Test–retest reliability of self-reported sexual health measures among US Hispanic adolescents.

Objective: Although Hispanic adolescents in the USA are often the focus of sexual health interventions, their response to survey measures has rarely been assessed within evaluation studies. This study documents the test–retest reliability of a wide range of self-reported sexual health values, attitudes, knowledge and behaviours among Hispanic adolescents from low-income, urban communities. Method: Intraclass correlation coefficients were calculated for sexual health scales (knowledge, values about condom use, parent communication, protection self-efficacy and condom-use intentions) in a 2-week sample (N = 86). Cohen’s kappa and the intraclass correlation coefficients were calculated for behaviours (romantic and sexual behaviours, and sexual health services history) in two separate 2-week samples (N = 86, N = 202). Selected behaviours were also examined for retractions in both 2-week samples and in a 1-year sample (N = 291). Results: Among sexual health scales, test–retest reliability ranged from .37 for condom-use intentions to .93 for parent communication. Among behaviours, reliability ranged from 0.27 for use of sexual health services to 1.00 for use of birth control at first sex. Retractions ranged from below 10% for ever having had a steady partner to 50% or more for sexual health services and testing. Conclusion: Test–retest reliability differed greatly across self-reported sexual health values, attitudes, knowledge and behaviours for the three samples of Hispanic adolescents. Overall, however, most items and scales showed good to excellent reliability (i.e. a coefficient greater than .60), suggesting that Hispanic adolescents are moderately to highly stable reporters in response to sexual health measures. [ABSTRACT FROM AUTHOR]