The application of the Things I worry About Scale to a sample of at-risk American adolescents: an examination of psychometric properties

Adolescence, Winter, 2007 by Irvin G. Esters, Anne Tracey, Rob Millar

The need to gain a better understanding of what adolescents worry about has been the focus of several studies (Anttila, Poikolainen, Uutela & Lonnqvist, 2000; Boehm, Schondel, Marlowe, & Manke-Mitchell, 1999; Esters, 2003; Gallagher & Millar, 1996; Millar & Gallagher, 1996; Porteous, 1979). Among the applications of the results of such studies is the augmentation of comprehensive needs assessments, including the content and degree of adolescent worries and concerns. The identification and quantification of worry can provide an invaluable resource to those responsible for the development of relevant and appropriate curriculum programs and guidance and counseling services within educational settings and beyond (Esters, 2003; Gallagher & Millar, 1996).

One attempt to ascertain what adolescents in Northern Ireland worried about and how frequently they did so gave rise to the development of the Things I Worry About Scale (Millar & Gallagher, 1996) hereafter referred to as the scale. As such, the aim was to design a survey instrument capable of reflecting personal, social, academic-related, and career worries (Millar, Gallagher, & Ellis, 1993). From a thorough review of existing instruments and the results of studies that investigated self-disclosed adolescent needs, Gallagher, Millar, Hargie, and Ellis (1992) initially formulated an 80-item scale. Pilot testing of the instrument with 28, 15 to 18-year-old students, produced a revised version of the scale consisting of 86 items reflecting eight a priori categories, namely, (1) At School/College; (2) Choosing a Job; (3) Job finding; (4) At Home; (5) Starting Work; (6) Opposite Sex; (7) Myself and Others. In order to examine the utility of the instrument in reflecting the worries and concerns of adolescents, Millar et al. (1993) sampled 378, 15-and 16-year-olds, across 10 schools in the North of Ireland. From a rigorous statistical analysis of the rating scale responses and a content analysis of 269 open-ended responses generated by 35% of the participants, a revised version of the instrument was compiled. In the study, 74% of the open-ended responses suggested additional constituent items for each proposed category and three new categories: (1) Money Matters, (2) School/College Work, and (3) Change. The revised scale was comprised of 138 items reflecting 13 content categories. Additional testing was necessary, however, to estimate, first, the reliability of the scale scores with an adolescent sample, and second, whether the content adequately represented the range of adolescent worries. A further and more thorough examination of the instrument was derived through a substantially larger study in which Millar and Gallagher (1996) surveyed 3,983 students ranging from 13 to 18 years of age who were attending schools in Northern Ireland.

The main findings from the study confirmed the utility of the scale as a reliable and valid indicator of the types and frequencies of adolescent worry in Northern Ireland. When the responses to the 138 items were analyzed, it was found that the factors identified in Millar et al. (1993) were generally supported, although some revision was necessary. The internal reliabilities of the scores on the 13 clusters were represented by alpha coefficients that ranged from .80 to .90 and by .98 for the total scale score (Millar & Gallagher, 1996). The content validity of the instrument was endorsed by 87% of the adolescents in the study who suggested that the scale items enabled them to express all of their major concerns. In the modifications that were required, two categories disappeared (Social Confidence and Powerlessness) and one, Social Efficacy, was added, reflecting concerns related to difficult social situations. The category At Home, included in the initial scale (Millar et al. 1993), was split into two distinct categories, Home Relationships and Communication at Home, in the revised scale.

Demographic differences in frequency of adolescent worries have been reported in studies that have employed the scale. One of the most consistent findings has related to the higher frequency of expressed worries among female, as compared with male adolescent respondents (Gallagher & Millar, 1998; Gallagher, Millar, Hargie, & Ellis, 1992). The reported differences have also been pervasive and generally evident in all categories of worry sampled. It was predicted that strong and pervasive sex differences would be evident in the present study, with females expressing significantly higher rates across all categories than their male counterparts.

Evidence has also been reported indicating age differences in self-reported worries among adolescents. The exact nature of the relationship between age and frequency of reported worry appears to be dependent on the age range of participants sampled. Where the ages ranged from 12 to 18 years, the trend approximates a negative linear relationship (Gallagher & Millar, 1998; Millar & Gallagher, 1996). However, there is evidence to suggest that the relationship may be characterized more as curvilinear with frequency of worry increasing in early adolescence up to a peak at around 16 years of age, and then declining through later adolescence (Gillies, 1989; Porteous, 1985). Given the age range of participants in the present study, it was predicted that reported frequency of worry would be greater in older respondents.