Document Type

Dissertation

Degree

Doctor of Philosophy

Major

Psychology

Date of Defense

12-17-2009

Graduate Advisor

Michael G. Griffin, Ph.D.

Committee

Matthew Taylor

Barbara Bucur

Mary Dugan

Abstract

Risk and protective factors for adolescent drug and alcohol abuse have gained prominence in the prevention field as a framework for prevention. Communities that Care® is the original survey that collected data on risk and protective factors in the adolescent community. The Missouri Student Survey is based upon the Communities that Care® survey and is administered to students across the state of Missouri every two years. This study examines the reliability and validity of the Missouri Student Survey in a sample of 126,923 students from across the state of Missouri. In addition, this study also considers the question of active versus passive consent and its influence on the generalizability of the data. Examining these issues will help illuminate the strengths and weaknesses of the survey as well as present some options for increasing the data quality in subsequent administrations. A confirmatory factor analysis revealed problems with the assumed factor structure of the survey (as shown in Appendix 1) so an exploratory factor analysis was performed to assess the possibility of an alternate factor structure. However, most of the original risk and protective factor scales demonstrated strong reliability and criterion-related predictive validity. An item analysis determined that students were less likely to complete the questions in the second half of the survey and a secondary analysis established that certain school districts were driving this pattern. Finally, an examination of the issue of active versus passive consent showed that passive consent indeed increased the participation rates as well as decreased the rates of students reporting using a fake drug. Evaluation of the effect of active or passive consent revealed mixed results for the number of risk factors reported. Implications of these findings are that the underlying factor structure should be re-examined with the 2010 data set before reports are generated. However, passive consent appears to be worthwhile and should be continued. Future directions and limitations of this study are discussed.

Included in

Psychology Commons

Share

COinS