Relationships between negative affectivity, emotion regulation, anxiety, and depressive symptoms in adolescents as examined through structural equation modeling

Abstract

The relationship between negative affectivity (NA) and emotion regulation (ER) in determining anxiety and depressive symptomatology was examined in a large (n = 1441) sample of adolescents (12–17 years old). Two models, diverging only as to inclusion or exclusion of a path from NA to negative ER, were analyzed through structural equation modeling; the goal was to explore the mediational or non-mediational role of ER in determining anxiety symptoms. The models yielded similar adequate fit to data, indicating that both NA and negative ER contribute to anxiety symptoms which, in turn, significantly determine depressive symptomatology. The mediational model better captures the relationships revealed in the data, with NA determining negative ER to a great extent. Additionally, most individuals scoring highly in NA also tend to score highly in negative ER, indicating that adolescents with heightened NA are prone to a dysfunctional style of ER.

Introduction

There is increasing recognition that anxiety and other emotional disorders have more commonalities than differences, from both a phenomenological and etiological perspective. This is perfectly captured in the well-known comprehensive triple vulnerabilities model proposed by Barlow, 1988, Barlow, 2000, Barlow, Allen, and Choate (2004) and in Craske's (2003) interactive layer model for the origins of anxiety disorders. In both models broad vulnerability factors, especially negative affectivity, are considered to contribute significantly to the development of anxiety. Data suggest that the connection between temperamental traits and other emotional disorders such as depression is mediated by anxiety symptoms (e.g., Muris, Merckelbach, Schmidt, Gadet, & Bogie, 2001). Both share the component of negative affect which has been claimed, from the tripartite model of anxiety and depression (Clark and Watson, 1991, Watson, 2005, Watson et al., 2008), to help explain the association between these psychopathological classes (see Anderson & Hope, 2008, for a review in adolescents). Moreover, there is substantial evidence that there may be a temporal relationship between anxiety and depression in youth and adult populations, with anxiety symptomatology and anxiety disorders tending to precede depression or, at least, increasing the risk for subsequent depression (e.g., Anderson and Hope, 2008, Bittner et al., 2004, Brady and Kendall, 1992, Chorpita and Daleiden, 2002, Kessler et al., 1996, Merikangas and Avenevoli, 2002).

The temperamental traits “negative affectivity” (NA) and “emotion regulation” (ER), as well as other related temperamental constructs such as effortful control, are key elements of theoretical models explaining the development of anxiety and depressive disorders (e.g., Lonigan and Phillips, 2001, Nigg, 2006, Rothbart et al., 1995, Rothbart and Rueda, 2005), from the developmental and personality research traditions. Although there have been significant advances in unlocking origins of anxiety in the last decade, including processes and mechanisms linking NA to the emergence of emotional disorders, “the field continues to lack a truly comprehensive theory of the developmental psychopathology of childhood anxiety” (Weems, 2008).

NA is the temperamental factor most commonly associated with anxiety and other emotional disorders (Lonigan and Phillips, 2001, Muris and Ollendick, 2005, Nigg, 2006). It is considered to be a common characteristic underlying these disorders as well as a well-established risk factor for their development. NA is an overarching construct including similar and overlapping concepts from different personality and motivation models, mainly those of neuroticism (Eysenck, 1967), sensitivity to punishment (Gray, 1987), and negative affect (Clark & Watson, 1991). NA can be broadly defined as “the proneness to experience an array of negative emotional states, and to activate defensive motivational systems” (Craske, 2003, p. 49). However, many individuals with heightened NA do not exhibit high levels of anxiety symptomatology or do not consequently develop anxiety disorders. This has led researchers to examine potential factors that mediate or moderate the relationship between the temperamental trait of NA and the presence of anxiety symptoms and the development of later anxiety pathology at clinical levels.

ER is perhaps the construct most frequently cited as playing this mediating or moderating role. A great deal of recent research has been devoted to this topic, highlighting its significance in understanding psychophatological processes. Some styles of ER, labeled as poor or negative, are considered to contribute to anxiety through mechanisms such as attentional bias to threat stimuli, increased unpredictability, and lack of contingency awareness; these ER styles elevate the threat value of negative affect and related stimuli (Craske, 2003, Lonigan and Vasey, 2009).

As of yet, there is no consistent definition of ER (e.g., Amstadter, 2008, Bridges et al., 2004, Cole et al., 2004, Eisenberg and Spinrad, 2004, Weems and Silverman, 2006). Several approaches to ER, from different psychological traditions, coexist, mainly (a) the process model of emotion that integrates the broad literature on psychosomatic medicine with that of coping and clinical psychology (Gross, 1998), (b) the developmental psychology literature (Thompson, 1994) as well as (c) studies grounded in the biobehavioral and personality (mainly temperamentalist) research traditions (e.g., Eisenberg et al., 2001, Eisenberg et al., 2005, Eisenberg and Spinrad, 2004, Lonigan and Phillips, 2001, Rothbart et al., 2000, Rothbart and Derryberry, 1981). Fortunately, there are some commonalities among these definitions (Valiente et al., 2003): all of them conceive of ER, in varying degrees, as the responses that influence the experience and expression of emotions. Divergences embrace, at least, that ER is a part of self-regulation or also including external regulation, to be voluntary/purposeful or involuntary/automatically controlled, and to be a conscious or unconscious process. In this article, we endorse a definition of ER as the individual style of responding to stressful events, specifically “the cognitive way of managing the intake of emotionally arousing information” (Garnefski, Kraaij, & Spinhoven, 2001).

Garnefski, Kraaij, and van Etten, 2005, Garnefski, Legerstee, Kraaij, van den Kommer, and Teerds (2002) and Garnefski, van den Kommer, et al. (2002) have reported that, in both adolescents and adults, use of negative forms of ER (self-blame, rumination, catastrophizing) and, to a lesser extent, not using positive ways to manage the intake of emotionally arousing information, predicts a significant number of symptoms of depression and anxiety. Poor ER has been also found in children with anxiety disorders (Suveg & Zeman, 2004) and has been associated with anxiety and depressive symptoms when it appears in conjunction with high frequency of negative affect (Suveg, Hoffman, Zeman, & Thomassin, 2009).

From a wider clinical perspective, there is strong evidence that some kinds of ER (such as attempts to avoid or to suppress expressions of emotion and unwanted thoughts) are closely related to most, if not all, anxiety and depressive disorders. In fact, these processes are considered to be a core feature in the maintenance of these problems (e.g., Barlow et al., 2004). Moreover, negative cognitions that can be considered negative forms of ER appear to be specifically associated with negative affect (Mausbach, Roepke, Depp, Patterson, & Grant, 2009). That is, it is accepted that some forms of ER that could generally be labeled as emotional avoidance are clearly counterproductive or negative in contributing to exacerbating and maintaining anxiety symptoms.

Additionally, evidence is growing that individuals who are low in effortful control, an important form of self-regulation considered central to cognitive and emotional regulation of children's behavior (Rothbart & Rueda, 2005), and who are also high in NA, are prone to exhibit greater anxiety symptoms and to develop anxiety and other emotional disorders. In summary, having ER deficiencies or having a negative style of ER are associated with higher levels of psychopathology; anxiety disorders appear to result from the combination of high NA and negative or poor ER.

Despite recent inroads, knowledge is still scarce about the relationship between NA and ER, and how both operate in the development of anxiety and other emotional disorders. Rothbart and Bates (1998) claim to examine the interactions between or among different temperamental characteristics is now starting to be addressed systematically in the literature. Some consistent data suggest that an attentional bias to threat stimuli, along with difficulties in executive and cognitive control processes, could be a central mechanism linking NA and ER (Lonigan and Vasey, 2009, Verstraeten et al., 2009).

The current study focuses mainly on the relationships between NA and ER. One question that remains to be answered and that has garnered little attention thus far is: to what extent can negative ER be determined, at least in part, by other temperamental features such as NA? The ER research tradition on effortful control hypothesizes that a failure in this process could be due, in part or in some cases, to temperamentally low capacity (Lonigan and Phillips, 2001, Rothbart, 2007). Nigg (2006) analyzed the distinction between effortful control and reactive control, a more automatic form of ER (Eisenberg & Spinrad, 2004); he indicated that reactive control would primarily be related to other temperamental traits such as extraversion or neuroticism (or NA); he considers this to be a key empirical question to be explored further. Moreover, Craske (2003) considers that some kinds of ER not only contribute to anxiety but also represent the manifestation of anxiety. Similarly, Davidson (1998) and Amstadter (2008) noted that emotions are rarely generated without accompanying regulatory processes. That is, ER is an inherent aspect of emotional response tendencies.

Lynch, Robins, Morse, and Krause (2001) used structural equation modeling to test a theoretical model relating negative affect intensity, emotion inhibition (a type of negative ER), and psychological distress in clinical and nonclinical samples. One especially relevant result of this study was that negative affect intensity determines, to some extent, the degree to which one attempts to control emotions by avoiding expressions of emotion and trying to suppress unwanted thoughts; furthermore, these ER strategies lead to increased psychological distress. Suveg et al. (2009) and Suveg and Zeman (2004) have also studied the relationship between experiencing emotions with high intensity and/or presenting a high frequency of negative affect with the difficulty of managing negative emotions such as worry, anger, sadness or hyperarousal in an adaptive way. That is, the manner in which emotions are experienced (with more or less intensity), which is widely associated with the NA temperamental trait, might determine the ability to regulate them. Suveg and Zeman (2004), and more recently Weems and Silverman (2006), discuss this in the more general context of conceiving of ER not only in terms of actual control but also as diminished perceived control and low self-efficacy with regulating emotions.

As the above studies suggest, some kinds of ER, those labeled as negative (thought suppression, rumination, etc.), could represent, at least partially, a reflection of NA. That is, could negative ER be, in some degree, a reactive control response that is greatly determined by heightened NA? In general, theoretical models view NA and ER as independent yet related temperamental factors.

The purpose of this study was to examine the relationship between NA and ER in determining concurrent anxiety and depressive symptomatology in adolescents.

In the present study, NA includes the concept of negative affect from the tripartite model (Clark and Watson, 1991, Watson et al., 1988) and also considers sensitivity to punishment as a measure of the behavioral inhibition system from Gray's (1987) reinforcement sensitivity theory. The aforementioned models have been systematically used in research on the relationship between personality and psychopathology; while negative affect has been studied in relation to ER and self-regulation, scarce research has been conducted on the relationship between the behavioral inhibition system and ER (Bijttebier, Beck, Claes, & Vandereycken, 2009).

Two overall theoretical models relating NA, ER, and anxiety and depressive symptomatology were evaluated using structural equation modeling in a large nonclinical adolescent sample. In the first model, the structural paths relating the latent variables included paths from negative affect, sensitivity to punishment, and negative ER to anxiety symptoms and an additional path from anxiety to depressive symptoms. In this model, NA and negative ER were each viewed as making a unique contribution to anxiety symptoms (non-mediational model). The second model (mediational model) considers whether NA determines negative ER. It includes paths from negative affect and sensitivity to punishment to negative ER, from which a path connects with anxiety symptoms; the last path extends from anxiety to depressive symptoms.

To further evaluate the relationship between NA and negative ER, the pattern of distribution of both variables was analyzed across the whole sample. This allowed us to explore the tendency of individuals with high and low NA to also exhibit high or low scores, respectively, in negative ER; furthermore we could then compare the resulting subsamples in terms of self-reported anxiety and depressive symptoms.

It was hypothesized that: (1) both high NA and negative ER, irrespective of the mediational or non-mediational role of ER, would be associated with anxiety symptoms, (2) that depressive symptomatology would be largely determined by anxiety, (3) that high NA would be related to an increased use of negative ER, and (4) that ER would be mediating the relationship between NA and anxiety symptoms.