The Role of Emotion Regulation in Distinct Measures of Emotional Resilience

Research conducted on emotional resilience has typically either conceptualised the construct as a trait or as an outcome of a dynamic process which is reflected in the different approaches to its measurement. First, earlier conceptualisations in the literature have defined emotional resilience as a stable trait or disposition that promotes positive adaptation against the negative emotional impacts of experiencing adverse events (Block & Kremen, 1996; Connor & Davidson, 2003). Research designs following this trait-based conceptualisation often use self-report measures to examine resilience, whereby scales with the best psychometric ratings included items typically focused on the measurement of protective factors thought to contribute to recovery from adversity (e.g., personal competence, acceptance, social support) or on one’s ability to ‘bounce back’ from stress (Windle et al., 2011). However, many theorists have criticised this conceptualisation and the way that it has been measured. Firstly, the static nature of trait-based conceptualisations limits the possibility that emotional resilience may vary moment-to-moment, across situations and further develop throughout the life span (Cahill et al., 2022; Rutter, 2012). Secondly, there appears to be a lack of ‘gold standard’ across the use of self-report measures raising concerns on the extent to which these scales are indeed measuring the true intended construct (Kalisch et al., 2017; Windle et al., 2011). Specifically, despite exposure to adversity being a core component to the conceptualisation of emotional resilience, many self-report scales do not adequately account for the variation in levels of adversity an individual may have experienced. Thus, it has been increasingly argued that these scales may be more reflective of one’s perception of their own resilience as opposed to a measure of the true intended construct (Britt et al., 2021; Notebaert et al., 2024).

To account for these concerns, an emerging consensus within the literature have begun to conceptualise emotional resilience as an outcome that is achieved through a dynamic process involving successful adaptation to adversity that is dependent on one’s context and resources rather than being a fixed attribute (Cicchetti & Blender, 2006; Kalisch et al., 2021; Rutter, 2012). Accordingly, this perspective has operationalised emotional resilience as the demonstration of more positive emotional outcomes than what is to be expected relative to the adversity experienced (Kalisch et al., 2021; Parsons et al., 2016).

Research designs following this operationalisation have indexed emotional resilience through a well-validated statistical approach, known as the residual approach (Booth et al., 2022; Cahill et al., 2022; Höltge & Ungar, 2022). This approach involves regressing a measure of emotional outcome onto a measure of experienced adversity and saving the residual scores as a proxy for each individual’s level of emotional resilience. These scores represent how much better or worse an individual has emotionally recovered relative to what would be predicted based on the adversity experienced. As positive emotional outcomes can manifest across various ways following experiences of adversity, this approach assumes a significant relationship between adversity type and post-adversity emotional outcome (Cahill et al., 2022). Specifically, according to a conceptual framework proposed by Kalisch et al. (2015), the selected measure must be sensitive to the expected changes in emotional outcomes for the type of adversity assessed within the given sample (see also Kalisch et al., 2021). For example, in line with this framework, researchers have indexed this outcome-based operationalisation in a sample of young adults by defining positive emotional outcomes as the absence of emotional difficulties following experiences of negative life events (Notebaert et al., 2024).

Moreover, in recognition of its dynamic nature, theorists have emphasised distinguishing between operationally defining emotional resilience as a distal outcome arising from positive emotional adaptation to multiple adversities over time (as per the residual approach), and emotional resilience as a proximal outcome following a single incident of acute stressor exposure (Bonanno & Diminich, 2013). For example, while transient perturbations in emotional functioning are expected during and immediately following an isolated event, individuals demonstrating high levels of emotional resilience demonstrate minimal to no lasting effects (Bonanno et al., 2015; Degering et al., 2023). Despite this recognition, research examining individual differences based on this operationalisation is lacking. Current literature has typically operationalised the construct in one particular way, either as a perceived trait via self-report scales or as an outcome of a process via the residual approach. This narrow approach limits the ability to compare candidate mechanisms across studies and across the different ways emotional resilience may be manifested (Luthar & Cicchetti, 2000; Luthar et al., 2000). Thus, to build a more comprehensive understanding of the cognitive factors underpinning individual differences in emotional resilience, future research must engage with and systematically compare these distinct operationalisations.

The cognitive process of emotion regulation has been proposed as one candidate mechanism underpinning why some individuals show more emotional resilience than others (Troy & Mauss, 2011). Emotion regulation is conceptualised as processes through which individuals modulate their emotions in response to environmental demands by deploying regulatory strategies to either modify the type of, or magnitude of, their emotional experience or of the emotion-eliciting event itself (Aldao et al., 2010). Examples of effective emotion regulation can be seen when an individual successfully reduces the intensity of their negative emotional experience through a process known as downregulation, or to increase the intensity of their positive emotional experience through a process known as upregulation (Gross, 2002, 2015).

Theoretical models conceptualise emotion regulation as a process that can be divided into a series of stages following exposure to an emotion-eliciting event (Bonanno & Burton, 2013; Gross, 2015; Sheppes et al., 2015). First, the need to emotionally regulate is identified, then a specific emotion regulation strategy is selected, implemented and its success monitored (Sheppes, 2020). Critically, as will be expanded below, individuals differ in their tendency to select specific emotion regulation strategies and the effectiveness of regulatory implementation, which has been described as two complementary, but discrete stages of emotion regulation (Rammensee et al., 2023). Given that adverse events are inherently emotional, an individual’s ability to effectively regulate their negative emotions during implementation may be a critical mechanism in supporting emotional well-being and enhanced emotional coping when exposed to stressful events (Kobylińska & Kusev, 2019). Conversely, the inability to effectively regulate emotions has been proposed to increase vulnerability to prolonged intense distress that can devolve into various psychopathologies (Aldao et al., 2010; Nolen-Hoeksema et al., 2008). A recent systematic review by Polizzi and Steven (2021) found substantial support across studies supporting a positive association between effective emotion regulation and emotional resilience. However, limitations relating to the use of self-report scales across studies for measuring emotional resilience and to infer effective emotion regulation were evident which warrant several considerations.

Firstly, whilst this review was comprehensive in examining articles defining emotional resilience across the different conceptualisations commonly used in the literature, majority of the studies included indexed emotional resilience using self-report questionnaires without measuring exposure to adversity, which as outlined above, may only provide evidence for operationalisations of this construct as a perceived trait. Secondly, most studies which directly assessed effective emotion regulation in this review also adopted this same approach (e.g., Banyard et al., 2017; Lee et al., 2019; Mestre et al., 2017; Polizzi et al., 2018; Poole et al., 2017; Powers et al., 2015; Vaughan et al., 2019). Self-report scales to infer effective emotion regulation tend to collapse different aspects of emotion regulation, such as perceived difficulties to successfully upregulate positive and/or downregulate negative emotions (e.g., DERS; Gratz & Roemer, 2004) and the perceived frequency of using strategies deemed as adaptive or maladaptive (e.g., ERQ; Gross & John, 2003), into a single global measure. However, as the effectiveness of emotion regulation can vary depending on the strategy deployed and the nature of the emotion-eliciting event (Gross, 2015), a critical issue with this approach is that individuals retrospectively report their behaviour without taking into account the decision-making process involved in strategy selection and the specific emotion-eliciting events being referred to when answering these questions (Sheppes, 2020).

Indeed, self-report questionnaires to examine emotional experience retrospectively have been associated with a range of biases (Barrett, 1997; Levine & Safer, 2002). Behavioural tasks that involve exposing participants to the same emotion-eliciting event and examining the difference in emotional experience before and after deploying a regulatory strategy may provide a more accurate indicator of effective emotion regulation (Sheppes, 2020; Troy et al., 2010). Despite this, studies adopting this approach to examine the relationship between effective emotion regulation and emotional resilience are lacking. The few studies that have examined effective emotion regulation using behavioural tasks have only examined its relationship with trait-based emotional resilience measures (e.g., Hildebrandt et al., 2016; Tugade & Fredrickson, 2004). Thus, whilst the review found substantial support for the relationship between effective emotion regulation and trait-based emotional resilience, more rigorous behavioural testing examining the relationship between effective emotion regulation and across the distinct operationalisations of emotional resilience is needed.

Now, we will consider the role of different emotion regulation strategies in relation to its effectiveness, and how the tendency to choose certain strategies may differentially relate to emotional resilience. As adverse life events can occur from a position where individuals are unable to change or avoid the negative emotion-eliciting event (e.g., death of a family member), cognitive strategies such as reappraisal and distraction that involve a change in emotional response to the situation, rather than a change to the situation itself, may be particularly useful in the context of adversity exposure. Reappraisal involves engaging with negative emotion by reinterpreting the meaning of an emotion-eliciting event to be more neutral or positive, whilst distraction involves disengaging from negative emotion by producing neutral thoughts unrelated to the emotion-eliciting event (Sheppes et al., 2011). However, as will be argued, these regulation strategies differ in their effectiveness for downregulation, which may influence how it is related to emotional resilience.

Historically, reappraisal has been considered the most effective emotion regulation strategy (Webb et al., 2012). As reappraisal involves reinterpreting the meaning of a negative emotional-eliciting event, such active engagement alters its subsequent emotional impact, and facilitates processing, evaluating, and remembering emotional information, all of which are essential for long-term coping (Folkman et al., 1986; Gross, 2002; Sheppes et al., 2014a, b). Empirical and meta-analytic evidence supports this, showing that reappraisal is associated with successful downregulation of negative emotions (Augustine & Hemenover, 2009; Wang et al., 2021; Webb et al., 2012), better mental health outcomes (Aldao et al., 2010; Gross & John, 2003) and greater levels of trait-based emotional resilience (Mestre, et al., 2017).

Additionally, the positive appraisal style theory of resilience (PASTOR; Kalisch et al., 2015), identifies reappraisal as a key mechanism underpinning emotionally resilient outcomes, as the ability to properly reappraise aversive situations involving appropriate adjustments of the initial negative appraisal and/or the generation of complementary positive appraisals, attenuates ongoing stress responses. Research has shown that tendency to use reappraisal can buffer the negative effects of adversity exposure. For example, in a sample of young adults, self-reported habitual use of cognitive reappraisal was associated with lower post traumatic and internalising symptoms following exposure to intense media coverage of a local terror attack and problems transitioning into university, respectively (Jenness et al., 2016; Zahniser & Conley, 2018). Together, these findings provide sound theoretical reasoning that greater reappraisal use is indeed associated with greater emotional resilience (Kalisch et al., 2015).

Conversely, while distraction can be effective for the downregulation of negative emotions, there is evidence suggesting that repeated use of distraction diminishes its efficacy (Kross & Ayduk, 2008; Paul et al., 2016; Thiruchselvam et al., 2011). Specifically, as distraction involves replacing negative emotional content in working memory with neutral content retrieved from long-term memory, such high cognitive cost can impair emotional processing of the triggering event, and lead to poorer future recall of the event (Richards & Gross, 2006; Sheppes & Meiran, 2007). As distraction does not address how emotional experiences are dealt with in the future, a habitual tendency to use distraction may thus be less effective than reappraisal for downregulation and more detrimental for long-term coping. For example, one study found that compared to instructions to simply attend to aversive stimuli, using distraction resulted in a significant increase in self-reported unpleasantness when participants were re-exposed to the same adversity within a single experimental session (Paul et al., 2016). Additionally, studies have demonstrated greater instructed and self-reported distraction use during stressful periods were significantly associated with a greater likelihood of developing depression in the future (Holahan et al., 2005; Kross & Ayduk, 2008).

Thus, given the effectiveness of reappraisal across situations and its benefits for long-term coping, one possible hypothesis is the Reappraisal Dominance Hypothesis. This hypothesis proposes that a greater tendency to choose reappraisal, relative to distraction, contributes to greater emotional resilience as reappraisal use leads to more effective negative emotion downregulation than distraction. One prediction arising from this hypothesis is that in a cross-sectional design, there will be an association between tendency to choose reappraisal over distraction and emotional resilience, which would be in part accounted for by effective downregulation. Thus, the first aim of the present study is to test this prediction across the distinct operationalisations of emotional resilience.

However, it is also possible that the proposed relationship between reappraisal choice tendencies and emotional resilience is more complex. The cognitive model of psychological resilience proposed by Parsons et al. (2016) posits that individuals showing emotionally resilient outcomes flexibly apply different cognitive processing strategies that are most conducive to current situational demands in achieving a desired goal. For instance, if applying a specific cognitive processing strategy in response to an adverse event does not promote the goal of effective coping with the elicited stress, these individuals will then adaptively change strategy to one that is more in line with achieving this goal. This perspective on the cognitive basis of emotional resilience is in line with more recent research conducted on emotion regulation flexibility, referred to as the ability to implement emotion regulation strategies that synchronize with contextual demands (Aldao et al., 2015). Specifically, studies by Sheppes and colleagues (2011, Sheppes et al., 2014a, b) examining emotion regulation choice in response to negative situations that varied in intensity demonstrated the emotion regulation asymmetry phenomenon whereby individuals tended to choose reappraisal for low intensity negative situations and distraction for high intensity negative situations.

The emotion regulation asymmetry phenomenon has been consistently replicated in the literature (Shafir et al., 2015, 2016; Van Bockstaele et al., 2020) and its effectiveness demonstrated in studies whereby instructed distraction use, relative to instructed reappraisal use, resulted in more effective downregulation of negative emotions for high intensity stimuli and equally effective for low intensity stimuli (Shafir et al., 2015; Sheppes et al., 2014a, b). Despite both strategies demonstrating equal levels of effectiveness in negative emotion downregulation for low intensity stimuli, choosing to deploy distraction relative to reappraisal in this context is suggested to be detrimental in promoting long-term mental health as it does not facilitate emotional challenges being dealt with in the future. So, choosing reappraisal for low intensity situations and distraction for high intensity situations, as is evident in the emotion regulation asymmetry phenomenon, may be the most effective pattern of emotion regulation strategy choice.

It is currently unclear whether a greater tendency to choose reappraisal relative to distraction in general, underpins variation in emotional resilience as this would lead to more effective downregulation, or whether greater reappraisal tendency alignment, underpins variation in emotional resilience, as this pattern of reappraisal tendency choice leads to more effective downregulation. Much of the current research within the emotion regulation literature has focused primarily on the effects of self-reported habitual use of reappraisal or distraction on mental health outcomes (Aldao et al., 2010; Holahan et al., 2005; Kross & Ayduk, 2008), or on people’s tendency to choose one strategy over the other when exposed to varying emotional contexts (). However, despite suggestions made regarding the critical role of reappraisal as a protective factor against negative emotional outcomes under stressor exposure (Kalisch et al., 2015; Riepenhausen et al., 2022) and a growing consensus that flexible emotion regulation strategy choice is crucial in promoting adaptive coping following stressor exposure and better mental health outcomes (Aldao et al., 2015; Bonanno & Burton, 2013; Kashdan & Rottenberg, 2010), the relationship of either possibility underpinning variation across the distinct operationalisations of emotion resilience remains untested.

In summary, the current study was designed to examine associations between reappraisal choice tendencies and individual differences in emotional resilience, to test the cross-sectional predictions generated by two key hypotheses. The Reappraisal Dominance Hypothesis proposes that a greater tendency to choose reappraisal relative to distraction, contributes to greater emotional resilience as reappraisal use leads to more effective downregulation than distraction. Second, the Reappraisal Tendency Alignment Hypothesis proposes greater tendency to choose reappraisal aligned to the emotion regulation asymmetry phenomenon contributes to greater emotional resilience, as this pattern of emotion regulation choice leads to more effective downregulation.

To examine these hypotheses, the current study design focuses on a sample of emerging adults attending university as this sample reflects a critical transitional stage of development. This stage includes many positive but also stressful challenges involved in establishing future trajectories in relation to educational, occupational, and social attainments uniquely central to this cohort (Arnett, 2003, 2007; Sandhu, 1994). Furthermore, as most mental health disorders emerge by emerging adulthood, this sample reflects a crucial time for effective interventions in preventing risks and promoting positive development (Arnett, et al., 2014; Catalano & Kellogg, 2020). To assess emotional resilience measured to reflect distinct operationalisations, indices of past adversity, current emotional difficulties and perceived trait resilience were obtained, followed by exposure to a standardised lab-based adverse event. This allowed for the derivation of three different indices of emotional resilience: a Depression Anxiety and Stress Scale (DASS) Residual Index reflecting outcome-based operationalisations consisting of less emotional difficulties than expected relative to past adversity experienced (Notebaert et al., 2024), a Brief Resilience Scale (BRS) Index reflecting trait-based conceptualisations of perceived emotional resilience (Windle et al., 2011), and a Lab Based Stressor Recovery Index reflecting a transient measure of emotional resilience (Bonanno & Diminich, 2013) (see Methods). Participants completed an Emotion Regulation Assessment Task in which they were exposed to stimuli of high and low intensity and chose either reappraisal or distraction to downregulate negative emotions elicited by these stimuli (Sheppes et al., 2011). Participants also rated the intensity of their negative emotions across each trial, which rendered a measure of the effectiveness of their downregulation attempts.

The Reappraisal Dominance Hypothesis generated the following predictions. First, the tendency to choose reappraisal over distraction will be positively associated with emotional resilience. Secondly, downregulation effectiveness will partially account for the statistical relationship between the tendency to choose reappraisal over distraction and emotional resilience. Conversely, the Reappraisal Tendency Alignment Hypothesis generated the following predictions. First, the tendency to choose reappraisal aligned to the emotion regulation asymmetry phenomenon (i.e., reappraisal tendency alignment) will be positively associated with emotional resilience. Secondly, downregulation effectiveness will partially account for the statistical relationship between reappraisal tendency alignment and emotional resilience. To comprehensively examine each hypothesis, these predictions were tested on each measure of emotional resilience reflecting distinct operational definitions¹ entered as the dependent variable.

Participants were recruited from a pool of undergraduate Psychology students from The University of Western Australia and Curtin University in exchange for course credit. A total of 205 participants were recruited, however 12 participants did not complete the study and a further 92 participants were excluded from tests of hypotheses (see Results) for not meeting the experimental requirements. The final sample included 113 participants, with 80 identifying as female, 32 as male and 1 as non-binary, and an age range of 17 to 26 (M = 19.95, SD = 1.96). Participants’ ethnicity within this sample included 50% Oceanian, 19% South-East Asian, 13% North-West European, 7% Southern and Central Asian, 5% North-East Asian, 3% North African and Middle Eastern, 2% Sub-Saharan African and 1% Southern and Eastern European.

Following the inverse square root method for assessing statistical power (Kock & Hadaya, 2018), this sample size was deemed sufficient to detect the predicted direct and indirect associations between reappraisal choice tendencies and emotional resilience (N_min = 69). Due to the novelty of the effects under test and an aim to detect effects of practical significance, this calculation was made based on the upper boundary of a moderate effect size (β_min = .30) with 80% power at a significance level of 0.05.

The current study was approved by the University of Western Australia’s Human Research Ethics Office (reference ET000074). Data was collected across two sessions. All participants provided informed consent prior to the commencement of each session. Participants completed both sessions privately in a small room with one computer.

In the first session, participants completed a demographics questionnaire, the NLESS-25, the DASS-21, the PANAS and the BRS, followed by delivery of the standardised adverse event. Participants then returned for a second experimental session within two weeks of attending the first session. In the second session, participants first viewed four example images obtained from the IAPS as an example of the type of content they would be exposed to. Informed consent was collected again before participants proceeded with the Emotion Regulation Assessment Task. Participants then completed the assessment of experimental task adherence, which marked the end of the experimental study. Participants were then thanked and debriefed about the purpose of the experiment.

As the total testing duration of the experimental study would take approximately two hours to complete with a heavy burden on participants, and to avoid any carry-on effects of induced negative emotions from the standardised stressor task on responses to the Emotion Regulation Assessment Task and vice versa, it was decided it would be appropriate to divide the experiment across two sessions. In deciding which session was to be completed first, consideration was given towards an order that would minimise participant drop-out. As the Emotion Regulation Assessment Task has been associated with high drop-out rates in the past, we collected emotional resilience measures in the first session and measures of emotion regulation in the second. Note that the ordering of these two sequences does not carry implications around causality as this is functionally a cross-sectional design. Thus, given that the aim of the current study was to test for cross-sectional associations that were predicted by the proposed causal hypotheses, the division of testing sessions in this specific order was based on practical convenience. All analyses were conducted using IBM SPSS Statistics for Windows (v29.0.2.0, 2023) except for the test of hypotheses which were analysed using R Statistical Software (v4.3.3; R Core Team, 2024). Specifically, all predictions were tested via Partial Least Squares Structural Equation Modelling (PLS–SEM) in the SEMinR package (v2.3.4; Ray et al., 2024) as this method is less constrained to assumptions of causality and temporal ordering than other tests involving an explanatory variable (Sarstedt et al., 2020).

We report this study following JARS, including how we determined our sample size, all data exclusions, all manipulations, and study measures (Cooper, 2008). The raw dataset has been made publicly available on the Open Science Framework. This study was not pre-registered.

The residual approach to calculating the DASS Residual Index rests on the assumption that number of past exposures to adversity is positively associated to level of current emotional difficulties (Kalisch et al., 2015; McLaughlin et al., 2010). As assumed, total NLESS scores and total DASS scores were significantly and positively correlated, r = .257, p = .002, permitting calculation of each participant’s DASS Residual Index score. Total DASS scores were first regressed on to total NLESS scores, yielding a significant model, F(1, 147) = 10.36, p = .002, R² = .066, which confirmed experiencing past adversity significantly predicted an increase in emotional difficulties (B = 1.30, SE = 0.40, t = 3.22, p = .002, 95% CI = [0.50, 2.10]). The residual scores were then saved and reverse scored to represent each participant’s DASS Residual Index, where higher scores indicate fewer emotional difficulties than expected relative to the adversity experienced, thus reflects greater levels of the outcome-based measure of emotional resilience.

To permit calculation of each participant’s Lab Stressor Recovery Index, it was first confirmed that participants’ State-Anxiety scores significantly increased following exposure to the standardised stressor task (see Supplementary Materials for details of analysis and results). Participants’ Lab Stressor Recovery Index was then derived by calculating the amount of perseveration in state anxiety following exposure to the standardised stressor task. Given that reactivity and recovery has been shown as dissociable dimensions of a stress response and may have distinct underlying mechanisms (Degering et al., 2023; Linden et al., 1997), coupled with conceptualisations of emotional resilience that emphasise this construct as an outcome that can only be observed following stressor exposure (Bonanno & Diminich, 2013; Bonanno et al., 2015), only participants’ State-Anxiety scores measured during the recovery period were considered when deriving each participant’s Lab Stressor Recovery Index score. This index thus permits examination of mechanisms associated with individual differences in stress recovery as a separate, independent dimension from stress reactivity.

As per Rudaizky and MacLeod (2014), this was calculated as the mean in State-Anxiety scores measured at mid-recovery and end-recovery assessment points, where higher scores reflect greater preservation of state-anxiety following completion of the standardised stressor task. The mean State-Anxiety score measured across mid-recovery and end-recovery assessment points were subsequently inversed to derive participants’ Lab Stressor Recovery Index score. Higher Lab Stressor Recovery Index scores indicate less perseveration of state-anxiety following exposure to the standardised adversity delivered in the lab, thus reflects greater levels of a transient measure of emotional resilience operationalised as a dynamic construct.

To evaluate the relationship between measures of emotional resilience reflecting distinct operationalisations, the correlation coefficients between measures were inspected. There is considerable consensus within the literature that a correlation at or above .80 between two distinct measures indicate they may be measuring synonymous constructs (Hodson, 2021). Pearson correlation analyses with 95% bootstrapped confidence intervals set at 1000 replications suggested that the DASS Residual Index appeared distinct from the Lab Stressor Recovery Index, r(147) = .34, p < .001, 95% BCI = [.19, .47], and distinct from the BRS Index, r(147) = .47, p < .001, 95% BCI = [.33, .59]. The Lab Stressor Recovery Index also appeared distinct from the BRS Index, r(147) = .24, p = .003, 95% BCI = [.08, .39]. These findings suggest each measure reflected distinct operationalisations of emotional resilience, and thus permit analyses and comparison of the proposed models on each measure separately entered as the dependent variable.

The Reappraisal Tendency Alignment hypothesis assumes that most participants would demonstrate the emotion regulation asymmetry phenomenon. Specifically, this hypothesis assumes that in the Choice block, participants would demonstrate a greater tendency to choose reappraisal for low intensity trials and distraction for high intensity trials.

To examine this assumption, the proportion of trials participants chose each strategy in the Choice block was subjected to a 2 × 2 repeated measures ANOVA with within-subjects factor of Strategy (Reappraisal vs. Distraction) and Intensity (Low vs. High). As predicted and depicted in Fig. 2a, there was a significant Strategy × Intensity interaction, F(1,112) = 175.52, p < .001, η_p² = .610. In line with previous research (Shafir et al., 2015; Sheppes et al., 2011; Van Bockstaele et al., 2020) follow-up paired-samples t-tests revealed participants chose reappraisal significantly more in low intensity trials (M = .66, SD = .25) than in high intensity trials (M = .29, SD = .20), t(112) = 13.24, p < .001, d = 1.25, and distraction significantly more in high intensity trials (M = .71, SD = .20) than in low intensity trials (M = .34, SD = .25), t(112) = 13.25, p < .001, d =  − 1.25. Thus, the assumption based on the emotion regulation asymmetry phenomenon underpinning the Reappraisal Tendency Alignment Hypothesis was met.

To explore participants negative emotion downregulation effectiveness when instructed to use each of the strategies to downregulate from high and low intensity images, participants’ instructed downregulation effectiveness scores derived from trials at each intensity level in the reappraisal and distraction sub-blocks of the Instruct block were examined. Instructed Downregulation Effectiveness scores were subjected to a 2 × 2 repeated measures ANOVA, with within-subjects factor of Strategy (Reappraisal vs. Distraction) and Intensity (Low vs. High). As depicted in Fig. 2b, there was a significant interaction of Strategy × Intensity, F(1, 112) = 26.29, p < .001, η_p² = .190. Follow-up paired samples t-test revealed that for high intensity trials, participants were significantly more effective at downregulating when instructed to use distraction (M = 1.75, SD = 1.33) than when instructed to use reappraisal (M = 1.28, SD = 1.14), t(112) = 4.97, p < .001, d = .47. For low intensity trials, there was no significant difference in the effectiveness of downregulation when instructed to use distraction (M = 1.31, SD = 0.93) compared to reappraisal (M = 1.24, SD = 1.00). Additionally, when examining the effectiveness of each strategy at each level of intensity, participants were significantly more effective at downregulating when instructed to use distraction in high intensity trials than in low intensity trials, t(112) = 4.86, p < .001, d = .46, but did not differ when instructed to use reappraisal in low intensity trials compared to high intensity trials, t =  − .46, p > .05.

Descriptive statistics and inter-correlations pertaining to all subsequent analyses are presented in Table 2. To test the predictions generated by both hypotheses under scrutiny required evaluation of direct and indirect associations between reappraisal choice tendencies, downregulation effectiveness and measures of emotional resilience reflecting distinct operationalisations. Therefore, a series of PLS–SEM analyses were tested. Additionally, exploratory analyses examining whether the predicted associations posited by both hypotheses would be conditional upon the extent to which individuals achieved effective downregulation when implementing reappraisal and/or distraction in general, or for specific stimuli was also conducted. Results of these exploratory analyses conducted on each emotional resilience measure are reported in Supplementary Materials. All models were bootstrapped with 1000 repetitions. Evidence of statistical significance for each pathway was indicated when bootstrapped confidence intervals did not pass through zero.

Despite its strengths, our study is not without limitations. Firstly, the current study was cross-sectional by design, as no variables were directly manipulated nor repeatedly measured. Thus, the results of the PLS–SEM analyses provide correlational evidence only and no inferences about the causal role and causal ordering of reappraisal choice tendencies as a candidate mechanism underpinning emotional resilience can be made. Future research seeking to elucidate causality and directionality between reappraisal choice tendencies and emotional resilience via downregulation effectiveness should adopt longitudinal cross-lagged panel designs involving repeated measures across sessions. Alternatively, future studies may seek to directly manipulate reappraisal choice tendencies and examine whether such manipulation directly influences differences in emotional resilience.

Secondly, the current sample consisted of young undergraduate students recruited from a Western, educated, industrialised, rich, and democratic (WEIRD) society (Henrich et al., 2010). Whilst our aim was to examine emotional resilience shown by young adults, the findings of this study may not be broadly applicable due to the homogeneity of the sample, which limits generalizability across diverse cultural contexts. Moreover, given that emotional resilience has shown to exhibit sensitivity to cultural and contextual influences (Ungar, 2008), future studies may wish to focus on addressing these limitations by adopting culturally sensitive methodologies and exploring intercultural differences in emotional resilience. For example, cultural-specific definitions and demonstrations of what constitutes as healthy emotional functioning following adversity, and assessments of culturally relevant stressors specific to the study’s target population should be considered (Blessin et al., 2022). Addressing these limitations and advancing emotional resilience research in this direction will aid in developing effective and culturally sensitive interventions aimed to promote emotional resilience across diverse populations.

Finally, while the study achieved strong experimental control, the study design was limited in achieving a high degree of ecological validity. As previously stated, the Emotion Regulation Assessment Task constraints the participants to select a strategy from two regulatory options to downregulate whilst viewing negative images. It is possible that the way individuals responded in this task is not a true reflection of how they may respond in real life. Future studies aiming to enhance ecological validity may benefit from examining emotion regulation choice tendencies in more naturalistic settings such as through ecological momentary assessments (see Boemo et al., 2022).

The current study adds to the existing literature by examining the contribution of emotion regulation, specifically reappraisal choice tendencies, as a potential mechanism underpinning individual differences across three distinct operationalisations of emotional resilience. In line with the PASTOR framework (Kalisch et al., 2015), our findings indicate that greater tendency to choose reappraisal over distraction was associated with greater emotional resilience operationalised as a transient state and a perceived trait. Additionally, our findings support theoretical models of emotional resilience as a multi-dimensional construct, suggesting that different mechanisms may relate to distinct operationalisations of emotional resilience. Critically, the cross-sectional design of the current study limits inferences concerning causality and directionality, thus future research aiming to replicate and extend on these findings are warranted. Nevertheless, our findings highlight the role of reappraisal use as a potential candidate mechanism underpinning emotional resilience and the need for future research investigating individual differences to engage with and compare the distinct operationalisations of emotional resilience. By doing so, a more comprehensive understanding of this construct can be achieved which can ultimately lead to the development of more targeted interventions for promoting emotional resilience in young adults.