Research studies in the alcohol field are concerned with understanding the causes and effects of alcoholism, alcohol abuse, and alcohol-related problems for the purpose of diagnosing, testing, and preventing such problems. Alcohol studies often formulate questions in terms of individual development over time and investigate these questions using repeated measurements on groups of individuals. Currently, the generality of the methodology and statistical software available for developmental studies is limited in comparison to the complexity of the theories that need to be tested. Most researchers examine longitudinal data using random coefficient models such as the multilevel model and the mixed linear model. The possibility of doing longitudinal modeling with random coefficients within the latent variable structural equation modeling framework has recently been proposed. Both approaches have their strengths and weaknesses, and neither alone is sufficient for the analysis of longitudinal data. This SBIR Phase II contract will develop a statistical software program designed specifically for the longitudinal analysis of alcohol data.
