These recommendations for fixed-effects IDA are drawn once again from our experience with the Cross Study (see Table 1 for sample description). Although we do not provide analytic details regarding the steps in hypothesis testing here given space constraints, we do consider the value added of an IDA approach for hypothesis testing based on applications from this work. For example, IDA allowed us to examine trajectories of internalizing and externalizing symptoms with a larger sample size and over longer periods of development than captured in any one of our contributing studies (Hussong et al., 2007; Hussong, Flora et al., 2008). Moreover, IDA allowed us to compare subgroups of children with small representation in the contributing studies to determine, for example, the relative risk for symptomatology among children of antisocial alcoholic parents versus children of depressive alcoholic parents. Given that these sub-populations are rare, important distinctions in the risk profiles of these groups of children of alcoholic parents could not be tested in contributing studies and were only evident in the larger and more highly powered pooled data analysis.
