Since Newton (1642-1727) continuous time modeling by means of differential equations is the standard approach of dynamic phenomena in natural science. It is argued that most processes in behavioral science also unfold in continuous time and should be analyzed accordingly. After dealing with the essentials of stochastic differential equation modeling of panel data by means of structural equation modeling (SEM) and the exact discrete model (EDM) as well as with the relevance of cross-sectional research for continuous time modeling (Coleman), several paradoxes speci cally related to the cross-lagged panel design in discrete time are addressed and solved in continuous time. These paradoxes have been the cause of a lot of confusion in the interpretation of cross-lagged panel analyses in the literature. Educational research data illustrate and evaluate continuous time modeling of cross-lagged effects by means of different models and methods using SEM.

Item does not contain fulltext