Journal of Personality and Social Psychology Copyright 1998 hy Ihc American Psychological Association, Inc. 1998, Vol. 74, No. 5, 1380-1397 0022-3514/9843.00 Attachment Styles, Emotion Regulation, and Adjustment in Adolescence M. Lynne Cooper Phillip R. Shaver University of Missouri—Columbia University of California, Davis Nancy L. Collins State University of New ^brk at Buffalo Attachment style differences in psychological symptomatology, self-concept, and risky or problem behaviors were examined in a community sample (N = 1,989) of Black and White adolescents, 13 to 19 years old. Overall, secure adolescents were the best-adjusted group, though not necessarily the least likely to engage in risky behaviors. Anxious adolescents were the worst-adjusted group, reporting the poorest self-concepts and the highest levels of symptomatology and risk behaviors. In contrast, avoidant adolescents reported generally high levels of symptomatology and poor self-concepts but similar levels of risk behaviors to those found among secures. Mediation analyses suggested that the observed differences in problem behaviors were at least partially accounted for by the differential experience of distress symptoms (primarily hostility and depression) and by social competence. Finally, patterns of attachment effects were similar across age, gender, and racial groups, with some important exceptions. During adolescence, the hierarchy of attachment figures tic of this developmental period. The present study, therefore, (Bowlby, 1969, 1982) is gradually reshuffled as young people examined individual differences in attachment styles as pre- increasingly direct their attachment behaviors and concerns to- dictors of adjustment in a representative community sample of ward peers rather than parents (Furman & Buhrmester, 1992; Black and White adolescents and tested three broad sets of Hazan & Zeifman, 1994). Although parents are generally not hypotheses linking attachment styles to psychological symptom- completely displaced as attachment figures during this period, atology, self-concept, and a range of developmentally relevant or perhaps ever, they slowly become what Weiss (1982) called risk or problem behaviors. "attachment figures in reserve." By the end of this period, sometime in early adulthood, most people settle on a single Attachment Theory and Attachment Styles romantic partner who will serve for years, if not for the remain- der of life, as a primary attachment figure. While making this John Bowlby (1969, 1973) was the first to present a coherent transition, many adolescents alter their conceptions of and feel- model of the process by which the bond between mother and ings about themselves and experiment with a range of explor- infant develops and the functions that this bond serves. He ar- atory behaviors (e.g., sex and substance use) that may be devel- gued that, because of the prolonged dependence of an infant opmentally functional but nonetheless carry substantial risk of on its mother, behavioral mechanisms evolved to protect the harm (Baumrind, 1987). Despite the co-occurrence of these immature offspring and to increase its chances of survival to phenomena during adolescence, little is known about how at- reproductive age. Essentially, mother and infant are thought to tachment patterns are related to the emotional experiences, at- have evolved a coordinated relationship in which the infant's tempts at self-definition, and exploratory behaviors characteris- signals of distress or fear are noted by the mother, who in turn offers comfort and protection, as well as a secure base from which the infant can explore the environment. According to M. Lynne Cooper, Department of Psychology, University of Missouri— Bowlby (1969), these early caregiving experiences are internal- Columbia; Phillip R. Shaver, Department of Psychology, University of ized as working models that not only serve as a prototype for California, Davis; Nancy L. Collins, Department of Psychology, State future relationships with significant others but also provide un- University of New Tfork at Buffalo. written rules for how one experiences, expresses, and copes This research was supported by Grant AA08047 from the National with distressing emotions. Institute on Alcohol Abuse and Alcoholism; manuscript preparation was Ainsworth and her colleagues (Ainsworth, 1973; Ainsworth, also partly supported by a grant from the Alcoholic Beverage Medical Blehar, Waters, & Wall, 1978) subsequently developed a system Research Foundation. We thank Jeremy Skinner for his thoughtful com- for identifying and describing individual differences in attach- ments on a draft of this article, Robb Peirce for assistance in running some of the analyses, and Kim Canavan and Cynthia Mercado for help ment among mother-infant dyads. They found that infants dif- in manuscript preparation. fered in the way they handled the stress of being left alone by Correspondence concerning this article should be addressed to M. their mother in a strange situation—a laboratory room equipped Lynne Cooper, Department of Psychology, University of Missouri, Co- with a host of novel toys. The majority of infants, called securely lumbia, Missouri 65211. Electronic mail may be sent to psymlc@ attached, became somewhat subdued or distressed in their moth- showme.missouri.edu. er's absence but expressed warm, relieved greetings and were 1380 ATTACHMENT AND ADJUSTMENT IN ADOLESCENCE 1381 quickly soothed by her when she returned. The remaining infants negative outcomes, and that distressing emotions cannot be ef- coped in two strikingly different ways, both of which Ainsworth fectively regulated. On the basis of these early experiences, et al. called insecurely attached. Some—labeled anxious-am- Bowlby (1969) argued, expectancies regarding the experience bivalent—protested and cried when their mother left, as well of negative emotions and preferred styles of coping with these as while she was gone. They acknowledged their mother's return emotions are internalized in the form of working models. Work- and sought to be held, but surprisingly (given their obvious ing models are thought to include both conscious and uncon- distress at her departure) continued to seem angry and distraught scious schematic elements that guide perceptions and trigger when she tried to calm them. The third group, called avoidant, characteristic emotions, as well as defense mechanisms, or rules seemed undisturbed by their mother's departure and cool, if not for regulating emotion and for processing or failing to process disinterested, when she returned. They did not seek physical certain kinds of attachment-relevant information. (See Shaver, cuddling or comforting and appeared to be prematurely self- Collins, & Clark, 1996, for an overview of the internal working reliant. models construct.) These models are thought to persist across Hazan and Shaver (1987; Shaver, Hazan, & Bradshaw, 1988) time and exert pressure toward continuity in affective experience suggested that these same patterns are evident in adolescent and and behavior. Kobak and Sceery (1988) provided a good sum- adult romantic and marital relationships, and they developed a mary of attachment theory's account of individual differences simple measure (based on Ainsworth's descriptions of the three in emotion regulation: infant types) to assess them. From scores of studies with college students and adults (see Shaver & Hazan, 1993, and Rothbard & Secure attachment [is] organized by rules that allow acknowledg- Shaver, 1994, for reviews), a portrait has emerged of the three ment of distress and turning to others for support, avoidant attach- ment by rules that restrict acknowledgment of distress and the kinds of individuals identified by this attachment measure. Se- associated attachment attempts to seek comfort and support, and curely attached adults are self-confident, socially skilled, open [anxious-] ambivalent attachment by rules that direct attention to and interested in close relationships with romantic partners, toward distress and attachment figures in a hypervigilant manner and likely to form relatively stable and satisfying long-term that inhibits the development of autonomy and self-confidence, (p. relationships. Anxious, or anxious-ambivalent, adults lack self- 142) confidence; are worried about rejection and abandonment; are prone to bouts of jealousy and anger at relationship partners who In other words, secure individuals should be able to acknowl- are perceived as untrustworthy; are eager to become involved in edge and then cope effectively with negative emotions, avoidant romantic relationships despite their perils; and are likely to en- individuals try not to acknowledge negative emotions and conse- gage in inappropriately intimate serf-disclosures, to fall in love quently may act emotionally without full knowledge of the rea- quickly and perhaps indiscriminately, and to experience frequent sons, and anxious individuals are highly emotionally expressive breakups and reunions. Avoidant adults may or may not be but often cannot regulate their emotions or emotionally driven interested in close relationships, but nevertheless they are un- behavior effectively in line with personal interests or social comfortable with closeness, are disinclined to become involved norms. in long-term romantic relationships, are uncomfortable with Consistent with these notions, a number of recent studies self-disclosure, and are relatively inhibited and socially un- provide evidence of attachment style differences