Age Bias, Missing Data, and Declining Response Rates in the National Youth Risk Behavior Survey and Their Influence on Estimates of Trends in Adolescent Sexual Experience, 2011-2023
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Introduction
The national Youth Risk Behavior Survey (YRBS) has experienced considerable declines in response rates, increases in missing data on sexual experience, and shifts in data collection – all of which raise questions about bias in estimates for sexual experience among US high school students.
Methods
We used weighted data from the YRBS for 2011-2023 (n=110,409). We explored the impact of declining school and student response rates, missing data, and shifts in age distribution in 2021 on reported sexual intercourse. Using statistical decomposition, we estimated the percentage change in this outcome between 2019 and 2021 due to change in the age structure versus changes in reported behavior.
Results
From 2011 and 2023, school and student survey response rates declined (school 81% to 40%, student 87% to 71%, and overall, 71% to 35%). Missing data on ever had sex increased over time from 7.0% in 2011 to 29.5% in 2019 and 19.8% in 2023. The age structure in the YRBS national sample was similar from 2011-2019 and in 2023, but substantially younger in 2021. Statistical decomposition estimated that 50% of the change in sexual experience among adolescent women between 2019 and 2021 and 30% of the change for adolescent men was due to a change in the age distribution.
Implications
Declining response rates, increased missing data, and changes in the age structure of 2021 YRBS raise serious concerns about the validity of trends in the YRBS. A concerted national effort is needed to build support for the collection of YRBS and other public health surveillance data.
