RESUMO
PURPOSE: The Youth Risk Behavior Survey (YRBS) monitors behaviors, experiences, and conditions affecting the health of high school students nationwide. This study examined the test-retest reliability of the 2021 national YRBS questionnaire. DESIGN: Respondents completed a Time 1 and Time 2 paper-and-pencil questionnaire approximately 2 weeks apart during February to May 2022. Data were linked in such a way as to preserve anonymity. SETTING: Convenience sample of high schools. SUBJECTS: High school students (N = 588). MEASURES: Health risk behaviors and experiences assessed on the 2021 national YRBS questionnaire. ANALYSIS: Time 1 and Time 2 responses were compared for each questionnaire item using the McNemar's test. Then, Cohen's kappa coefficients tested the agreement between Time 1 and Time 2 responses overall, and by sex, grade, and Black, White, and Hispanic race and ethnicity. RESULTS: Among the 74 items analyzed, 96% had at least moderate reliability, and 73% had substantial or almost perfect reliability. The mean Cohen's kappa was .68. McNemar's test findings showed Time 1 and Time 2 data significantly differed (P < .01) for 9 items (12%). CONCLUSION: Reliable health behavior measures are important in the development of youth-focused public health programs and policies. Findings suggest the national YRBS questionnaire is a reliable instrument. Such findings lend support to relying on adolescent self-reported data when monitoring health behaviors using the YRBS.
RESUMO
BACKGROUND: School Health Profiles assesses school health policies and practices among US secondary schools. METHODS: The 2020 School Health Profiles principal and teacher questionnaires were used for a test-retest reliability study. Cohen's kappa coefficients tested the agreement in dichotomous responses to each questionnaire variable at 2 time points. The aggregate prevalence estimates between time 1 and time 2 were compared for each questionnaire item via overlapping 95% confidence intervals. Chi-square tests examined whether the prevalence at time 2 differed between paper and web administration for both questionnaires. RESULTS: For the principal (N = 50) and teacher (N = 34) data, there were no significant differences in the prevalence of any items between time 1 and time 2. For the principal survey, the mean kappa for 191 variables was 0.49. For the teacher survey, the mean kappa for 260 variables was 0.65. Overall, 60.7% of principal and 91.1% of teacher questionnaire items had at least "moderate" reliability. CONCLUSIONS: School Health Profiles offers education and health agencies a reliable tool to monitor school policies and practices.
