Relationships Among Latent Growth Curve Model, Statistic Power, and Sample Size

In many social and behavioral sciences, longitudinal study has become an important research method. Since the data is collected more than one specific point of time, the longitudinal study allows researchers to realize the growth trajectories of the individual in an organization, and even the changes in the whole organizational behavior. Also, the researcher can get the support of the causal relationship for his theory hypothesis by identifying the temporal precedence of variables. Moreover, the data collected at different points in time may avoid common method variance. Although the longitudinal study has such advantages, some researchers have misused it and made their model structure too complicated. Adding too many temporal differences in their model may conclude in the wrong results. Random Coefficient Model, one of the longitudinal study methods, is apt to have an incorrect model specification and result in a spurious regression relationship. This may inflate the Type I error and cause spurious results. In the present study, we attempt to apply the framework of the Latent Growth Curve Model as the basis of our research design. We also use Monte Carlo simulation to calculate the sample size and to test statistic power and effect size. At last, hopefully, we may provide a different aspect to the researchers who are intended to conduct longitudinal studies.