Separating age, cohort, and time of measurement: analysis of variance or multiple regression.

This paper empirically compares the relative advantages of analysis of variance (ANOVA) and multiple regression (MR) approaches to the separation of age, cohort, and time of measurement effects in sequential research designs. The comparison utilizes four synthetic data sets, designed to have specific characteristics. The results support Adam's recent claim that standard ANOVA procedures, as described by Schaie, do not permit the development of accurate decision rules for age-period-cohort analysis. A modified dummy variable regression procedure developed by Mason, et al. is demonstrated to permit accurate attribution of variance among age, cohort, and time of measurement effects in the developmental model.