Developmental research often entails cross-sectional or longitudinal designs involving such categorical variables as early, on-time, or late maturation, family type, or problem status. In order to identify and extend data analytic methods useful in such studies, this article discusses the use of design matrices for log-linear analysis of developmental hypotheses in cross-sectional and longitudinal designs. The »-parameters in log-linear models are introduced as weights for vectors of a design matrix, X. In exploratory or computer algorithm-controlled application of log-linear modeling, X is set up by the computer program, irrespective of the researcher’s specific hypotheses concerning main effects and interactions. The chapter describes means to create design matrices to model hypotheses about development during any period of the life span. Data examples illustrate the analysis of cross-sectional designs, therapy outcome, and the testing of trends. The article also discusses strategies of log-linear modeling and problems with the interpretation of the parameter size.