Introduction

Self-conscious emotions are a special class of emotions that involve people's reactions to their own characteristics and behavior [1]. Shame and guilt are negatively valenced self-conscious emotions, typically experienced in situations of failure or in which behavioral standards are violated [1, 2]. Although they are elicited by similar types of situations, shame and guilt differ in terms of how individuals appraise transgressions or errors and in terms of the action tendencies they elicit [1, 3]. Shame usually involves negative evaluations of the global self, and is accompanied by a sense of inferiority and worthlessness, and the desire to escape or hide [1]. Guilt, on the other hand, involves the negative evaluation of a specific behavior, is characterized by remorse and regret over the bad thing done, and motivates reparative behavior [1].

Normative levels of shame and guilt are functional and serve social goals [2]. They have been linked to empathy towards others [4, 5], prosocial behavior [6–8], self-improvement motivation and behavior [9], and lower levels of aggression and antisocial behavior [10, 11]. However, when shame and guilt are disproportionate to the situation, and/or triggered very easily and experienced in a variety of contexts, becoming the dominant way of emotional responding, they can be problematic [2]. Studies on children and adolescents have linked shame-proneness to negative outcomes such as anxiety and depression [12–14], eating disorders, delinquent behavior and substance use (for review see [2, 15, 16]). Data regarding the association between guilt-proneness and psychological problems are less consistent [2]. They seem to indicate that guilt over specific behaviors is not associated with poor psychological adjustment [1, 17], and that guilt becomes maladaptive when it is fused with shame, when people develop a distorted sense of responsibility for events beyond their control, and when opportunities for reparation are blocked [1, 18, 19].

In order to be elicited, shame and guilt require self-awareness, the ability to form stable self-representations, to reflect on those representations, and to generate self-evaluations [20]. Being the product of complex cognitive operations, they emerge later in development than basic emotions, and are gradually shaped and refined through childhood and adolescence [10, 21]. So far, research on the development of self-conscious emotions has mainly focused on childhood [22, 23], and relatively little is known about their course across other life periods [24]. Existing data suggest that adolescence may be a particularly salient time in the development of shame-proneness and guilt-proneness. A previous study has shown that shame-proneness decreases from adolescence to middle adulthood [24]. However, another study reported that shame temporarily decreases in 10th grade adolescent girls and increases in 10th grade adolescent boys compared to 8th graders, but these changes are reversed in first year college students [25]. In which guilt-proneness is concerned, there seems to be a steady increase from adolescence to old age [24, 25]. Clearly, additional studies are needed in order to characterize age- and sex-related changes in shame-proneness and guilt-proneness in adolescence.

Several studies have also sought to understand the influence of childhood trauma on dispositional shame and guilt and found that neglect is associated with higher shame-proneness, but not guilt-proneness in children [26] and adults [19, 27]. Similarly, a recent longitudinal study has reported that harsh parenting in childhood is related to increased shame-proneness, but not guilt-proneness in adolescence [28]. Other childhood traumatic events such as parental conflict and sexual abuse were not associated with proneness to shame and guilt [28, 29]. Another recent study showed that shame-proneness may be increased in adolescents with a history of serious illness or injury [29]. Research focusing on situational shame and guilt has also documented their relation to childhood trauma. For instance, Alessandri and Lewis [30] found that maltreated children show higher levels of shame when they fail on a task, and Donatelli, Bybee, and Buka [12] found that adolescents whose mothers have a history of depression report more guilt over failing to meet maternal expectations. Overall, evidence on the impact of childhood trauma on shame and guilt in adolescence is heterogeneous, and this issue needs further clarification [17]. Crucially, studies on childhood trauma and shame and guilt need to control for traumatic intensity in order to ascertain that exposure to a childhood stressful event has a substantial negative impact on personality and life course [31], while also distinguishing between dispositional (i.e., proneness to shame and guilt) and domain or situation-specific shame and guilt.

Recent research suggests that the long-term influence of childhood trauma on shame-proneness and guilt-proneness in adolescence may involve other individual differences [28, 29]. One obvious candidate is emotion regulation, considering that it undergoes major maturational changes during adolescence (e.g., [32]), and plays a central role in emotional adaptation and risk for psychopathology (e.g., [33]). Adolescence may be characterized by changes both in the habitual use of emotion regulation strategies and the efficiency of these strategies, as reflected in their relations with emotional difficulties [34]. To our knowledge, there is only limited evidence regarding the links between emotion regulation and proneness to shame and guilt. For example, a recent study [35] has found that higher use of suppression (i.e., inhibiting emotional expressions) is associated with increased shame-proneness, whereas higher use of reappraisal (i.e., changing the meaning of a situation) is associated with increased guilt-proneness in adolescence. These results suggest that the preference for maladaptive emotion regulation strategies, which are less efficient in reducing negative affect (e.g., suppression), may be related to shame-proneness, whereas preference for adaptive, more efficient strategies (e.g., reappraisal) may be related to guilt-proneness. Indeed, emotion regulation efficiency (i.e., impulse and anger control; tendency to down-regulate negative affect and up-regulate positive affect; appropriateness of emotional responses relative to the situation) was found to be negatively correlated with shame-proneness, and positively correlated with guilt-proneness [35, 36]. However, adolescents use many emotion regulation strategies when facing negative events (e.g., [37]) and the analysis of the relations between habitual emotion regulation and proneness to shame and guilt should be extended. Research shows that other dimensions of negative affect (e.g., depression, anxiety) are positively associated with higher use of emotion regulation strategies such as rumination, self-blame and catastrophizing, and negatively associated with higher use of strategies such as positive refocusing and positive reappraisal [34, 37], but the links between these strategies and proneness to shame and guilt have not been examined until now.

The present study investigated the independent contributions of age, sex, childhood trauma and individual differences in emotion regulation to shame-proneness and guilt-proneness in a large community sample of 13 to 17-year-old adolescents. Following recommendations in the field [28, 38], we used derivate measures of shame-proneness and guilt-proneness, which control for the interrelations between these dispositional dimensions. In addition, we focused on childhood negative events that were perceived as highly traumatic, in order to reduce heterogeneity in stressor intensity [31]. Moreover, we linked shame-proneness and guilt-proneness with depression and anxiety symptoms, to highlight their contribution to vulnerability to psychopathology [28]. Finally, individual differences in emotion regulation were assessed using a multidimensional scale that captures the habitual use of a wide range of adaptive and maladaptive cognitive emotion regulation strategies [37].

Materials and Methods

Results

Age, Sex and Emotional Symptoms

First, we investigated the relations between age and the habitual use of emotion regulation strategies. Age was positively associated with CERQ Rumination (two-tailed r = 0.13, p = 0.001). There were no other significant correlations between age and CERQ emotion regulation scores. Moreover, the correlations between age and DASS depression (two-tailed r = 0.06, p = 0.090) and anxiety scores (two-tailed r = 0.03, p = 0.313) were not significant.

Table 1 shows the correlations between emotion regulation and emotional symptoms. CERQ Self-Blaming, Acceptance, Rumination, Catastrophizing and Blaming Others scores correlated positively with both DASS depression and DASS anxiety scores. In addition, CERQ Putting into Perspective was positively associated with DASS anxiety. In contrast, CERQ Positive Refocusing, Refocus on Planning and Positive Reappraisal scores correlated negatively with DASS depression scores.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Correlations between emotion regulation and depression and anxiety symptoms.

https://doi.org/10.1371/journal.pone.0167299.t001

Using Student t-tests for independent samples, we also checked for sex differences in emotion regulation and emotional symptoms. Girls reported significantly higher levels of CERQ Putting into Perspective (t[644] = 2.25, p = 0.024) and significantly lower levels of CERQ Other Blaming (t[644] = -3.01, p = 0.003) in comparison to boys. There were no other significant sex differences in the other CERQ emotion regulation scores or in DASS depression and anxiety scores (all ps> 0.105).

Predictors of Shame-Proneness and Guilt-Proneness

Two separate regression analyses focused on shame-proneness residual scores and guilt-proneness residual scores as outcomes, and age and sex (Step 1), childhood trauma (Step 2) and CERQ emotion regulation (Step 3) as predictors. Sex (boys = 0; girls = 1) and childhood trauma (no traumatic event = 0; one or more traumatic events = 1) were dummy coded. Tolerance (all ≤ 0.99) and variance inflation factor (all ≤ 2.21) coefficients did not indicate multicollinearity. Given that two regression models were tested, a Bonferroni-corrected threshold of statistical significance (p/2 = 0.025) was adopted for these analyses.

The model in which shame-proneness was used as outcome was not significant in Step 1 (F[2, 637] = 1.34, p = 0.262) and Step 2 (F[3, 636] = 0.90, p = 0.439), which indicated that neither age and sex, nor the history of childhood trauma were significantly related to shame-proneness (Table 2). The model became significant (F[12, 627] = 15.60, p< 0.001) in Step 3, after CERQ emotion regulation scores were added, and accounted for an additional 22.57% of shame-proneness (F-change[9, 627] = 20.41, p< 0.001). As shown in Table 2, CERQ Self-Blaming, Positive Refocusing and Catastrophizing scores were significant positive predictors of shame-proneness, whereas CERQ Refocus on Planning and Positive Reappraisal scores were negative predictors of shame-proneness.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Coefficients from the multiple regression in which shame-proneness was regressed on age and sex, childhood trauma and individual differences in emotion regulation.

https://doi.org/10.1371/journal.pone.0167299.t002

The model in which guilt-proneness was used as outcome was not significant in Step 1 (F[2, 637] = 3.18, p = 0.042). Neither age, nor sex was significantly related to guilt-proneness (Table 3). The model became significant (F[3, 636] = 5.56, p = 0.001) in Step 2, after adding the history of childhood trauma as predictor, and accounted for an additional 1.57% of the variance of guilt proneness (F-change[1, 636] = 10.22, p = 0.001). The history of childhood trauma was a significant positive predictor of guilt-proneness. The model remained significant (F[12, 627] = 14.59, p< 0.001) in Step 3, after entering CERQ emotion regulation scores as predictors, and accounted for an additional 19.27% of the variance of guilt-proneness (F-change[9, 627] = 17.17, p< 0.001). CERQ Refocus on Planning and Positive Reappraisal scores were significant positive predictors of guilt-proneness, and CERQ Positive Refocusing, Catastrophizing and Blaming Others scores were negative predictors of guilt-proneness.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Coefficients from the multiple regression in which guilt-proneness was regressed on age and sex, childhood trauma and individual differences in emotion regulation.

https://doi.org/10.1371/journal.pone.0167299.t003

Discussion

The present results show that individual differences in emotion regulation are related to shame-proneness and guilt-proneness in adolescents, independent of the influence of age, sex and childhood trauma. Childhood trauma was related to guilt-proneness, but it explained only a relatively minor portion of its variance compared to emotion regulation. This study also indicates that the habitual use of rumination increases with age and described positive and negative associations between emotion regulation strategies and emotional symptoms in adolescence.

Individual differences in the habitual use of emotion regulation strategies explained around 20% of both shame-proneness and guilt-proneness. Remarkably, largely the same emotion regulation strategies contributed to both emotional dispositions, but with opposing roles. That is, we found that the more often one thinks about positive issues instead of confronting negative events (i.e., Positive Refocusing) and thinks about how bad negative events are (i.e., Catastrophizing), the higher s/he is on shame-proneness and the lower on guilt-proneness. In addition, the habitual use of putting the blame on oneself for negative events (i.e., Self-Blaming) is associated, as one would expect, with higher shame-proneness, whereas putting the blame on the situation or on other people (i.e., Blaming Others) is related to lower guilt-proneness. Extending the contrast, tendencies to confront negative events by taking the necessary steps to handle the situation (i.e., Refocus on Planning) and to look for a positive meaning of negative events (i.e., Positive Reappraisal) are linked with lower shame-proneness, and higher guilt-proneness in this study. Overall, the present results indicate that maladaptive emotion regulation strategies, such as Catastrophizing, Self-Blaming and Blaming Others, which have been associated with depression and anxiety symptoms in previous studies (e.g., [49, 50]) as well as in the present sample, are also related to shame-proneness, whereas low levels of these strategies are associated with guilt-proneness. In contrast, adaptive emotion regulation strategies such as Positive Reappraisal and Refocus on Planning, which are known to mitigate against symptoms of depression and anxiety (see [49] and this study), worry and fearfulness [50] are associated with reduced shame-proneness and higher guilt-proneness. Positive Refocusing is an exception to this pattern considering that, while it is viewed as an adaptive emotion regulation strategy based on its negative relation to depression symptoms (see [50] and this study), it has been associated with enhanced shame-proneness and reduced guilt-proneness in the present study. This relation may seem less surprising if we consider that both Positive Refocusing and shame involve disengagement from unpleasant situations, one by distraction with the help of positive thoughts about unrelated matters [37] and the latter, by an intrinsic desire to escape or hide in negative situations [1]. Importantly, given the correlational and cross-sectional nature of the present study, the direction of the relations between emotion regulation and proneness to shame and guilt cannot be identified. While the influence of emotion regulation on dispositional shame and guilt is more plausible considering evidence from prospective studies (e.g., [51]), which showed that emotion regulation predicts subsequent emotional adjustment and not the other way around, this study cannot rule out alternative models in which dispositional shame and guilt drive habitual emotion regulation or they influence each other.

The present results also show that guilt-proneness is increased in adolescents with a history of childhood trauma. Previous studies have reported that neglect [26], harsh parenting [28] and severe illness or injury [29] are associated with enhanced shame-proneness, but not guilt proneness. Our findings may thus seem at odds with this literature, but we argue that the discrepancy rests in methodological differences. The present study assessed a variety of childhood negative events, most of which were not investigated in previous research [26, 28]. We employed the same measure in one of our previous studies [29], but the analysis in that study did not control for traumatic intensity and therefore, a whole range of childhood negative events, from mild to traumatic, were included. In order to limit the heterogeneity of childhood stressors, the present study focused on traumatic events that were perceived by participants as having had a substantial impact on their personality and life course. As expected, only a minority of adolescents (i.e., 15%) reported such trauma, and we found that they had higher levels of guilt-proneness. The association between childhood trauma and guilt-proneness echoes previous observations that adolescents with depressive mothers tend to feel guiltier over failing to meet maternal expectations, compared to adolescents with non-depressive mothers [12]. Considering that the available literature on this topic includes only a handful of studies, future research should systematically describe the relations between different aspects of childhood trauma (e.g., type, severity, chronicity, age of exposure) and proneness to shame guilt.

We found no evidence for an association between age and sex, and dispositional shame and guilt in adolescents. A previous longitudinal study [24] showed that shame-proneness decreased and guilt-proneness increased from adolescence onward, with the former reaching a minimum around age 50, and the latter reaching a plateau around age 70. Therefore, age-related changes in shame-proneness and guilt-proneness may start in adolescence, but they extend into adulthood and this might explain why we found no association between age and these emotional dispositions in adolescents between ages 13 and 17. In which sex is concerned, a recent meta-analysis [52] has suggested that sex differences in shame and guilt are small, and this may account for the failure to detect such differences in the present study.

An important assumption of this study was that adolescence is marked by changes in emotion regulation [32], with a potential impact on the development of shame-proneness and guilt-proneness (e.g., [24]). Taking advantage of the large sample of adolescents included in the study, we were able to show that the habitual use of Rumination increases with age. Our results also illustrate the maladaptive role of emotion regulation strategies such as Self-Blaming, Rumination, Catastrophizing and Blaming Others, positively associated with depression and anxiety symptoms, and the adaptive role of strategies such as Positive Refocusing, Refocus on Planning and Positive Reappraisal, negatively associated with depression. The positive association between Acceptance and emotional symptoms may be explained by its specific efficiency in situations of uncontrollable stress (e.g., [53]). Putting into Perspective was also positively associated with anxiety, which underscores the potential negative consequences of playing down the seriousness of stressful events.

As expected [28, 38], raw scores of shame-proneness and guilt-proneness correlated, and therefore, derivate measures (i.e., standardized residual scores) were used to control for this interrelation. Using these indices, which may have improved construct validity [28], we found that shame-proneness was positively associated with depression and anxiety symptoms and guilt-proneness was negatively associated with depression symptoms. This underscores the dysfunctional nature of shame-proneness, which has been consistently linked to psychopathology [1]. The relation between guilt-proneness and psychopathology is less straightforward [1]. However, when guilt-proneness pervades emotional responses, it can also become problematic and contribute to emotional dysfunction [2]. This study was not designed to identify moderators in the relation between guilt-proneness and mental health, mainly because it did not employ a comprehensive psychiatric assessment, but this issue is an important challenge for future research.

Our results suggest that emotion regulation may be targeted in interventions aiming to reduce shame-proneness and prevent associated dysfunctions. Previous studies have shown that neglected children [26] and those exposed to harsh parenting [28] may be prone to exaggerated levels of dispositional shame and subsequent depression. By assessing ways of reducing the use of maladaptive emotion regulation strategies such as Self-Blaming and Catastrophizing, and enhancing the use of adaptive strategies such as Refocus on Planning and Positive Reappraisal, future research could examine the social and emotional benefits of reducing shame-proneness in children and adolescents.

The major limitation of this study is related to its correlational and cross-sectional nature, which does not allow us to draw a conclusion on the direction of the relations between habitual emotion regulation and dispositional shame and guilt. Future longitudinal studies and interventions targeting emotion regulation (e.g., promoting the use of adaptive strategies and discouraging the use of maladaptive strategies) will be able to characterize the relations between emotion regulation and proneness to shame and guilt. Another limitation of the present study is that it relied exclusively on self-report assessments of childhood trauma, emotion regulation, shame-proneness and guilt-proneness, and these measures may be susceptible to social and memory biases. For instance, childhood maltreatment may be underreported in questionnaires [54], and self-deceptive enhancement may inflate the self-report of socially desirable emotion regulation strategies such as reappraisal [55]. Therefore, future studies could use interview-based measures of childhood trauma (e.g., [54]) and cognitive tasks that directly assess the efficiency of emotion regulation strategies (e.g., [56]). In addition, it could be useful to differentiate between emotional and behavioral aspects of shame and guilt (e.g., [4]), and employ candidate biomarkers of these emotional dispositions (e.g., [57]).

In conclusion, these results describe the relations between individual differences in emotion regulation and dispositional shame and guilt in adolescence. In light of its positive association with the habitual use of maladaptive emotion regulation strategies and emotional symptoms, shame-proneness appears to be more dysfunctional than guilt proneness. Therefore, future research should seek ways of reducing shame-proneness in adolescence and we suggest that the optimization of emotion regulation might offer a promising mean of achieving this goal.

Acknowledgments

We are grateful to Roxana Chitta and Ioana Bunea for help with data collection.