An Examination of the Reliability,Data Screening Procedures,and Extreme Response Patterns for the Youth Risk Behavior Surveillance Survey

Abstract

This article explores psychometric characteristics of the Youth Risk Behavior Surveillance Survey (YRBS), one of the most widely used instruments to assess the prevalence of violent and other high-risk behaviors in secondary school settings. Response patterns were analyzed for a subset of 414 youths who indicated that they had carried a weapon to school six or more times during the preceding 30 days, and were compared to a matched sample of youths randomly selected from the 13,610 participants in the 2001 national administration of the YRBS. Results indicated that extreme responders to the weapon-carrying item were considerably more likely than their counterparts to respond in an extreme fashion to all YRBS items, including risk items pertaining to school (e.g., physical fights on school property) and outside school (e.g., alcohol consumption) factors, as well as positive health behaviors (e.g., healthy eating habits). Overall, the results suggest an ex-treme response bias among some participants that may impact the validity of the YRBS instrument. More specifically, presence of this response bias may inflate estimates of the prevalence of school violence and related concerns. These findings are discussed in light of the need to carefully examine individual response patterns on future administrations of the YRBS in an effort to ensure maximum instrument utility.