On average, participants displayed more positive than negative affect (mean affect balance score = 7.63, SD = 9.23; one-sample t-test against midpoint: t(834) = 23.874, p < .001) and moderate life satisfaction (mean score = 62.83, SD = 20.46). Mean depressiveness (PHQ-9) score was 6.17 (SD = 6.21), reflecting mild (subclinical) depressive symptoms (Kroenke and Spitzer 2002).

The factor analysis of the MAX-R resulted in seven factors with an eigenvalue greater than 1. Based on the scree plot inflections, four factors were retained, explaining 43.9% of total variance. The first (14.85% variance explained), second (12.58%), and third (8.62%) factors resembled the adaptive, maladaptive, and avoidance scales reported by Moritz et al. (2016) and were named accordingly. The fourth factor (7.85%) was labeled “expressive suppression” as it was comprised only of variants of suppressing outward emotional expression, which is referred to as expressive suppression (Dryman and Heimberg 2018). Each of the four factors was used as a subscale in the final questionnaire. All items with loadings of 0.5 or higher were included in the respective subscale. The full pattern matrix is displayed in Table 1. Context-related allocations are described in Sect. 3.3.

Two previously (according to Moritz et al. 2016) adaptive and one previously maladaptive item showed factor loadings below 0.5 and were subsequently not allocated to a subscale. None of the three items allocated to the avoidance subscale by Moritz et al. (2016) remained part of it; two of the items were then categorized as expressive suppression while one item’s loading did not exceed 0.5 for avoidance. Several of the newly included items did not exceed loadings of 0.5 for any factor.

Internal consistency of the newly formed subscales was good for adaptive (α = 0.86) and maladaptive (α = 0.88) coping and acceptable for avoidance coping (α = 0.73) and expressive suppression (α = 0.78). The latter two subscales show improved reliability compared to the original MAX’s avoidance subscale (α = 0.65; Moritz et al. 2016). Inter-item correlations of subscales ranged from r = 0.512 for adaptive, from r = 0.120 to 0.667 for maladaptive, from r = 0.280 to 0.487 for avoidance, and from r = 0.338 to 0.751 for expressive suppression. Six items had commonalities lower than 0.4 and were therefore not included in any subscale: “Thinking over my problems again and again has helped me find solutions in the past,” “I use humor to help cope with my problems (e.g., imagining an absurd outcome to a situation or joking with a friend),” “I avoid things and situations that get me into trouble,” “It is easy to tell by looking at me that I am suffering in a certain situation,” “I avoid problems,” and “When I have a problem or feel distressed, I feel better after I burst out with emotion.”

Mean item endorsement scores were highest for expressive suppression (2.89, SD = 0.62), followed by adaptive (2.88, SD = 0.53), maladaptive (2.52, SD = 0.66), and avoidance (1.93, SD = 0.60) coping. Bivariate correlations of coping subscales, SWB measures, and depressiveness can be found in Table 2.

The factor analysis of the MAX-R essentially replicated the previous structure yielding adaptive, maladaptive, and avoidance subscales (Moritz et al. 2016). The avoidance subscale then comprised new context-capturing items (e.g., “I do not try to change a negative situation even if I could actually change it”). We additionally identified a fourth subscale, which we labeled expressive suppression in accordance with the literature (e.g., Dryman and Heimberg 2018).

Bivariate correlations between coping subscales, SWB measures (affect balance and life satisfaction), and depressiveness emerged in the expected directions. Specifically, both affect balance and life satisfaction were significantly positively associated with adaptive coping and significantly negatively associated with all other coping subscales and depressiveness. Depressiveness was significantly negatively associated with adaptive coping and significantly positively associated with the other three coping subscales.

Multiple linear regression showed that—after the influence of sociodemographic variables, psychiatric diagnoses, current mood, and depressiveness was accounted for—adaptive, maladaptive, and expressive suppression coping, but not avoidance coping, contributed significantly to the prediction of affect balance and life satisfaction in the expected directions. After accounting for sociodemographic variables, psychiatric diagnoses, current mood, and SWB measures, adaptive, maladaptive, and avoidance coping, but not expressive suppression coping, significantly contributed to the prediction of depressiveness. All associations with depressiveness were in the expected direction with the exception of adaptive coping, which yielded a small positive contribution to depressiveness scores. In summary, differential patterns of small but significant contributions of coping to SWB and depressiveness emerged.

Of the covariates, gender significantly predicted SWB (but not depressiveness) in that women showed higher SWB scores than men. Younger age was associated with higher scores on all three outcomes. Higher education was associated with higher life satisfaction (but not affect balance). Individuals who had started but not finished college showed higher depressiveness than individuals who had completed high school or less; no differences were found for college or professional degree compared to high school degree. Participants who identified as White showed higher affect balance scores than participants who identified as non-White. Current mood predicted SWB in a positive direction but not depressiveness.

We added several items that put the coping strategies into different contexts to the MAX questionnaire to test whether some coping strategies are maladaptive in some contexts but adaptive in others. For some strategies, for example catastrophizing, all items loaded on the factor maladaptive regardless of context. Hence, catastrophizing appears to always be maladaptive. The former (MAX) and new (MAX-R) expressive suppression items (see Table 1) also did not present as differing in terms of adaptiveness. Instead of being part of the previously established adaptive versus maladaptive subscales, they now constitute a new subscale. This new subscale significantly contributed to the prediction of the two SWB measures (negative association) but not to the prediction of depressiveness.

Interestingly, acceptance appears to be less clear-cut in terms of functionality. We found that the original noncontextual acceptance item (“I accept a situation with which I am not totally happy and try to make the best of it”) was replicated as adaptive. Yet, it appears that the question of whether one is actually able to influence the situation at hand plays a role in the functionality of the coping behavior. According to our data, accepting situations is most adaptive if the situation presumably cannot be changed. When it is clear that the individual does have a subjective influence over the unfavorable situation (“When I am unhappy with a situation, I try to change it if it can be changed”; “I do not try to change a negative situation even if I could actually change it”), then accepting the situation appears to represent an avoidance strategy and accordingly was not associated with SWB; instead, a weak association with depressiveness emerged. To conclude, our results essentially agree with previous work (e.g., Aldao 2013; English et al. 2017; Tang and Huang 2019) emphasizing the need to consider context factors when judging strategies’ functionality. However, context seems to be of varying importance for different styles.

Overall, our findings speak to the hypothesis that well-being as an expression of mental health correlates with, but is not the polar opposite of, mental illness (Keyes 2005). We found differential associations of coping subscales with affect balance, life satisfaction, and depressiveness. Although maladaptive coping predicted SWB via negative associations (B =  − 0.26, p < .001 for affect balance; B =  − 0.32, p = .009 for life satisfaction) and depressiveness via a positive association (B = 0.22, p < .001) at similar magnitudes, the other coping subscales did not show inverse correlations with SWB and depressiveness as one would expect if SWB and depressiveness were at opposite ends of one bipolar construct.

We found that both adaptive and maladaptive strategy use predicts both SWB and depressiveness scores. However, avoidance coping was found to predict depressiveness but not SWB, while expressive suppression was found to predict SWB but not depressiveness. Numerically, adaptive coping has the highest explanatory value for life satisfaction, maladaptive coping has the highest explanatory value for depressiveness, avoidance coping has significant explanatory value only for depressiveness, and expressive suppression has significant explanatory value only for SWB (more so for life satisfaction than for affect balance).

Thus far, few studies have investigated the link between adaptive and maladaptive coping strategies and well-being. Our finding that adaptive strategies show a positive correlation and maladaptive strategies a negative correlation with aspects of SWB is in line with a previous study by Balzarotti et al. (2016). They found that positive reappraisal and catastrophizing in particular were strongly associated with SWB positively or negatively, respectively. Similar to the results of Moritz et al. (2016), we also found the positive correlations of maladaptive (> 0.5) and avoidance (> 0.4) coping to be numerically higher with depressiveness than the negative correlation of adaptive coping (< − 0.4). When accounting for sociodemographic factors, current mood, SWB measures, and the other coping subscales, the unique contribution of adaptive coping to depressiveness was a small positive contribution. The adaptive coping score is a sum score of endorsement for all items on this subscale. This result thus may reflect the previous finding that a higher total sum of the maladaptive and adaptive strategies used is associated with a higher depressiveness score (Moritz et al. 2016) and the conclusion that a larger repertoire of coping strategies does not directly translate to decreased depressiveness. A possible explanation is that people who use certain coping strategies ineffectively use a larger pool of strategies to compensate for the insufficient regulation; however this theory needs to be examined further. Overall, our findings concur with the suggestion that maladaptive and adaptive coping are not inverse and that addressing maladaptive coping seems more essential than enhancing adaptive coping for alleviating psychiatric symptoms (Moritz et al. 2016).

In summary, certain coping strategies, particularly cognitive reappraisal, seem to have the potential to contribute to SWB while others, particularly catastrophizing and expressive suppression, appear to undermine SWB. This appears to be an add-on effect to the aggravating or appeasing role coping strategies play in psychopathologies such as depression, which further supports the notion that well-being is more than the absence of psychopathology (Franken et al. 2018; Keyes 2007).

We need to acknowledge several limitations. Due to the online design of our study, there are several aspects of the data-gathering process that we could not fully control, such as whether participants responded in a distraction-free environment. However, the online setting enabled us to recruit a large community sample and, as a precaution, we implemented several attention checks and additional measures to ensure good data quality (Buhrmester et al. 2018). In addition, we used self-report measures for the assessment of both coping strategies and SWB and our study design is cross-sectional, which does not allow for extended causal conclusions. Our results thus need to be corroborated by longitudinal designs, preferably using ecological momentary assessment. Analyzing our data, we further found that our sample was more highly educated (64% college or professional degree) than the general population of the U.S. aged 18 and older, of whom about 43% held a post-high school degree in 2019 (U.S. Census Bureau 2020). As we accounted for education when conducting our analysis, we do not believe that this difference is a significant concern. However, future studies should differentiate the associations with coping and SWB across various levels of education. We also do not know for certain how relevant the context provided in the items was for the participants. Future studies should refine contexts and replicate findings, especially concerning items that were intended to distinguish between cognitive or emotional suppression with close social contacts versus strangers. Although the additional items were able to tease out only a few contextual differences, the revised version of the questionnaire in our opinion does provide additional value for future studies due to its improved internal consistency. Finally, in our investigation we aimed to differentiate between subjective well-being and depressiveness. We have to acknowledge that SWB most likely does not cover all aspects of mental health and that depressiveness, as measured by the PHQ-9, is only one facet of mental illness or psychopathology. However, as depressed mood occurs in various forms of mental illness, we believe that it is an appropriate measure for this exploratory study. Further studies comparing undiagnosed individuals to those who have verified diagnoses of mental disorders are needed to explore in more depth the differential impact of coping strategies on mental health while accounting for mental illness.

Categorizing coping strategies into generally helpful versus harmful seems unwarranted as the functionality of at least some strategies appears to be dynamic and influenced by contextual factors (e.g., Aldao and Tull 2015; Aldao and Wisco 2015; Catterson et al. 2017; De France and Evans 2020; Rottweiler et al. 2018). Yet, there are countless components of context, such as certain characteristics of the person executing the emotion regulation (e.g., sleep quality or perceived stress level), features of the emotion-eliciting stimuli/situations (e.g., anxiety due to moving to a new home versus anxiety due to chronic illness), implementation of regulatory processes, and types of outcomes (Aldao 2013) that have yet to be fully explored.

Identifying strategies that are always adaptive or always maladaptive also serves as a valuable contribution to psychotherapy research as conveying versus disputing certain coping strategies is a substantial part of many psychological treatment approaches. However, the situation may not be clear-cut as depressed mood may in fact be associated with engaging in many strategies—both supposedly adaptive and supposedly maladaptive—and not being able to successfully regulate one’s response to stress, rather than with simply not engaging in adaptive coping.

While interventions aimed at fostering coping with everyday stressors have recently been developed for nonclinical populations such as students (e.g., Houston et al. 2017), more research is needed to determine whether stimulated use and non-use of coping strategies can directly translate into improved subjective well-being.

In conclusion, some coping strategies appear universally maladaptive or adaptive (e.g., catastrophizing), but for other coping strategies the impact on well-being depends on the situation (accepting versus attempting to change difficult situations). If replicated, this knowledge should be incorporated into psychotherapeutic treatment approaches that already target coping. Analyzing further if, when, and how coping strategies contribute positively to both mental health and well-being is essential for improving individuals’ lives beyond reducing the detrimental effects of particular psychological disorders.