Construct validation of the health belief model

Abstract

A multitrait-multimethod design was employed to assess the construct validity of the Health Belief Model. The data were obtained from a nonrepresentative sample of 85 graduate students at The University of Michigan's School of Public Health. The traits consisted of the respondents' perceptions of: health interest, locus of control, susceptibility to influenza, severity of influenza, benefits provided by a flu shot, and the barriers or costs associated with getting a flu shot. Each trait was measured by three methods: a seven-point Likert scale, a fixed-alternative multiple choice scale, and a vignette. The results indicate that the Health Belief Model variables can be measured with a substantial amount of convergent validity using Likert or multiple choice questionnaire items. With regard to discriminant validity, evidence suggests that subjects' perceptions of barriers and benefits are quite different from their perceptions of susceptibility and severity. Perceptions of susceptibility and severity are substantially but not entirely independent. Perceived benefits and barriers demonstrate a strong negative relationship which suggests the possibility that these two variables represent opposite ends of a single continuum and not separate health beliefs. These preliminary results provide the basis for developing brief health belief scales that may be administered to samples of consumers and providers to assess educational needs. Such needs assessment, in turn, could then be used to tailor messages and programs to meet the particular needs of a client group.