In the aftermath of exposure to potentially traumatic events, individuals frequently engage in counterfactual thinking (Olson et al., 2000; Roese et al., 2017). Counterfactual thinking is defined as mentally simulating a “what if” alternative to the reality of the past event (Roese et al., 2017). Individuals often revisit the past event by using conditional “if only” thinking to simulate what could or might have been (Dunning & Madey, 1995; Markman et al., 1993). Meta-analytical evidence suggests a positive association between counterfactual thinking and posttraumatic stress disorder (PTSD) symptoms (r = 0.46, k = 24; Hoppen et al., 2020). Recent theoretical work (Morina, 2021) and subsequent scale developments (Hoppen & Morina, 2021) aimed at advancing the understanding of counterfactual thinking by conceptualizing processes of posttraumatic counterfactual comparison (CFC). Specifically, counterfactual thinking involves implicit comparisons between the episodic memory of the event and a mentally simulated “what if” alternative to this event (e.g., “If only I had not gone to this gathering, the accident would not have happened”; De Brigard & Parikh, 2019). CFC, however, involves one further step, which requires thinking about the degree to which specific self-relevant dimensions (e.g., well-being) or possessions in the current situation would be different in a counterfactual world (e.g., “I would do much better at the moment, if only the accident had not happened”). Indeed, the frequency of engaging in CFC has been reported to be associated with PTSD symptoms and potentially play a role in the maintenance of this disorder (Hoppen & Morina, 2021; Morina, 2020). Understanding the process of CFC more thoroughly is thus of great relevance for psychological research and clinical practice (Morina, 2021). Accordingly, factors influencing the extent to which frequent CFC goes on to shape affective states need to be identified. To this end, the present research investigates whether affective styles (adjusting, concealing, tolerating) are related to CFC frequency and engendered affective valence.

Counterfactual comparisons

The functional theory of counterfactual thinking posits that this thinking type is an adaptive process serving goal striving and future action planning (e.g., identifying causal mechanisms that lead to different outcomes if a similar situation is encountered; Roese et al., 2017). However, the underlying cognitive processes may become maladaptive by resulting in excessive overthinking and ruminative regret (Roese et al., 2017). This is evinced in a positive association between counterfactual thinking and PTSD symptoms (Hoppen et al., 2020) and associations of CFC frequency with more PTSD symptoms and more self-blame (Hoppen & Morina, 2021). We posit that the CFC process can be better understood by examining whether the functionality and resulting affective valence of CFC differ as a function of (1) the comparison direction and (2) the reference to the actor. In terms of comparison direction, the mentally simulated alternative may be evaluated as more favorable than some current self-attribute (i.e., upward CFC). For instance, individuals may think that they would currently do better if the car accident had not happened. Alternatively, the mentally simulated alternative may be evaluated as less favorable than some current self-attribute (i.e., downward CFC). This may be the case when individuals think that they would currently feel worse had they injured somebody else in the car accident. Following traumatic events, upward counterfactual thinking tends to be more common than downward counterfactual thinking (Dalgleish, 2004; Roese, 1994; Sanna & Turley, 1996). The general comparative-processing model (gComp, Morina, 2021) suggests that the outcome of CFC (e.g., that the comparer would be doing better today if they had acted differently in the past) is processed with respect to its motivational significance, which then influences affective, cognitive, and behavioral reactions. The motivational significance of self-relevant upward and downward CFC differs. Accordingly, the motivational meaning of upward CFC is likely to be perceived as aversive (i.e., threatening the motives of the comparer) and the motivational meaning of downward CFC as appetitive (i.e., consonant with or challenging the motives of the comparer) in relation to the desired end state. Specifically, an upward CFC implies that the comparer would be doing better in the counterfactual world than they are doing currently. This outcome may be perceived as aversive as it signals that harm has taken place (i.e., the comparer is doing worse now) because of some significant event. On the other hand, the outcome of downward CFC that implies that the counterfactual world would be worse in comparison to current well-being is likely to be perceived as appetitive (Morina, 2021).

In terms of references to the actor, self-related CFC may have a different psychological impact relative to some counterfactual act referring to another person. According to theoretical accounts, referring to oneself during the cognitive process of upward CFC should be associated with self-blame, guilt, shame, lower self-esteem, feelings of incompetence, or a decrease of internal control beliefs (Foa et al., 1999; Mandel, 2003). Referring to another actor during upward CFC may be more strongly associated with a loss of interpersonal trust, anger, or a sense of helplessness (Foa et al., 1999; Mandel, 2003). Empirically, CFC frequency was associated with both more negative posttraumatic cognitions about the self and negative posttraumatic cognitions about others (Hoppen & Morina, 2021). Downward self-referent CFC may lead to pride or relief (“e.g., If I had not reacted fast enough during the accident and saved the child, I would do much worse at the moment”), while downward other-referent CFC may also lead to gratitude (e.g., “If that bystander had not helped me after the accident, I would do much worse at the moment”).

Affective styles

According to gComp, CFC is best understood as a process, in which self-relevant CFC is followed by negative or positive affect (Morina, 2021). Noteworthy, the more frequent self-relevant CFC is, the higher its overall affective impact will be. However, individuals differ in terms of how they regulate the engendered negative or positive affect following a CFC, which suggests that dispositional affect regulation styles may be implicated in this relationship. We refer to engendered affective valence in the context of CFC as the affective state that accompanies the respective CFC, whereas affect regulation styles are defined as dispositional strategies to regulate emotions. In line with cognitive PTSD theories (Ehlers & Clark, 2000; Schnyder et al., 2015), CFC may serve as a cognitive strategy to avoid thinking about trauma memories and the resulting aversive affect (Hoppen & Morina, 2021). Accordingly, different emotion regulation strategies may be centrally involved in this process. Indeed, maladaptive emotion regulation strategies emerged as a transdiagnostic variable across several mental disorders including PTSD (Aldao et al., 2010; Cludius et al., 2020; Gross et al., 2019). Literature identified several styles of affect regulation that may be relevant for the present research question. Here, we focus on the affective styles adjusting emotions, concealing emotions, and tolerating emotions (Hofmann et al., 2012).

Adjusting is defined as the capability to modulate unfolding negative emotions and adjust according to the situational demands to maintain psychosocial well-being (Hofmann & Kashdan, 2010; Totzeck et al., 2018). In the context of CFC, adjusting may be associated with a lower CFC frequency, especially for upward CFC. This is because upward CFC can be conceptualized as negative appraisal of some current self-attribute (e.g., well-being or appearance) that is accompanied by negative affect. In this regard, individuals with high levels of adjusting should be less likely to habitually engage in upward CFC, and more likely find adaptive ways to regulate the engendered affective valence effectively by adjusting to the situational demands. Accordingly, effectively adjusting affect as needed in the given context may be associated with lower CFC frequency and less negative affective valence.

Concealing represents a response-focused suppressive regulative strategy in an attempt of concealing and avoiding negative emotions (Hofmann & Kashdan, 2010; Totzeck et al., 2018). This affective style may play a crucial role in CFC given that CFC may serve as a means to avoid thinking about trauma memories and the resulting aversive affect (Hoppen & Morina, 2021). Altogether, emotional suppression constitutes a maladaptive regulation strategy that is associated with psychopathology (Aldao et al., 2010; Cludius et al., 2020; Gross et al., 2019). Thus, more concealing should be related to more negative affective valence after engaging in CFC. Also, concealing may be associated with a higher CFC frequency when individuals aim to avoid negative emotions and instead focus on CFC as a means to control perceived threat and affect associated with the respective event. In this vein, concealing may be associated with maladaptive frequent engagement in CFC, and this should especially account for upward CFC, which are perceived as aversive (Morina, 2021).

Tolerating represents an acceptance-based affective style (Hofmann et al., 2012). It appears that individuals engaging frequently in CFC have difficulties to maintain a tolerating or accepting attitude towards negative emotions resulting from the event. Therefore, a tolerating affective style should be associated with a lower CFC frequency because the affective impact of the event does not need to be avoided (Hoppen & Morina, 2021). As a consequence, the resulting affective valence after engaging in CFC should be more positive. Individuals with high levels of tolerating may be able to form more integrated contextual memories of the traumatic event (Ehlers & Clark, 2000; Schnyder et al., 2015). In this regard, they may be able to see the event as tragic and being part of one’s past, but not posing any imminent threat to the individual in the present moment.

Present study

Given the lack of research on CFC and affective styles, we examined whether affective styles (adjusting, concealing, & tolerating) are associated with CFC frequency and the engendered affective valence (i.e., impact of CFC on acute affect) in individuals with exposure to at least one potentially traumatic life event. To examine granulated aspects of these relationships, we conducted these analyses separately for (1) upward self-referent CFC, (2) upward other-referent CFC, (3) downward self-referent CFC and (4) downward other-referent CFC. To this end, we tested four a priori defined pathway models, in which CFC frequency was treated as a potential mediating variable between the three affective styles adjusting, tolerating, and concealing on CFC affective valence as a dependent variable (see Fig. 1). For upward comparisons, we expected that adjusting and tolerating would be negatively associated with CFC frequency but positively with more positive CFC affective valence. For concealing, we expected reverse relationships.

Fig. 1
figure 1

A priori specified pathway models with standardized coefficients, in which all pathways were specified. For visual interpretation, only significant pathways are depicted. Note. * p ˂ 0.05, ** p ˂ 0.01, *** p ˂ 0.001, ref = referent; CFC = counterfactual comparison

Full size image

Methods

Participants and procedure

This study was nested within a larger project focusing on the associations between trauma-related CFC and PTSD symptoms (Hoppen & Morina, 2021). In this larger project, the first measure of CFC, the posttraumatic counterfactual comparisons scale (P-CFC-S), was developed and validated and first results on the associations between trauma-related CFC and PTSD symptoms were reported. The present study focuses on a second (related though distinct) research question. Since the data for the present endeavours were collected as part of the larger project, the procedure was identical. The link to our study was disseminated online via psyweb (https://psyweb.uni-muenster.de), a non-commercial online panel for psychological research. Initially, 1,533 individuals took part in the study. Altogether, 977 individuals did not meet the inclusion criteria or did not provide informed consent and a total of 556 participants were included in the final sample. Inclusion criteria were exposure to at least one traumatic event according to the Life-Events-Checklist (Weathers et al., 2013), being fluent in the German language, having no acute suicidal ideation and no lifetime psychotic disorder. Participants were, on average, 51.01 years of age (SD = 14.28). Most participants reported to be female (62.95%), had high levels of education (76.60%, finished high school, 52.70% hold a university degree), were working (58.50% working full- or part-time), and in a relationship (67.30%). All materials, data, and the corresponding R code to reproduce the analyses can be found in the OSF (https://osf.io/rh27s/?view_only=7d4c3d9a8c4149f4882db8daaf7e8801).

Measures

Traumatic life events were assessed with the Life Events Checklist (LEC, Weathers et al., 2013). This scale was the premise for the inclusion criteria that participants experienced at least one traumatic event. The LEC assesses a total of 17 categories of traumatic events. Participants were asked to identify the index-trauma, and to state whether it involved violent or accidental death of someone else, life threat, sexual violence or serious injury. In addition, they were asked about single vs. repeated exposures to these events. The internal consistency of the LEC in the present analyses was α = 0.84.

Posttraumatic counterfactual comparisons were assessed using the Posttraumatic Counterfactual Comparisons Scale (P-CFC-S) (Hoppen & Morina, 2021). The P-CFC-S assesses the direction of comparison (i.e., upward vs. downward) and the reference of the actor in relation to the counterfactual event (i.e., self-referent vs. other-referent). Thus, this scale results in four distinct subtypes of CFC. Although the initial validation study suggested that CFC may be conceptualized as a unidimensional construct (Hoppen & Morina, 2021), we focus on the four different combinations and investigate our research questions on the item level. This was based on our rationale that we aimed to disentangle differential effects of upward self-referent CFC, upward other-referent CFC, downward self-referent CFC and downward other-referent CFC. In the present analysis, we focus on two questions for each type of CFC. First, the frequency of engaging in CFC during the last four weeks was assessed while referring to the index trauma as assessed with the LEC (see above). An example for upward self-referent CFC reads: “When considering your well-beingduring the past 4 weeks, how often did you think that you would do BETTER currently, if YOU had done something differently before/during/after the traumatic event?“. This question was assessed on a seven-point Likert scale ranging from 1 (“not at all”) to 7 (“very often”). When participants indicated that they had engaged in CFC as signified by a score above 1, they received a follow-up question tapping into CFC affective valence. The question reads: “How did you feel as a consequence of such thoughts during the past 4 weeks? (on average)”. This question was assessed on a scale ranging from 1 (“much worse”) to 7 (“much better”), whereby 4 represents a neutral middle category.

To assess affective styles, we applied the Affective Style Questionnaire (Hofmann & Kashdan, 2010). This questionnaire comprises twenty items assessing three dimensions of affective styles. The first dimension taps into a general ability to adjust one’s own affect to the situational demands and is coined adjusting (7 items, e.g., “I can get out of a bad mood very quickly”). The second dimension is defined as a tendency to conceal or suppress affect and is labeled concealing (8 items, e.g., “I often suppress my emotional reaction to things”). The third subscale taps into an accepting attitude toward emotions and is coined tolerating (5 items, e.g., “It’s ok if people see me being upset”). Items were assessed with a 5-point Likert scale ranging from 1 (“not true of me at all”) to 5 (“extremely true of me”). Internal consistencies were good for the concealing dimension (α = 0.85) and acceptable for adjusting (α = 0.78) and tolerating (α = 0.76).

Analysis procedure

Analyses were performed in R (R Core Team, 2020). We calculated descriptive statistics and intercorrelations among all constructs. To further examine the relationship between affective styles and posttraumatic CFC, we tested four a priori defined pathway models, in which CFC frequency was treated as potential mediating variable between the three affective styles adjusting, tolerating, and concealing on CFC affective valence as dependent variable. We conducted these models separately for (1) upward self-referent CFC, (2) upward other-referent CFC, (3) downward self-referent CFC and (4) downward other-referent CFC. Accordingly, we tested four different pathway models (see Fig. 1 for a visual representation of these models). All pathway models were estimated with the lavaan package in R using full-information maximum likelihood estimation (Rosseel, 2012). To test the statistical significance of the indirect effects in our models, we generated nonparametric confidence intervals using a bootstrap resampling procedure (Hayes, 2015). We generated 10,000 bootstrap resamples to estimate 95% confidence intervals (CIs) for the assessment of the indirect effects (Hayes, 2015).

Results

Table 1 shows descriptive statistics of all constructs along with gender differences in these constructs. Females reported higher upward self-referent CFC frequency and upward other-referent CFC frequency compared to males. Males reported more positive affective valence after engaging in upward self-referent CFC. Table 2 shows the intercorrelations of all included constructs among each other. Figure 1 depicts the a priori specified pathway models with standardized coefficients of significant effects.

Table 1 Descriptive statistics and gender differences
Full size table
Table 2 Intercorrelations among all constructs
Full size table

Upward self-referent CFC

In bivariate associations, more frequent upward self-referent CFC was associated with more frequent upward other-referent CFC, more frequent downward self-referent CFC, and more frequent downward other-referent CFC. Moreover, more frequent upward self-referent CFC was associated with more negative affective valence after engaging in all four types of CFC. More frequent upward self-referent CFC was further associated with lower levels of adjustment and tolerating, and higher levels of concealing.

In our pathway model, the affective styles tolerating and concealing were negatively associated with upward self-referent CFC affective valence. The affective style adjusting was positively associated with upward self-referent CFC affective valence. Upward self-referent CFC frequency was negatively associated with the affective styles adjusting and tolerating and positively with concealing (see Fig. 1). Moreover, higher upward self-referent CFC frequency was associated with more negative upward self-referent CFC affective valence. Yet, none of the indirect effects were significant. Only the indirect effect for adjusting and upward self-referent CFC frequency on upward self-referent CFC affect was close to being statistically significant, p = 0.053.

Upward other-referent CFC

In bivariate associations, more frequent upward other-referent CFC was associated with the frequency of the other three CFC and a more negative affective valence after engaging in upward other-referent CFC. More frequent upward other-referent CFC was further associated with lower levels of adjustment and tolerating, and higher levels of concealing. In the pathway model, upward other-referent CFC frequency was negatively associated with adjusting and positively with concealing. Moreover, upward other-referent CFC affective valence was negatively associated with upward other-referent CFC frequency (see Fig. 1). No other effects emerged in this pathway model referring to upward other-referent CFC.

Downward self-referent CFC

In bivariate associations, more frequent downward self-referent CFC were associated with the frequency of all three other CFC types and a more negative affective valence after engaging in downward self-referent CFC. More frequent downward self-referent CFC were also associated with lower levels of adjustment and tolerating. No significant effect occurred in the pathway model concerning downward self-referent CFC (see Fig. 1).

Downward other-referent CFC

In bivariate associations, more frequent downward other-referent CFC was associated with more negative affective valence after engaging in both types of self-referent CFC (upward & downward). In addition, more frequent downward other-referent CFC was associated with higher levels of concealing and lower levels of tolerating. In the pathway model, downward other-referent CFC frequency was positively associated with the affective style concealing. Further downward other-referent CFC frequency was negatively with the affective style adjusting.

Discussion

We investigated whether affective styles (adjusting, concealing, tolerating) play a role in the relationship between CFC frequency and resulting CFC affective valence. Analyses were conducted separately for (1) upward self-referent CFC, (2) upward other-referent CFC, (3) downward self-referent CFC and (4) downward other-referent CFC. The frequency of all four types of CFC correlated highly among each other, indicating a general tendency to engage in CFC in the aftermath of traumatic events, yet the frequency of downward CFC was lower than the frequency of upward CFC.

Upward self-referent CFC

In the pathway model, upward self-referent CFC affective valence was negatively associated with upward self-referent CFC frequency, tolerating, and concealing, as well as positively associated with adjusting. Upward self-referent CFC frequency was negatively associated with the affective styles adjusting and tolerating and positively with concealing. These results were mainly in line with our expectations concerning the main effects of affective styles and CFC frequency and engendered affective valence (Hofmann & Kashdan, 2010; Hoppen & Morina, 2021; Totzeck et al., 2018). However, we observed a sign flip in the relationship between CFC affective valence and the affective style tolerating. In bivariate associations, higher levels of tolerating were non-significantly yet positively associated with more positive affective valence, which we initially expected based on theoretical notions. This relationship was reversed in the pathway model where we found a significant negative effect, which may represent a spurious effect when the three affective styles are considered together. Although none of the mediation effects were significant, strong effects were found for adjusting and its relationship with upward self-referent CFC frequency and upward self-referent CFC affective valence. The resulting indirect effect was close to being statistically significant, hinting at the importance of adjusting in the relationship between upward self-referent CFC frequency and CFC affective valence. Thus, individuals with lower levels of adjusting reported both a higher upward self-referent CFC frequency and more negative CFC affective valence. This appears logical insofar as upward self-referent CFC are based on negative appraisals of the event and accompanying emotions (Hoppen & Morina, 2021). Individuals with a lower ability to use adjustment as an antecedent-focused strategy in the early phase of the emotion regulation process may thus more often engage in upward self-referent CFC, which results in more engendered negative affective valence (Hofmann et al., 2012). As expected, concealing was related to more negative affective valence after engaging in upward self-referent CFC. Also, concealing was associated with a higher upward self-referent CFC frequency. Individuals may aim to avoid negative emotions arising from the traumatic event and instead focus on CFC to control perceived threat and affect associated with the respective event (Morina, 2021). It is noteworthy that overall, most effects were found for upward self-referent CFC compared to the other possible combinations. This is in line with theoretical notions of the gComp model that upward counterfactual comparison represent an aversive state (Morina, 2021), and empirical findings that posttraumatic self-blame cognitions are associated with CFC frequency (Hoppen & Morina, 2021). This may result from changes in individuals with PTSD symptoms surrounding the belief system concerning the self in that one feels incompetent and worthless in the aftermath of traumatic events (Ehlers & Clark, 2000; Foa et al., 1999; Schnyder et al., 2015). In this context, engaging in upward rather than downward self-referent CFC may negatively influence perceptions of the self. More specifically, upward self-referent CFC may foster beliefs around deviations from the ideal self since they (per definition) involve judgments in which one’s own behavior is perceived as a violation against one’s own expectations or ideals (e.g., in terms of moral standards). Research indicates that such self-discrepancies are also influenced by the affective style adjusting (Schlechter et al., 2022).

Upward other-referent CFC

We found that a more negative upward other-referent CFC affective valence was associated with a higher upward other-referent CFC frequency and that upward other-referent CFC frequency was negatively associated with adjusting and positively with concealing. Although in bivariate associations upward other-referent CFC frequency and affective valence were associated with both affective styles, no mediation effects occurred in the pathway model. Based on the bivariate associations, our considerations concerning the affective styles and upward self-referent CFC may also account for the relationship with upward other-referent CFC. However, when controlling for each other in the pathway model, affective regulation may not play a central role for resulting CFC affective valence. This suggests that other variables surrounding interpersonal trust and a sense of helplessness play a more pivotal role (Ehlers & Clark, 2000; Foa et al., 1999; Schnyder et al., 2015). Dysfunctional appraisals about the trauma and its sequelae concerning other actors may surround beliefs about the world as a dangerous place (McNally & Woud, 2019).

Downward self-referent CFC

In terms of bivariate associations, we found that adjustment and tolerating were negatively associated with downward self-referent CFC frequency and positively with downward self-referent affective valence. Adjustment and tolerating may help individuals to process the event as an integrated contextualized experience (Ehlers & Clark, 2000) and thus reduce the need to engage in counterfactual thinking. However, no significant effect occurred in the pathway model when controlling for all variables. This indicates that affective styles are not centrally involved in the relationship between CFC frequency and affective valence in this model. Caution is thus necessary in any interpretation of the effects.

Downward other-referent CFC

We found in the pathway model downward other-referent CFC frequency to be positively associated with the affective style concealing and negatively with the affective style tolerating. Again, those who aim to conceal emotions related to the event may be more likely to engage in CFC as this may serve as a form of mental avoidance. Yet, avoiding trauma-related emotions hinders the integration of the traumatic event into autobiographic memory and thus increases the likelihood of frequently engaging in CFC as an attempt to process the event. In contrast, those with high levels of tolerating do not need to revisit the event anymore as they have established an accepting attitude toward the event and its sequalae.

Clinical implications

Clinical studies suggest that a reduction of dysfunctional appraisals precedes PTSD symptom reduction in trauma-focused cognitive behavioral therapy (Brown et al., 2019; Kleim et al., 2013; Schumm et al., 2015). Accordingly, focusing on CFC may be helpful in addressing the cognitive sequalae involved in psychopathology. In general, research indicates that maladaptive appraisals and repetitive negative thinking present a modifiable cognitive mechanism in the context of PTSD (Ehring, 2021; Woud et al., 2021). Accordingly, challenging dysfunctional CFC directly or modifying CFC by integrating the trauma memory into autobiographical memory and by focusing on the present reality and needs is advisable. Relatedly, it may be especially critical to focus on cognitions regarding self-blame (Kip et al., 2022). This study further indicates that therapeutic work on affective regulation may be necessary to deal with the trauma and its sequalae to help the patient refrain from engaging in CFC as a means of mental avoidance. This approach can be integrated into different trauma-focused therapies (Schnyder et al., 2015). Furthermore, mindfulness and acceptance-based therapeutic approaches may be used to assist in functional emotion regulation following CFC. For example, in acceptance and commitment therapy, aversive CFC can be defined and treated as a form of experiential avoidance (Hayes et al., 1999). Prospective research needs to systematically examine how to best implement these considerations into therapeutic work.

Limitations and further research

Some limitations to our study and potential implications are noteworthy. The study design was cross-sectional and does not allow for any causal interpretation. Our models were based on theoretical considerations that affective styles represent relatively stable tendencies and that these tendencies were already present before individuals engaged in CFC. While the traumatic events may have changed affective styles, which needs further testing in subsequent longitudinal studies, affective styles are likely to be involved in CFC frequency and engendered affective valence. Hence, we decided to test these proposed models. We acknowledge that other directionalities are feasible in this regard. Therefore, longitudinal research and prospective studies are warranted to allow for more solid conclusions. For instance, it could also be that downward CFC serve as a coping mechanism to regulate the engendered affective valence after engaging in upward CFC. Further, more research is needed to understand the differential effects on the different CFC types in terms of comparison directionality and actor more thoroughly. It also remains unclear whether the associations in the present study differ as a function of trauma type. It may have a different impact whether a traumatic event was man-made or not or whether individuals were exposed to the event once or repeatedly. In addition, more diverse samples are warranted to investigate the generalizability of our results.

Conclusion

The present study sheds light on the complexity of the relationship between CFC frequency and the engendered affective valence and the role of affective styles therein. We found differential effects of the affective styles depending on whether CFC was upward self-referent, upward other-referent, downward self-referent CFC or downward other-referent CFC. Most effects occurred for upward self-referent CFC pointing to the clinical importance of this comparison style, but more research is warranted to understand the underlying mechanisms.