Biosocial models of adolescent problem behavior: extension to panel design

S M Drigotas,J R Udry
DOI: https://doi.org/10.1080/19485565.1993.9988831
Abstract:We extended the biosocial model of problem behavior tested by Udry (1990) to a panel design, following a sample of over one hundred boys in adolescence for three years. We found the expected results for sociological variables, but weaker effects for testosterone than Udry found on cross-sectional data. Using panel models with lagged hormone effects, we identified relationships between Time-1 testosterone and problem behavior one year or more later. The relationship between testosterone and problem behavior was not present for subsequent measures of testosterone, either in cross-section or with time-lagged models. Therefore we cannot interpret the results as showing testosterone effects on problem behavior. Rather it appears that testosterone level in early adolescence is a marker for a more general growth trajectory of early development.
What problem does this paper attempt to address?