This document is part of a series of chapters described in SO 011 759. This chapter offers guidelines for studying change in dynamically interdependent variables, even when a model that ignores interdependence is estimated. Many sociological studies concern dynamically interdependent variables; for example, in a study of female employment, a woman's employment may influence and be influenced by her marital status. The report addresses these issues: Under what circumstances do inferences based on a misspecified model (one that ignores interdependence among variables) lead to errors that are comparatively small and what kinds of research designs minimize errors and what kinds accentuate them? Separate sections examine assumptions of models, implications of bivariate and univariate models, research design, and procedures and findings for the 28 bivariate processes considered. Results indicate that biases from ignoring dynamic interrelationships tend to increase with the degree of interdependence of the variables. However, under some circumstances, biases are not large when interdependencies are ignored. If the variable being studied changes more slowly than the ignored variable, biases are relatively small. For example, research on marital stability may ignore changes in employment status, but not the reverse. Advantages of using event-history and panel analyses are also discussed. (Author/KC)