Measurement invariance in growth models of substance use

University Of North Carolina Chapel Hill, Chapel Hill NC

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): The purpose of the proposed project is to explore new methods for testing measurement invariance in latent growth models (LGM). LGMs are a collection of statistical methods with the aim of modeling complex systems of behavior such as substance abuse, AIDS and HIV prevention, functional brain imaging, and academic achievement over time. However, much of the research utilizing LGMs has been conducted by modeling individual sum scores over a series of assessments; although at times appropriate, this technique necessarily assumes that the underlying measurement model remains constant over assessments. That is, it is assumed that the condition of measurement invariance has been retained. Although there is extensive literature discussing the importance of measurement invariance in longitudinal research, little quantitative work has been performed examining the appropriateness of this assumption within the LGM framework. In fact, I am unaware of any empirical research specifically exploring the minimum measurement invariance conditions required to validly apply a LGM. This is critically important because failure to maintain measurement invariance in practice can reduce the internal validity of a study as well as lead to biased or misleading results. Thus, the proposed study will aim to 1) examine the impact of violating the assumption of measurement invariance within LGM; and 2) explore the necessary conditions for the retention of measurement invariance in practice. The results of the proposed project will help insure greater validity in applied substance use research.