Testing the Question-Behavior Effect of Self-Administered Surveys Measuring Youth Substance Use

Introduction: Does simply asking a student about their current drug use lead to an increase in drug use one year later? The idea that asking about specific behaviors could elicit the behavior, a question-behavior effect or mere-measurement effect, is often cited for why communities and schools should not administer a survey about youth substance use. The belief in question-behavior effect has created opposition to surveys in schools but there is little evidence to support the effect. This presentation tests the validity of the question-behavior effect in youth outcomes in a longitudinal panel of 4,355 elementary and middle school students surveyed in the 5^th and 6^thgrades.

Methods: The sample of students was drawn from the Community Youth Development Study (CYDS), a community randomized test of the Communities That Care (CTC) prevention planning system.  The initial 5^th grade sample (n=3229) was recruited and surveyed in 2004 and 2005.  A second recruitment effort a year later in 2005 when the initial cohort was in 6^th-grade added an accretion sample of previously unconsented students (n=1126). The accretion sample was surveyed for the first time in 2005.  Responses to questions about substance use on the Communities That Care Youth Survey (CTCYS) were compared for the initial and the accretion samples, controlling for gender, parental education and racial and ethnic background.  If the prevalence of substance use in the initial sample surveyed in 5^th and 6^th grade was higher than in the accretion sample surveyed only in 6^thgrade, the difference could indicate a question-behavior effect. 

Results: Results from logistic regression analyses showed no significant difference between the initial sample and the accretion sample in 30-day or lifetime substance use prevalence of alcohol, tobacco, and marijuana reported in 6^thgrade.    

Conclusions: The results indicate that simply asking youths about substance use in a survey did not increase their substance use as measured by self-reports a year later. The absence of any evidence of a question-behavior effect should ease the concerns of communities and schools when administering surveys measuring youth substance use questions.