Prof. Christian Dormann
Professor Johannes Gutenberg-University Mainz
Longitudinal designs provide the opportunity to better understand psychological stress processes over time. As various types of longitudinal designs have become increasingly common, several scholars have suggested that the adoption of a temporal approach is needed for the advancement of our field. However, confusion exists regarding the best combination of designs and statistical analyses for testing different sorts of hypotheses that propose temporal relations. In this presentation, an overview is provided of different commonly applied analyses, their assumptions regarding the role of time, the kind of data/design required for their applications, the interpretations that can be derived, and their limitations. Special emphasis is placed upon the mutual influence of different psychological processes across time, that is, their possible reciprocal effects and moderated effects. Although temporal effects can be investigated using various designs and analyses, I also discuss new developments and their advantages, such as designs with interindividually and intraindividudally varying time intervals, and continuous time structural equation modeling (ctsem), which is a very useful extension of common (discrete time) structural equation models.
