Leveraging latent growth models to better understand MIS theory: a primer

Abstract

The purpose of this didactic study is to demonstrate how latent growth models (LGM) can be utilized to measure changes in student's computer self efficacy (CSE) over time. LGM is a special application of structural equation modeling (SEM), an analytic tool that is popular among MIS researchers. LGMs have been used to study longitudinal changes in observed and/or latent variables over time in several other fields such as psychology, sociology, and management. To promote its use within MIS research, this paper provides a primer on the application of LGM using CSE data gathered from freshmen enrolled in the introduction to MIS class. We illustrate unconditional and conditional LGMs, and highlight the types of research questions such models can address. We discuss issues related to data requirements, model identification, estimation methods, sample size requirements, and model fit assessment statistics for LGMs, and conclude by providing avenues of further longitudinal research in MIS that can benefit from the use of LGMs.