As soon as someone experiences emotions, one has to deal with them in one way or the other. This is also referred to as emotion regulation (ER) and has been intensely studied for the last few decades (Gross, 2014). The process model of emotion regulation (Gross, 1998, p. 275) defines ER as “[…] the processes by which individuals influence which emotions they have, when they have them, and how they experience and express these emotions. Emotion regulatory processes may be automatic or controlled, conscious or unconscious, and may have their effects at one or more points in the emotion generative process.” More specifically, there are five stages where the emotion-generative process might be altered by ER: the situation selection (e.g., avoidance), the situation modification (e.g., behavioral problem-solving), attentional deployment (e.g., distraction, rumination), cognitive change (e.g., acceptance, reappraisal) and response modulation (e.g., suppression). Initially, research was primarily focused on the use of single ER-strategies (ERS) trying to identify them and differentiate their consequences and efficacy by experimental emotion-induction (Webb et al., 2012). However, recent studies found that often multiple ERS are used concurrently or sequentially, which is also referred to as emotion-polyregulation (Aldao and Nolen-Hoeksema, 2013; Brans et al., 2013; Ford et al., 2019). Ford et al. (2019) have only recently integrated polyregulation into the process model of ER by Gross (2015).

It is important to study temporal dynamics of the involved processes in order to capture them in an ecologically valid way. In naturalistic observation studies using for instance Ecological Momentary Assessment (EMA) (Koval et al., 2020; for a comprehensive overview see Koval and Kalokerinos, 2024) and experimental studies, research has shown a crucial impact of contextual factors on ER (Kobylińska and Kusev, 2019). For example, the type and intensity of emotion (Aldao, 2013; Barrett et al., 2001; Dixon-Gordon et al., 2015), motivation (Tamir et al., 2020), timing of implementation (Diedrich et al., 2016), and perceived controllability of the situation (Troy et al., 2013) influence not only the ER-success but also ERS-choice (Koval et al., 2015; Matthews et al., 2021; Sheppes et al., 2014). Interestingly, ER-success and ERS-choice are differently influenced by psychopathology: A recent study found that strategy selection rather than the implementation of those strategies is associated with mental disorders (Houben et al., 2023).

Several strategies have been identified that are typically used in case of higher stressor intensity or more emotionally loaded situations reflected by higher negative affect (putatively maladaptive ERS, e.g., distraction, rumination) and conversely others in case of lower stressor-intensity or lower levels of negative affect (putatively adaptive ERS, e.g., acceptance, reappraisal), which is probably a consequence of decreased cognitive resources and higher cognitive load (Aldao, 2013; Barrett et al., 2001; Blanke et al., 2022; Broderick, 2005; Dixon-Gordon et al., 2015; Sheppes, 2020; Sheppes et al., 2011, 2014). Further, ERS have been found to vary in their temporal deployment, with suppression and rumination occurring more at the beginning, and reappraisal and distraction occurring more toward the end of a negative emotional episode (Kalokerinos et al., 2017).

Of particular interest for this investigation is the association of ER with rumination. Originally, rumination was seen as a cognitive vulnerability to develop depressive disorders (Nolen-Hoeksema and Morrow, 1991). It is regarded as an abstract negative thinking-style where thoughts revolve around the past while having no goal-orientation (Teismann et al., 2012). In the context of Gross’ model (2015), rumination is categorized as a form of attentional deployment where there is repetitively a passive, self-immersed focus on the emotional features and consequences of a situation. Ruminative processes and other forms of repetitive negative thought have been found not only in depression but also other psychopathologies such as anxiety, eating and substance-related disorders (Arditte et al., 2016; McLaughlin and Nolen-Hoeksema, 2011; Svaldi et al., 2012). Rumination is also observed in healthy individuals, for instance in response to stressful life events (Moberly and Watkins, 2006; Robinson and Alloy, 2003; Ruscio et al., 2015). Ruminative thinking has been found to be highly persistent by predicting future rumination, increased negative affect and depressive symptoms (Bean et al., 2020; Boemo et al., 2022; Connolly and Alloy, 2017; Kircanski et al., 2018; Moberly and Watkins, 2008; Rosenbaum et al., 2021, 2022; Rosenbaum et al., 2018a, 2018b; Ruscio et al., 2015). Interestingly, in an EMA-study by Koval et al. (2012) rumination and emotional inertia both independently predicted depression severity in healthy undergraduates as well as depressed patients.

The scope of this study was to investigate different patterns of ERS used in response to daily life stress and their association with stress-reactive rumination and ER-success and perceived self-efficacy. Ten predefined ERS were assessed in a large community sample. As the pattern of ERS use showed high interrelationships, we performed an exploratory multilevel factor analysis in order to reduce data complexity. We aimed to find patterns of which ERS are used most commonly at the same time window and how those groups were associated with ER-success (direct effect measure), self-efficacy, reduction in stress (indirect effect measure) and reduction in momentary rumination (indirect effect measure). We set up the following hypotheses: (I) We expected to find decreased ER-success in case of higher stress and higher state-rumination and higher success in case of putatively adaptive (and lower in case of maladaptive) ERS. (II) We expected to find decreased self-perceived self-efficacy in case of higher stress and higher state-rumination and higher self-efficacy in case of putatively adaptive (and lower in case of maladaptive) ERS. (III) Concerning stress, we expected to find higher stress in the evening in case of higher stress in the morning and higher rumination (Rosenbaum et al., 2022). Further, we expected to find adaptive ERS to be associated with reduced stress and rumination. (IV) Analogously, state rumination should be elevated in case of higher previous state rumination and higher stress. Likewise, we expected adaptive ERS to be negatively associated with rumination.

The aim of the current online diary study was to investigate the use of emotion regulation strategies (ERS) in everyday life and their associations with perceived ER-success, self-efficacy, stress and state rumination. For this, after an assessment of baseline data, participants completed an online diary for 2 weeks. After exploration of the data and the observation of clustered use of multiple ERS at the same measurement time points, we firstly performed an exploratory multilevel factor analysis and a principal component analysis and identified four factors: cognitive perspective change, cognitive-behavioral problem-solving, suppression-distraction, body-social ERS. Although those factors already reduced data complexity, they were still rarely used alone. This underlines the importance of investigating emotion-polyregulation, which assumes that multiple ERS are implemented simultaneously/sequentially (Ford et al., 2019). Going further, this questions ERS-differentiation: our results show that ERS which were originally defined as altering different stages of the emotion-generative process (Gross, 1998), differentiated in the literature but all associated with cognitive perspective change (self-compassion, acceptance, reframing and mindfulness), loaded on one factor. Furthermore, the use of these ERS is almost always accompanied by the use of cognitive-behavioral ERS (cognitive and behavioral problem-solving). This gives rise to the idea that all of them might be facets of cognitive and behavioral ER which aims for cognitive-behavioral flexibility, for instance by cognitive perspective change. We further fitted mixed models with time-lagged variables investigating ER-success, self-efficacy, stress and state rumination dependent on ERS-factors.

Our results show that increased ER-success was significantly associated with lower stress, lower rumination, an increased use of cognitive perspective change, cognitive-behavioral problem-solving and body-social ERS. These results support our initial hypothesis (I). Interestingly, people using suppression-distraction habitually more often do not perceive their ER as less successful, however, when people use it more often than their person-mean, ER-success is perceived as less successful.

When we fitted the same models for self-efficacy as dependent variable, higher WP-rumination was associated with lower self-efficacy, which was in line with our hypothesis, whereas, contrary to our expectation, stress did not have a corresponding impact (II). We only found WP-ERS associations, namely cognitive perspective change and cognitive-behavioral problem-solving being associated with increased self-efficacy. The latter did no longer yield significance when ER-success was included. This finding suggests that the effect of ERS on self-efficacy is moderated by how successful the ER is perceived. However, the concrete causal relationship between ER-success and self-efficacy is unclear and difficult to investigate. Please note that BP-effects of ER-success and self-efficacy only correlate moderately, r(142) = 0.328, p < 0.001; however, this might also be influenced by a different item format (Likert-scale vs. slider). Disentangling these complex associations and potentially causal mechanisms will be an interesting endeavor for future research.

We further found a significant positive association of current state rumination with stress. Also, an increased WP-use of suppression-distraction was associated with higher stress. Further, individuals who use cognitive perspective change ERS habitually more often had lower stress ratings. These findings are in line with our hypotheses (III) and easily integrated into our previous findings showing an increased ER-success for cognitive perspective change as compared to the use of suppression-distraction. However, the effect of cognitive perspective change and WP-suppression-distraction no longer yielded significance when ER-success was added to the model. Higher ER-success was then associated with lower stress indicating that ER-success is more relevant than the respective strategy that was applied. However, it is important to keep in mind that there was quite some overlap between the strategies and their success. That is, cognitive perspective change was associated with more ER-success and suppression-distraction with lower success. However, in order to understand which strategies are perceived as more helpful for whom and in which situations, it still is important to get a better understanding of the effects of the different regulation strategies. When self-efficacy was added to the model, it was positively associated with stress. Please note that this seemed to be an artifact due to multicollinearity as this effect diminished when ER-success was not added to the model. In this context it is important to note that we did not randomize the order of items but it remained the same throughout the assessment. This might have an impact on how the items are answered that cannot be controlled in our analysis. Studies that randomize the order of items would not need to identify the potential effects that the order has on the data or responses to the items, but rather control for them independently.

Interestingly, despite the fact that cognitive-behavioral problem-solving was positively associated with ER-success on the WP level, we found an increased use of cognitive-behavioral problem-solving (WP and BP) to be associated with increased stress. This was originally not hypothesized but could be explained by situational demands that make the ERS-use more likely under high/low stress. In case of higher stress, cognitive-behavioral problem-solving might be used more frequently, which is why it is associated with increased stress but also increased ER-success. The stress-dependent use of ERS, namely cognitive perspective change and cognitive-behavioral problem-solving, might be explained by situational constraints such as limited cognitive resources and different needs for resources (Aldao, 2013; Barrett et al., 2001; Blanke et al., 2022; Broderick, 2005; Dixon-Gordon et al., 2015; Sheppes, 2020; Sheppes et al., 2011, 2014). As previously noted, our design does not allow us to determine whether ERS are used in a stress-dependent manner or if stress influences their use.

Please note that according to our hypotheses, we only used data entries where participants reported to have used at least one ERS when fitting our models. Additionally, we excluded participants with fewer than six consecutive data entries from the analysis to avoid distortions from less reliable subjects and to estimate parameters based on a sufficiently large data set. However, these criteria were not pre-registered. Future studies should pre-register their analyses to ensure unbiased exclusion of data points.

One question arising from these results is the definition of ER-efficacy and the appropriate measurement. This could be either done using a direct measure of ratings of ER-success or the actual difference in stress. The measure used might impact the results concerning ERS-efficiency; however, as ERS are most probably selected by situational constraints, direct measures of ER-success provide important additional information.

Investigating state rumination, we found higher stress to be associated with higher rumination on a BP- and WP-level, which supported our hypothesis (IV). Furthermore, we found a WP-increased use of cognitive perspective change to be associated with decreased state rumination. This effect, however, also diminished when ER-success was added to the model which was negatively associated with state rumination. This further resulted in WP-cognitive-behavioral problem-solving to yield a significant positive association, which also supports the idea of a substantial amount of shared variance in the model. State-of-the-art psychotherapeutic interventions already propose stressor-intensity-dependent ERS-implementation and aim to tackle ruminative processes (Goldberg et al., 2019; Hayes et al., 2007; Mennin and Fresco, 2013; Segal et al., 2008; Watkins, 2018) as they have shown to play an important role in various psychopathologies (Aldao et al., 2010). Our findings underscore the importance of rumination in everyday-life also in healthy individuals. As habitual rumination is associated with increased risk for elevated stress and negative affect, which is associated with increases in state rumination, this results in a self-sustaining loop of ruminative inertia (Bean et al., 2020; Connolly and Alloy, 2017; Kircanski et al., 2018; Moberly and Watkins, 2008; Rosenbaum et al., 2022; Ruscio et al., 2015).

The most profound limitation of the current study is the non-experimental design and correlative data which does not allow the investigation of causal mechanisms. As participants could choose ERS freely, this includes complex temporal dynamic processes of ERS selection and implementation and the reciprocal impact of situational constraints which are not controlled for in our set-up. Future studies might investigate these associations using an experimental EMA-design where participants are instructed to use ERS dependent on their current stress. Like this, it is possible to directly evaluate the situation-specificity of ER because with our data, no systematic variation of stressor-intensity has been implemented. As experimental laboratory studies with emotion-inductions often differ in terms of various crucial aspects compared to ecologically valid Ecological Momentary Assessment (EMA) data (e.g., the controllability and importance of negative events), EMA is promising for investigating these associations and it would be beneficial to combine the best of both study-designs.

A very promising and only recently developed approach is the situated assessment method (SAM²) (Dutriaux et al., 2023). It can be utilized to identify individual differences in habitual behavior and combined the evaluations of target behaviors and their situational influences to develop a comprehensive profile of an individual’s behavior across various situations.

Another important point to note is the very unequal sex distribution which might potentially have caused biases or make our findings less generalizable to men. First evidence regarding this comes from studies investigating coping styles and everyday stressors finding higher stress and more emotion-focused coping styles in women (Almeida and Kessler, 1998; Matud, 2004). On the other hand, however, effect sizes are low (Matud, 2004) and other studies find no gender differences (Porter et al., 2000).

Nevertheless, our results underpin the idea of situation-specific ERS-use. The present study advances the insufficiently explored interplay between emotion regulation, stress, and state rumination while emphasizing the importance of cognitive emotion regulation strategies in overcoming momentary rumination. Additionally, the findings challenge the previously distinct categorizations of emotion regulation strategies in the literature. Future research should abstain from the artificial investigation of the use of single ERS and instead focus on the use of multiple ERS in order to improve psychological theories and treatments. The categorization of cognitive ERS should be reconsidered as they might be aspects of one construct.