Anxiety sensitivity (AS) is the fear of anxiety-related sensations due to beliefs that such sensations will lead to catastrophic outcomes such as physical illness, social embarrassment, or mental incapacitation (Reiss & McNally, 1985). Longitudinal research supports AS as a vulnerability factor in anxiety disorders development (e.g., Schmidt, Lerew, & Jackson, 1997). Evidence from factor analytic studies on the adult Anxiety Sensitivity Index (ASI; Peterson & Reiss, 1992) supports a hierarchical structure in which three intercorrelated lower-order AS factors (i.e., Physical, Social, and Psychological Concerns) load onto a single higher-order Global AS factor (see Zinbarg, Mohlman, & Hong, 1999 for review). A similar hierarchical structure has been found in recent factor analytic studies of the Childhood Anxiety Sensitivity Index (CASI; Silverman, Fleisig, Rabian, & Peterson, 1991), although there is debate on whether there are two, three, or four lower-order factors (Chorpita & Daleiden, 2000; Muris, Schmidt, Merckelbach, & Schouten, 2001; Silverman, Ginsburg, & Goedhart, 1999).
In adults, women score higher on average on the ASI than men (Peterson & Reiss, 1992). While similar hierarchical factor structures can be extracted from the ASI in men and women, there are gender differences on the particular dimensions of AS in both nonclinical (Stewart, Taylor, & Baker, 1997) and clinical populations (Stewart & Baker, 1999). For instance, Stewart et al. (1997) found that women scored higher than men on the lower-order Physical Concerns factor but not the Social or Psychological Concerns factors of the ASI. In addition, within-gender comparisons showed that women score higher on Physical Concerns relative to their scores on Social and Psychological Concerns dimensions. In contrast, men score higher on Social and Psychological Concerns dimensions relative to their scores on Physical Concerns.
In children, girls score higher overall than boys on measures of childhood AS (e.g., Deacon, Valentiner, Gutierrez, & Blacker, 2002; Muris et al., 2001; van Widenfelt, Siebelink, Goedhart, & Treffers, 2002). Moreover, consistent with adult studies, Deacon et al. (2002) found that girls scored higher than boys on a Physical Concerns subscale, but not a Psychological Concerns subscale, derived from factor analysis of a 12-item ASI for Children (ASIC). In contrast, van Widenfelt et al. (2002) found that girls had higher scores than boys on all three CASI subscales (i.e., Physical, Psychological, and Social Concerns).
To date, no studies have examined CASI factor structure as a function of gender and none have compared girls and boys on CASI dimensions using factor scores as opposed to subscale scores (cf., Stewart et al., 1997).1 Examination of the factor structure as a function of gender would ensure that the previously observed gender differences in childhood AS are due to actual differences in the relative focus of concerns between girls and boys and not due to gender differences in CASI factor structure. Moreover, it would not be appropriate to compare CASI dimension scores for boys versus girls if the structure of the scale differs across gender.
There were three main goals in the present study. The goals were to examine: (1) the hierarchical structure of the CASI in a large nonclinical sample of children and adolescents; (2) whether similar lower-order and higher-order factor structures could be extracted across genders; and (3) whether there are gender differences in higher-order and lower-order AS factor scores. Consistent with several recent factor analytic studies, a three- or four-factor lower-order solution was predicted with lower-order factors reflecting Physical, Psychological, and Social and/or Control Concerns. Moreover, it was hypothesized that these lower-order factors would load onto a single higher-order Global AS factor. It was also hypothesized that a similar hierarchical factor structure would be observed in girls and boys, and that girls would score higher than boys only on the higher-order Global AS and the lower-order Physical Concerns factors. Within-gender groups, girls were hypothesized to score higher on lower-order Physical Concerns than Social and Psychological Concerns factors, while boys were hypothesized to score higher on Social and Psychological Concerns factors than on Physical Concerns (cf., Stewart et al., 1997). No predictions were made for gender differences on the possible fourth factor of Control Concerns.
