Our statistical approach permits simultaneous analysis of data drawn from the two longitudinal studies. These methods have recently been referred to as Integrative Data Analysis (Curran & Hussong, 2009). As noted in Table 1, the MLS and AFDP samples differ in a number of respects on key study variables. However, these differences do not undermine the value of integrative data analysis but rather augment it. By combining samples we increase heterogeneity among participants in developmental range, cohort representation, sampling and measurement strategies, and demographic features of the participants. We are careful to take into account these differences in harmonizing measurement (i.e., making our measures comparable across studies) and in conducting inferential tests of our hypotheses. Importantly, by combining the samples and including tests of differences in how our predictor variables relate to outcomes as a function of sample membership, we are able to directly test the generalizability of our findings across the two samples (akin to meta-analysis). These two features (increasing sample heterogeneity and testing generalizability of findings across the samples) are key advantages of integrative data analysis. The combined
