|
Traditional Approaches for Longitudinal Data
|
| Presenter(s):
|
| Mende Davis, University of Arizona, mfd@u.arizona.edu
|
| William Becker, University of Arizona, beckerwj@email.arizona.edu
|
| Abstract:
Evaluation studies have measured outcomes at multiple time points for several reasons. For evaluation purposes, multiple time points allow us to at least obtain some data in between, and how participants change over time. In traditional approaches, such as t-tests and analysis of variance, the dependent variable will be the change score or post-test score. When a study has multiple time points and covariates, these simple analyses can only provide results regarding two time-points (e.g., pre-post). With a complete dataset without any missing data, a pre-post approach may provide us with a screen-shot of how people change from the first time point to the last time point. There are riches to be found in longitudinal data; however, using two observations at a time does not do these data justice. In this presentation, we will cover the traditional approaches for longitudinal data.
|
|