Related Experiment Videos
Planning for the appropriate analysis in school-based drug-use prevention studies
1Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis 55455.
Journal of Consulting and Clinical Psychology
|August 1, 1990
Summary
School-based drug prevention studies face analysis challenges due to intragroup dependence. Researchers must use alternative methods and recalculate sample sizes for accurate school-based drug use research.
Area of Science:
- Public Health
- Biostatistics
- Educational Psychology
Background:
- School-based drug-use prevention studies frequently implement interventions at the school level.
- Assigning intact social groups (schools) to study conditions can lead to intragroup dependence, complicating data analysis.
Purpose of the Study:
- To highlight the impact of intragroup dependence on statistical analyses in school-based drug-use prevention studies.
- To review alternative analytical methods and sample size calculations that account for intragroup dependence.
Main Methods:
- Analysis of data from two school-based drug-use prevention studies.
- Calculation of intraclass correlation coefficients (ICCs) for common drug use measures.
- Review and illustration of alternative statistical methods and sample size computations.
Main Results:
- Intraclass correlation coefficients between 0.02 and 0.05 were found for drug use measures, indicating modest intragroup dependence.
- Traditional fixed-effects analyses may be invalidated by this level of dependence.
Conclusions:
- Researchers must adopt alternative statistical methods that acknowledge intragroup dependence in school-based studies.
- Sample size requirements may be substantially greater when accounting for intragroup dependence, necessitating careful study planning.