ABSTRACT
Transitions in families are frequently marked by specific dates such as birth of a child, marriage, divorce, or death. Such transitions, however, are frequently of long duration, and their effects may be variable over long periods of time. The transition to parenthood, for example, begins before the actual birth of the child and the effects continue to change for at least 18 months (Cowan, Cowan, Heming, & Miller, this volume). Recognition of these over-time processes has lead to increased interest in the use of longitudinal research designs. Data generated by these designs create unique data analysis problems which have been identified as problems in measuring change (Harris, 1963). These problems range from questions regarding the reliability of difference scores to the conceptual definitions of change itself.
