Introduction
Dysregulated emotions are thought to play a significant role in the etiology and maintenance of many forms of psychopathology (e.g., Bradley
2000; Cole and Hall
2008). In fact, negative or dysregulated emotions are diagnostic symptoms of many disorders as described in the DSM-IV (American Psychiatric Association [APA],
1994). For instance, excessive anxiety and worry, and/or intense fear are implicated in the anxiety disorders, while irritability, anger, and mood lability are implicated in the disruptive behavior disorders (conduct disorder and oppositional defiant disorder; APA,
1994). High levels and prolonged duration of negative emotions and heightened emotional variability may be signs of emotional dysregulation (Cole and Hall
2008). Individual differences in emotion regulation and their relation with the development of psychopathology become especially relevant during the developmental period of adolescence. Adolescence is characterized by an increase in the intensity and frequency of (negative) emotions (e.g., Larson and Lampman-Petraidis
1989), heightened levels of emotional variability (Larson et al.
1980), increases in several types of psychopathology (e.g., Moffitt et al.
1996; Wight et al.
2004), and increased demand for independent self-regulation. However, longitudinal studies that address the role of emotion dysregulation in the development of psychopathology in adolescence are scarce. In consequence, little is known about the role of emotion dysregulation in general, and about the role of emotional dynamics (the intensity and variability of emotions; Silk et al.
2003) in the development of adolescent behavioral and emotional problems in particular. The present study focused on the role of the dynamics of four basic emotions (happiness, anger, anxiety, and sadness) in the 1 year change or stability of internalizing and externalizing psychopathology in early adolescence.
Emotional Dynamics and Psychopathology in Adolescence
The term “emotion dysregulation” refers to maladaptive patterns of emotion regulation, that is, to patterns of emotion regulation that are costly in the pursuit of long-term goals such as maintaining social relationships and well-being (Cole and Hall
2008). The dysregulation of emotions may be studied at all different levels of emotion experience, cognition and regulation, such as emotional dynamics (Silk et al.
2003), emotion knowledge (e.g., not knowing that one may experience different emotions at the same time and believing that emotional experiences cannot be modulated; e.g., Meerum-Terwogt and Olthof
1989), difficulties with the use of emotion regulation strategies (e.g., distraction, cognitive reinterpretation; Gross and Thompson
2007), and meta-emotion experiences (e.g., nonacceptance of emotional responses; Gratz and Roemer
2004). We consider emotional experiences and behaviors that are too little, too much, or not appropriate for the situation at hand as signs of dysregulation. This study is focused on emotion dynamics, as these reflect the outcome of regulatory efforts and processes. Specifically, we focus on adolescents’ daily experiences of emotions. At the level of emotion dynamics, emotion dysregulation may be reflected in the form of heightened and prolonged negative emotions, and heightened variability of both negative and positive emotions (Cole and Hall
2008; Silk et al.
2003). We define emotional variability (EV) here as “the
frequency with which an individual’s emotions change, and the
extremity of these changes”. It should be noted that some variability of emotions, as well as the experience of negative emotions is clearly expected as a result of normal emotional reactivity to environmental stressors. Heightened and prolonged negative emotions and heightened variability, however, may be indications of extreme emotional reactivity and ineffective attempts at emotion regulation.
As indicated previously, early to mid adolescence is an opportune time to study emotional dynamics, and its links with psychopathology. EV was found to be heightened in adolescents compared to both younger children and adults in a study, which assessed EV several times a day over the course of 1 week (Larson et al.
1980). Although heightened EV may partly be a sign of normative developmental changes in adolescence (Larson et al.
1980), comparatively high levels of EV in adolescents may indicate emotional dysregulation, and be indicative of future psychopathology (Schneiders et al.
2006).
Several cross-sectional studies demonstrate that in adolescence, high levels of negative emotions, and high levels of EV are related to symptoms of depression (Larson et al.
1990; Silk et al.
2003), and externalizing problems (Silk et al.
2003). These findings are commonly interpreted as suggesting that emotion dysregulation influences the development of psychopathology. However, given the cross-sectional nature of these studies, the role of emotion dysregulation and EV in the development of psychopathology needs further clarification. Therefore, in the present study, we first tested the hypothesis that high levels of negative emotions, low levels of positive emotions, and high variability of both positive and negative emotions add to the growth or development of different types of psychopathology. In addition, we addressed questions on the importance of intensity versus variability of emotions regarding the development of behavioral/emotional problems, on the specificity of the emotion by psychopathology type association, and the potential gender-specificity of these associations.
A first question is whether it is emotion dysregulation in general or dysregulation of specific emotions that is linked to the development of specific forms of psychopathology. For example, does dysregulated anger relate specifically to aggressive problems, and dysregulated sadness relate specifically to depression? ‘Functional continuity’ between discrete emotions and specific forms of psychopathology is often assumed (e.g., Cole and Hall
2008; Malatesta and Wilson
1988; Muris and Ollendick
2005). That is, it is assumed that patterns of emotional responding become consolidated over time and consequently lead to specific forms of psychopathology (Zahn-Waxler et al.
2000). Accordingly, individuals who are often fearful are thought to be more likely to develop anxiety-related than other psychological disorders, individuals who are often sad, and experience little happiness will show depressive symptoms, and individuals who are easily angered will display symptoms of disruptive behavior disorders (e.g., Muris and Ollendick
2005). Empirical evidence for specific links between emotions and specific forms of psychopathology in adolescence is rather mixed and may heavily depend on the method used to assess emotions. One cross-sectional study addressing links between observer-rated facial expressions of emotions (anger, fear, sadness, and embarrassment) and internalizing and externalizing problems in early-adolescent boys, provided support for the idea of specific links (Keltner et al.
1995): boys with externalizing problems showed more anger, and boys with internalizing problems showed more fear. By contrast, studies focusing on the internal experience rather than the expression of emotion tend to find nonspecific associations between discrete emotions and forms of psychopathology. High and variable levels of sadness have been associated with depressive symptoms, and high and variable levels of anger with aggressive problem behavior (Larson et al.
1990; Silk et al.
2003). However, both high and variable anger and anxiety also correlate with depressive symptoms, and high and variable levels of sadness and anxiety correlate with aggressive problem behavior scores in adolescents (Silk et al.
2003). As the present study used adolescent daily reports of experienced emotions we expected to find mainly nonspecific associations between levels and variability of emotions and change in psychopathology.
The study of basic emotional processes in the development and maintenance of psychopathology is not just of theoretical interest, but may also inform preventive intervention and treatment efforts. For instance, the question whether links between discrete negative emotions and psychopathology are better characterized as specific or as general, has potentially important implications for preventive interventions and treatment: if fear, as well as anger and sadness turn out to all play a role in anxiety disorder symptoms, teaching anxious youths how to deal with feelings of fear is insufficient.
Secondly, comparatively little is known about the role of positive emotions in the development of psychopathology. It has long been assumed that elevated negative emotions are a common feature of anxiety and mood disorders, whereas diminished positive affect is specific for mood disorders (e.g., Clark and Watson
1991). A recent review and meta-analysis however, has shown significantly diminished positive affect in social anxiety (Kashdan
2007). Regarding externalizing problems, results of studies that have included indices of positive emotions are mixed: no differences in the display of happiness between aggressive and non-aggressive youth (Orobio de Castro et al.
2005), lower state but not trait happiness in delinquent youth than a comparison group (Plattner et al.
2007), and higher happiness in response to antisocial acts in adolescent males with conduct disorder have all been reported (Cimbora and McIntosh
2003). We hypothesize that, in addition to high levels of negative emotions, low levels of positive emotions are related to increased levels of anxiety, depression, and aggressive behavior.
A third question concerns the gender-specificity of the hypothesized associations. It has been reported that males and females differ in levels of specific types of psychopathology (for a review see Zahn-Waxler et al.
2008), as well as in mean levels and variability of negative emotions (Silk et al.
2003). While it is uncertain whether adolescent sex needs to be taken into account when studying the role of emotional dynamics in the development of internalizing and externalizing problems, it might be that sex-related differences in emotional dynamics in part underlie the established gender differences in levels of internalizing and externalizing problems. For instance, emotion dysregulation may be more strongly associated with internalizing problems for females, and more strongly with aggressive behavior for males. The present study tested whether the associations between emotional dynamics and psychopathology are moderated by adolescent gender.
The Present Study
The present study addressed the role of the intensity and variability of happiness, anger, anxiety, and sadness in the development of internalizing and externalizing psychopathology in 452 adolescents followed from age 13 to 14. Adolescents’ experiences of emotions were obtained from three 5 day periods, during which adolescents reported on their emotions every day. These intensive assessments of emotional experience contain two strengths: they minimize recall bias, inherent in many other self-report methods of subjective experience, and they maximize ecological validity (Shiffman et al.
2008). We expected that high levels and high variability of negative emotions (anger, anxiety, and sadness), predict the stability of adolescents’ anxiety, depression, and aggressive behavior scores from age 13 to age 14, in non-specific ways, thereby extending earlier cross-sectional findings (Silk et al.
2003). Further, we expected that positive emotions play a role in the continuity of symptoms of psychopathology. Specifically, we predicted that low levels of—and high variability in—happiness partly predict the 1 year continuity of adolescents’ anxiety, depression, and aggressive behavior scores. An additional question that was tested is whether the level and variability of emotions both contribute uniquely to the development of psychopathology. This is an important question, because the intensity and variability of emotions tend to correlate, but their contributions to symptoms of psychopathology have not been tested in one model (e.g., Silk et al.
2003). Finally, we tested whether the associations between emotional dynamics and psychopathology hold across adolescent gender.
Discussion
The present study examined the role of adolescent emotional dynamics in the development of psychopathology from age 13 to 14. Results showed that individual differences in the dynamics of happiness, anger, anxiety, and sadness partly account for the development of adolescent anxiety disorder and depressive symptoms, and that the dynamics of the negative emotions, but not happiness, partly account for the development of aggressive behavior. These findings revealed three overarching results. First, most links between the four basic emotions and the three problem areas were non-specific. That is, the dynamics of the four emotions were similarly related to the forms of psychopathology studied. This was particularly true for anxiety disorder and depressive symptoms, while associations between emotion dynamics and aggressive behavior were a bit more specific. In particular, anger played a more consistent role than the other emotions in aggressive behavior. Second, the variability, but not the levels of emotions predicted the development of anxiety disorder symptoms, while the levels, but not the variability, of emotions predicted depressive symptoms and aggressive behavior, when effects of both variability and levels of an emotion were studied in one model. Third, despite the fact that females had higher levels of anxiety and depression, the role of emotional dynamics in the development of psychopathology was similar for both sexes; with the main exception that aggressive behavior was predicted by levels of sadness and anxiety for female, but not for male adolescents.
Building on the assumption that elevated levels of negative emotions, diminished levels of happiness, and elevated emotional variability are all indices of emotion dysregulation, the results add to a growing body of evidence showing that emotion dysregulation predicts symptoms of anxiety, depression and aggressive behavior in children and adolescents (Beauchaine et al.
2007; Bosquet and Egeland
2006; Yap et al.
2008). These findings are consistent with theory proposing that emotion dysregulation underlies the development of psychopathology (Bradley
2000). Our results are also consistent with earlier cross-sectional research linking adolescent negative emotions and emotional variability with symptoms of depression and problem behavior in adolescents (Larson et al.
1990; Silk et al.
2003).
The finding that the dynamics of happiness, anger, anxiety, and sadness contributed in mostly non-specific ways to different problem areas, may be surprising from the view of functional continuity between emotions and psychopathology. However, our findings are consistent with the notion of a general emotion dysregulation factor (Silk et al.
2003; Steinberg and Avenevoli
2000), and are in line with findings that most forms of psychopathology include dysregulation in more than one discrete emotion. For instance, increased anger (Moscovitch et al.
2008), and diminished happiness (Kashdan
2007) have been implicated in the anxiety disorders, and depression involves the dysregulation of both positive and negative emotions (Gross and Levenson
1997). The fact that, in contrast to the current findings, some studies report specific links between discrete emotions and forms of psychopathology in adolescents (Keltner et al.
1995) and children (e.g., Eisenberg et al.
2001), may be due to differences in the assessment of emotions. While we studied the adolescents’ own emotional experience, the above studies used videotaped observations of emotional displays (Keltner et al.
1995) or observer ratings of emotions (Eisenberg et al.
2001). For an emotional experience to be observable, it must have been translated into some form of behavior (e.g., facial or verbal expression). The way in which an internal experience is expressed may very well be an important factor in determining the specific form of psychopathology that one develops. For instance, when both anger and anxiety are often experienced, but the expression of anger is usually suppressed, while anxiety is shown more freely, findings based on observations of these expressions may suggest that the individual is more likely to develop an anxiety disorder than antisocial personality disorder. As our results show, however, at an experiential level, high levels and variability of negative emotions (and low levels of positive emotions), are implicated in non-specific ways in adolescent mental health problems. This suggests that the specificity hypothesis does not hold for experiential data. Future research needs to explicitly address the question of what determines emotional expression—candidate variables may be intrapersonal variables, such as behavioral inhibition versus behavioral activation (Gray
1982) as well as social variables, such as cultural and gender specific display rules for emotions (e.g., Brody
2000).
This non-specificity may have implications for our understanding of comorbidity. Anxiety disorders, mood disorders, and disruptive disorders show strong co-occurrence in childhood and adolescence. Caron and Rutter (
1991) suggest that one possible reason for overlap between two disorders is that they share the same risk factor or factors. This possibility arises from the fact that many psychiatric disorders are multifactorial in origin and that many causal factors are not diagnosis-specific. However, it is not known whether the shared risk factors mechanism of several factors mentioned by these authors (e.g., temperamental variables, family adversity) does in fact account for patterns of comorbidity. It may well be that emotional dysregulation is one of the risk mechanisms underlying multiple disorders, including the ones addressed in the present study.
Some specificity in links between the type of emotion dynamics and forms of psychopathology was still found: increased variability in all emotions was specifically related to anxiety disorder symptoms, while elevated levels of negative emotions and diminished levels of happiness, but not their variability were specifically related to depression. This finding is consistent with reports that the development of anxiety and depressive disorder symptoms of adolescents from the general community, though parallel, occurs as two distinct disorders (Hale et al.
2009), and may have implications for our understanding of depression versus anxiety disorder. What differentiates individuals at risk for depression from individuals at risk for anxiety disorders, may be that the former experience “learned helplessness” (Abramson et al.
1978), while the latter alternate between hope and fear. Thus, depression may be characterized by the belief that there is little that one can do about one’s situation in general, and about one’s affective state in particular. In contrast, inherent in the excessive worry experienced by individuals at risk for anxiety disorders, may be not only the fear that things will not turn the way one wishes, but also the hope that they will—leading to rapid changes in emotional states. This interpretation, though speculative, is consistent with findings that symptoms of anxiety often precede symptoms of depression (Cole et al.
1998). While there is still hope in anxiety, it is lost in depression, as reflected in the present study by heightened variability of emotional states in anxiety, and rather high and stable levels of negative emotions, and low happiness in depression. A recent study of emotional inertia (defined as ‘the degree to which emotional states are resistant to change’; Kuppens et al.
2010, p. 984) in 16 year olds, showed that emotional inertia was significantly higher in adolescents who met DSM-IV criteria for current major depressive disorder than in a control group. Taken together with the results of the present study, these findings suggest, that depression in younger adolescents is characterized mainly by high levels of negative emotions, while depression in older adolescents with clinically high symptom levels, is (additionally) characterized by high emotional inertia. Future research may investigate the role of emotional dynamics in the development of anxiety and depression using longer timeframes, comparing clinical and non-clinical groups, and including more cognitive predictors such as hope and helplessness.
While some associations between anxiety and sadness dynamics on the one hand, and the development of aggressive behavior on the other hand, were found, anger clearly played the most prominent role in adolescents’ aggressive behavior. Compared to the internalizing problems, aggressive behavior thus appears less emotional. This is in line with research showing that self-reported emotion dysregulation (Neumann et al.
2010) explains more of the variance of internalizing than of externalizing problems. However, it is still possible, that certain forms of externalizing problems, which were not studied in the present investigation, are more emotional. It has been suggested that reactive aggression is more emotional than proactive aggression, and it has indeed been found that reactive aggression is related to poorly regulated responses to emotional stimuli (Vitaro et al.
2002), while proactive aggression is related to callousness and emotional shallowness (Frick et al.
2003; Marsee and Frick
2007).
In addition to sex differences in the levels of emotion variability and anxiety and depression, some sex differences in associations between psychopathology and emotion dynamics were found. Anxiety disorder symptoms predicted happiness and anger variability more strongly for males, and sadness and anxiety predicted aggressive behavior for female, but not male adolescents. These gender differences seem to suggest that more emotions are involved in the less gender-typical disorders, akin to the gender paradox as proposed for conduct problems (Tiet et al.
2001). However, breaking this pattern, depressive symptoms predicted happiness variability for females, but not males. The gender differences that we found are thus not entirely conclusive, and it should also be noted, that most paths between emotion dynamics and psychopathology were not moderated by gender. It might be that sex differences in emotion variability and in psychopathology, and their potential associations are not yet distinct enough in early adolescence. Research has shown that sex-differences in (especially internalizing) psychopathology emerge/widen throughout adolescence (Zahn-Waxler et al.
2008). Further research across a larger age range is needed to adequately test effects of gender as a moderating variable in associations between emotion dynamics and psychopathology.
Some limitations of this study need to be mentioned. First, the sample consisted of white, mostly middle-class Dutch adolescents; thus caution should be exercised in generalizing the results to the general adolescent population. It is also unclear how the present results generalize to clinical populations. Further, the sample represented only a small age-range of adolescence. Another limitation is that EV and internalizing and externalizing problems were assessed by adolescent self-report only. Associations between EV and psychopathology might have been inflated by shared source variance. Subjects who experience high levels of negative emotions and heightened emotional variability, might also report more symptoms of psychopathology. The subjective experience of emotions and psychological well being is important; nevertheless, it would be interesting to see if the associations found in the present study hold, if reports by other informants, diagnoses, or observational methods are used. Also, EV is generally assessed over several days, with multiple assessments of emotional intensity per day (e.g., Larson et al.
1980; Silk et al.
2003). In the present study, EV was indexed as EV over the course of three one-week periods, with one assessment of emotional intensity per day. It would be interesting to see if the present results replicate with EV measured intensively over the course of one or several days. However, again, our findings were in accordance with theoretical assumptions and earlier empirical studies. Therefore, not having a within-day variability measure of EV did not seem to overly affect our results.
Clearly, especially with regard to the variability of emotions, more work is needed to understand its nature and development in adolescence, and how it relates to emotional reactivity to stressors, and emotion regulation strategies. For instance, how do emotional reactivity and emotion regulation strategies interact, i.e., can some highly emotionally reactive individuals modulate their emotions in such effective ways that emotion dysregulation does not result? Further, concerning the link to psychopathology, an important issue would be: when does EV become so strong that it becomes a risk for the development of emotional problems? Clearly, emotions and moods that resist change are maladaptive as well (Cole and Hall
2008), and to some extent EV is normative in response to events that are common in adolescence, such as difficulties with peers, or major disappointments in school. Since our measure of EV indexed 3 weeks, separated by 3 months, we are confident that such occurrences have not had a major impact on our results, but it would nevertheless be informative to study EV in the context of adolescents’ daily lives.
Future studies of the sources and consequences of EV may also want to include pubertal status, as pubertal status has been related to both depressive and aggressive affect (Brooks-Gunn et al.
1994), and recent evidence shows that pubertal status and emotional reactivity to experimentally induced stress interact in the prediction of internalizing symptoms (Leen-Feldner et al.
2007).
In sum, results of the present study imply that when studying the emotional underpinnings of (internalizing) psychopathology, researchers may want to focus less on the specific emotions, and more on the general form the dysregulation takes, as indicated by high levels of negative, and low levels of positive emotions, or highly variable emotions. In addition to application to research and theory, the study of basic emotional processes in adolescence is also informative for prevention and intervention efforts, as early forms of emotion dysregulation can indicate risk for psychopathology (Cole and Hall
2008). An important message regarding intervention from the present study, then, is to consider the entire emotional spectrum.