The Adolescent Invulnerability Scale: Factor Structure and Construct

This paper reports on the first examination of the factor structure and construct validity of the Adolescent Invulnerability Scale in a sample of young adolescents. Previous research (Duggan, Lapsley & Norman, 2000) using samples of older adolescents showed that the AIS consisted of two factors, “danger invulnerability” and “psychological invulnerability.” These factors also demonstrated adequate internal consistency and construct validity. In the present study we administered the AIS, along with indices of risk behaviors, use of substances, depressive symptomatology and positive adjustment, to 238 young adolescents (Mage = 13.19 years). Principal components factor analysis revealed three factors: general invulnerability, danger invulnerability, and interpersonal invulnerability. Each factor, and the total AIS, demonstrated strong internal consistency. The total AIS score, and each subscale, was positively correlated with delinquent risk behaviors. However, general and interpersonal invulnerability also counterindicated depressive symptoms, and each dimension of felt invulnerability was positively correlated with mastery and coping. Hence invulnerability plays a dual role in adolescent development. The appearance of third “interpersonal” factor in early adolescence is attributed to the relatively greater peer focus evident among younger than older adolescents. Paper presented at the biennial meeting of the Society for Research in Child Development, April 21, 2001, Minneapolis. Introduction Two theoretical traditions suggest that adolescence is marked by a heightened sense of felt invulnerability. The psychodynamic tradition holds that invulnerability is adopted as a defensive posture during separationindividuation (Blos, 1962). Similarly, the adolescent egocentrism theory suggests that teenagers entertain a personal fable of invulnerability as a result of cognitive egocentrism that attends the emergence of formal operations (Elkind, 1967). Both traditions assert that this “fable” impairs the judgement of adolescents in critical situations. Indeed, it is widely believed that teenagers engage in more risk behaviors, and take more risks, just because of their heightened sense of invulnerability (Arnett, 1992), although this view has also been contested (Furby & Beyth-Marom, 1992; JacobsQuadrel et al., 1993). Surprisingly, however, the widely assumed link between adolescent felt invulnerability and risk behavior has never been empirically evaluated, probably because there are no measurement options currently available for assessing the felt invulnerability of adolescents. To address this limitation we have conducted a series of studies in order to develop an Adolescent Invulnerability Scale (AIS) The development of the AIS followed standard scale development procedures. In Phase I an item pool of 56 items was generated by a team of seven late adolescent volunteers. This list was edited for duplicates and grammatical consistency, yielding a final list of 25 candidate items. In Phase II, this list was evaluated by four expert raters, who included a general expert on adolescent development, two nationally visible experts on allied constructs (sensation-seeking and narcissism), and a pioneer in the assessment of personal fable ideation. On the basis of expert commentary three invulnerability items were deleted and 12 were re-written, yielding a 22-item scale. In the third phase of scale development (Duggan, Lapsley & Norman, 2000) the revised AIS was administered to a sample of 228 late adolescents (Mage = 21.85) in order to determine a parsimonious factor structure, and to derive estimates of internal consistency and construct validity. Exploratory factor analysis (extracting principal components) revealed two factors. One factor ( = .85, 12 items) appeared to represent felt invulnerability to external danger, and was therefore labeled “danger invulnerability.” The second factor ( = .79, 9 items) appeared to represent an invulnerability to psychological distress, and was therefore labeled “psychological invulnerability.” The total AIS scale also demonstrated strong internal consistency ( = .87). Duggan et al. (2000) showed, too, that the total AIS score was significantly and positively correlated with risk behaviors (r = .41), as were the “danger invulnerability” (r = .43) and “psychological invulnerability” (r = .26) sub-scales. This supports, of course, the hypothesized relationship between felt invulnerability and the tendency to engage in risk behaviors. Moreover, males reported more feelings of invulnerability that did females. The scale items and factor loadings are reported in Table 1. This first study showed, then, encouraging preliminary support for the psychometric integrity and construct validity of the Adolescent Invulnerability Scale. The fourth phase of research was an attempt to explore the measurement properties of the AIS in a sample of young adolescents. We also wanted to determine whether felt invulnerability would also predict risk behaviors in this sample, including use of substances, and whether felt invulnerability was associated with internalizing symptoms and indices of positive adjustment. This latter question was motivated by theoretical considerations. Lapsley (1993) has argued that “personal fable” ideation might play a dual role in adolescent development. Although such ideation might dispose the adolescent to engage in grandiose risk-behaviors, it might also fortify the adolescent against depressive affect, and even be associated with the sort of “positive illusions” that social psychologists have associated with positive mental health (e.g., Taylor & Brown, 1988).