After our extensive qualitative interviews but before developing the final questionnaire, we engaged in several additional stages of data collection. First, we worked with six of the younger girls interviewed during the qualitative study to establish the comprehensibil- ity, readability, and relevance of our initial items for girls under 15. This step was especially important for the item development of the new SSI scale. Next we conducted several pretests, first establishing readability with 27 adolescent girls (using an athletic youth group as respondents) and then pretesting the full survey with 60 girls between the ages of 12 and 19. This pretest sample was obtained by contacting several youth church groups and attending their meetings, as well as by recruiting teens outside a retail chain focus- ing on adolescents in a local mall. To ensure variance in the responses, approximately half of the sample evalu- ated stores they had shopped at “several times before,”
while the other half evaluated stores they had shopped at “only a few times before.”
We analyzed the pretest data for internal consistency with coefficient alphas and exploratory factor analysis (with principal axis factoring and oblique rotation) to assess the reliability of each construct and its loadings (Bearden, Netemeyer, and Teel 1989; Churchill 1979). Items with low loadings or cross-loadings were deleted, with the items maintained for the main study all pro- ducing reliable scales. Further details are provided below.