Does Being Old for Grade in High School Predicts Sexual Risk Behaviors, Beyond Grade-Level Effects?

The prevalence of adolescent risk behaviors varies by age and grade level in school, both of which are proxies for physical and social development. Adjusting for both age and grade would be ideal but is ill advised because of collinearity. We developed a variable to assess "relative age" (i.e., old for grade vs. age normative) and estimated grade and old-for-grade status in association with sexual risk behaviors.

Data are from the 2017 National Youth Risk Behavior Survey. We conducted weighted multivariable logistic regression models and reported prevalence ratios to estimate associations between grade, old for grade, and sexual risk behaviors among 9th-11th graders.

Being old for grade (vs. age normative) was associated with a higher prevalence of sexual risk behaviors, particularly for ninth graders.

Adjusting for relative age is a feasible way to increase precision in estimation of grade-related differences in sexual risk behaviors.