Social and generalized anxiety symptoms and alcohol and cigarette use in early adolescence: The moderating role of perceived peer norms

Abstract

This study prospectively examines the association between social and generalized anxiety symptoms and alcohol and cigarette use in early adolescence and how injunctive (perceived peer approval of use) and descriptive (perceived peer use) norms may moderate the association. Sex differences were also examined. Data were taken from a longitudinal study investigating problem behavior and adolescent substance use. The community sample (N = 387) was assessed annually, and data from the first two waves of assessment were used for this study. Early adolescents were between the ages of 11 and 13 at the first assessment (mean age = 11.05, SD = 0.55, 55% female). Peer norms moderated the association between both social and generalized anxiety symptoms and the likelihood of alcohol and cigarette use for girls, but not for boys. Specifically, girls with elevated levels of generalized anxiety symptoms were at risk for use when perceived peer use was low, and protected from use when perceived peer use was high. Girls with elevated levels of social anxiety symptoms were at risk for use when perceived peer approval of use was high, and protected from use when perceived peer approval of use was low. Past studies have found inconsistent support for an association between anxiety and adolescent substance use, and our findings provide some clarity regarding for whom and when anxiety operates as a risk/protective factor. Social context and sex are critical for understanding the role of different forms of anxiety in the etiology of adolescent alcohol and cigarette use.

Introduction

Alcohol and cigarette use are initiated and escalate during adolescence (Colder et al., 2002, Colder et al., 2001, Johnston et al., 2011), and both are associated with a variety of negative outcomes (Aung et al., 2003, Cooper et al., 1994, Ellickson et al., 2001, Hill et al., 2000, Patten et al., 2000, Spirito et al., 1997). Investigating the etiology of alcohol and cigarette use in early adolescence is important because such research may inform interventions. Some studies suggest that high levels of anxiety symptoms are associated with adolescent alcohol and cigarette use, and hence, anxiety has been considered an important risk factor (e.g., Clark and Sayette, 1993, Deas-Nesmith et al., 1998, DiFranza et al., 2004, Kessler et al., 2005, Rohde et al., 1996). The mechanism often purported to account for this association is self-medication (Kassel et al., 2003, Wills and Hirky, 1996). That is, substance use is motivated by a desire to reduce emotional distress. Indeed, negative reinforcement is frequently endorsed as a motive for adolescent alcohol use and cigarette smoking (Cooper, 1994, Lewis-Esquerre et al., 2004).

Although there is evidence for anxiety as a risk factor, there is also evidence that anxiety is protective (e.g., Costello, Erkanli, Federman, & Angold, 1999) or not associated with adolescent substance use (e.g., Woodward & Fergusson, 2001). Unmeasured moderators may account for these inconsistencies. The epigenetic theory suggests that individual differences affect liability for substance use depending on the characteristics of the environment (Tarter et al., 1999, Zucker et al., 2011). According to this perspective, it would be important to consider potential person × environment interactions when examining associations between anxiety symptoms and substance use. A notable dearth of studies has considered potential moderators of the impact of anxiety symptoms on adolescent cigarette and alcohol use.

The goal of the current study was to clarify the role of anxiety in the development of early adolescent alcohol and cigarette use by considering how the peer context in the form of descriptive and injunctive norms and sex might moderate these associations. Some forms of anxiety may be more germane to adolescent alcohol and cigarette use than others (Colder et al., 2010, Hussong et al., 2011). Moderators may be particularly relevant for social and generalized anxiety as the literature has been equivocal on links between these domains of anxiety and early adolescent alcohol and cigarette use. Accordingly, we focus on social and generalized anxiety in this study.

Socially anxious individuals avoid social situations and have excessive concerns about negative evaluation and interpersonal rejection (Kearney, 2005). Although high levels of social anxiety are associated with adult alcohol and cigarette use (for reviews, see Morissette et al., 2007, Morris et al., 2005), there is evidence to support social anxiety as a risk (Buckner et al., 2008, Deas-Nesmith et al., 1998, Zimmermann et al., 2003) and a protective factor (e.g., Fröjd et al., 2011, Myers et al., 2003, Stewart et al., 2006) in adolescence. Adolescent alcohol and cigarette use largely takes place with peers (Verkooijen, de Vries, & Nielsen, 2007), and alcohol and cigarette use may serve as a way to self-medicate worries about social evaluation and make it easier to join with peers for socially anxious youth (Blumenthal, Leen-Feldner, Frala, Badour, & Ham, 2010). On the other hand, socially anxious youth are likely to be avoidant of social situations and may be protected from peer contexts where alcohol and cigarette use typically take place. Thus, social anxiety may operate as a both a protective and risk factor.

Generalized anxiety is characterized by excessive worry across a variety of domains (e.g., athletic performance, schoolwork, health) that is difficult to control (Wilmshurst, 2005), and often accompanied by a physical symptoms (e.g., restlessness or feeling keyed up or on edge, being easily fatigued, and muscle tension) (Merrell, 2008). Generalized anxiety symptoms are associated with adult alcohol and cigarette use (Bolton et al., 2006, Covey et al., 1994, Robinson et al., 2009). However, the empirical support for a link between generalized anxiety symptoms and adolescent alcohol and cigarette use has been equivocal (Kaplow et al., 2001, Wu et al., 2010). Generalized anxiety symptoms may increase risk for early adolescent alcohol and cigarette use via a self-medication pathway. This is consistent with evidence suggesting that self-medication is a prominent motive for a subgroup of youth characterized by high levels of anxiety sensitivity (Buckner et al., 2006, Morissette et al., 2007, Stewart and Zeitlin, 1995). On the other hand, early adolescents who are inclined to worry may be concerned about the consequences of engaging in an illicit and dangerous behavior like substance use, and such concerns may serve a protective function. Accordingly, generalized anxiety symptoms, like social anxiety symptoms, may plausibly increase or decrease liability for substance use.

Peers become particularly influential (Brown, Dolcini, & Leventhal, 1997) as youth spend more time with their peers and distance themselves from their parents during early adolescence (Larson & Richards, 1991). Changing peer dynamics may have implications for the role of anxiety in the development of substance use. Adolescents may become especially susceptible to social norms that convey information about what behaviors and attitudes are generally valued and accepted (Ward, 2011).

Descriptive norms refer to the prevalence of a particular behavior (e.g., the frequency and quantity of cigarette and alcohol use) while injunctive norms refer to peers' attitudes or beliefs surrounding a particular behavior (e.g., approval or disapproval) (Borsari & Carey, 2003). Both types of norms are considered theoretically important for adolescent substance use (Petraitis, Flay, & Miller, 1995). Descriptive norms may provide naïve early adolescents with the knowledge required to successfully adopt substance use behaviors and injunctive norms may provide expectations for positive substance use-related outcomes, leading to behavioral modeling (Borsari & Carey, 2001). Indeed, prospective research has demonstrated that peer substance use (descriptive norms) and peer approval of use (injunctive norms) predict later substance involvement (Ennett and Bauman, 1994, Larimer et al., 2004, Perkins, 2002, Schulenberg et al., 1999, Urberg et al., 1997, Wills and Cleary, 1999).

Socially anxious early adolescents might be particularly vigilant of descriptive and injunctive norms so that they can tailor their own attitudes and behaviors to gain acceptance and affirmation from peers. Indeed, studies of college drinking have found that socially anxious youth who perceived high levels of peer use and peer approval of use were at increased risk for alcohol use and problem drinking (Buckner et al., 2011, Neighbors et al., 2007). Such moderational effects of norms have not been tested in earlyadolescence, but peer norms may be important moderators during this period given the emergent importance of peers. We propose that high levels of social anxiety symptoms may decrease early adolescent alcohol and cigarette use when peer approval of use and peer use is perceived to be low, and may increase use when peer approval or peer use is perceived to be high.

The association between generalized anxiety symptoms and substance use may also depend on norms. To our knowledge, no studies have tested these potential interactions in adult or adolescent samples. Early adolescents with high levels of generalized anxiety symptoms may model peer use in order to alleviate distress, and perceived peer approval may provide reinforcement for coping motivated use. However, in the absence of a peer context supportive of substance use, anxious early adolescents may fear potential negative outcomes of use (e.g., health or legal consequences), and thus high levels generalized anxiety symptoms in this social context may be associated low levels of use.

Although adult studies suggest that anxiety symptoms in general, and social anxiety in particular, may be more strongly linked to alcohol and cigarette use outcomes for females (Buckner and Turner, 2009, Merikangas et al., 1996, Norberg et al., 2011), it is unclear if such sex differences generalize to early adolescence. Most studies examining adolescent substance use combine anxiety symptoms or disorders into one category. For example, reviews of psychiatric co-morbidity suggest no consistent pattern of association between an anxiety disorder and adolescent substance use/abuse (Armstrong and Costello, 2002, Roberts et al., 2007). Sung, Erkanli, Angold, and Costello (2004) found that anxiety symptoms increased risk for substance use disorders more strongly for girls than boys, but only after age 16.

Studies that have examined specific domains of anxiety suggest a similarly mixed picture regarding sex differences. For example, Zimmermann et al. (2003) examined prospective associations between several anxiety disorders and adolescent alcohol use and dependence, and findings suggested few sex differences with the exception of Phobia Not Otherwise Specified, which was more strongly linked to alcohol outcomes for females. Wu et al. (2010) found social phobia to be related to cigarette smoking in boys, but not girls, and reported a positive association between generalized anxiety and cigarette smoking in girls, but not boys. Kaplow et al. (2001) did not find sex differences in the association between generalized anxiety and alcohol use. Overall, these studies present a mixed picture regarding sex differences in the association between generalized and social anxiety symptoms and adolescent substance use. When sex differences emerge, they most often suggest that anxiety symptoms are more strongly associated with substance use for girls. The equivocal nature of sex differences suggests that it may be important to consider sex along with other moderators to clarify the role of anxiety in the etiology of early adolescent substance use.

The present study examines whether social and generalized anxiety symptoms will interact with perceived peer use (descriptive norms) and perceived approval of use (injunctive norms) to predict early adolescent alcohol and cigarette use. We offer the following hypotheses: (a) High levels of social and generalized anxiety symptoms will be associated with an increased likelihood of alcohol/cigarette use when perceived peer approval/use are high; and (b) high levels of social and generalized anxiety symptoms will be associated with decreased likelihood of use when perceived peer approval/use are low. We also examine sex as a potential moderator, and test 3-way norms × anxiety × sex interactions given some evidence for sex differences in associations between anxiety symptoms and early adolescent substance use. The proposed norm × anxiety interactions are hypothesized to be stronger for females than males, but this is a tentative hypothesis given the equivocal nature of sex differences in the literature. These hypotheses are tested using a prospective design to minimize potential interpretative difficulties regarding direction of the effects between anxiety and substance use. We also statistically controlled for depressive symptoms given the known overlap between anxiety and depression (Brady & Kendall, 1992).