Models for Hierarchical Structures in Differential Psychology

This paper provides descriptions and applications of models for investigating hierarchical structures in differential psychology. These models are sub-classified into common- source, person-situation and state-trait models according to the characteristic restrictions of the relationships between manifest variables and combinations of manifest and first-order latent variables. The models include one level of manifest variables and two or more levels of latent variables. It is the advantage of this framework that a few characteristic models can be adapted to a large number of structures by minor modifications. Cross-sectional and longitudinal approaches are suggested for investigating different aspects of the person-situation issue. While the cross-sectional approach is concerned with behavior in different situations, the longitudinal approach deals with behavior observed at different occasions. Structural equation modeling provides the framework for the formal representation of these models. Some models are applied for investigating the structure of social optimism. For a sample of N=462 participants it is demonstrated that the model including three first-order latent variables and one second-order latent variable provides an appropriate representation of the data.