As noted earlier, IDA is not so much a set of statistical techniques as it is a framework for delineating, controlling for, and exploring sources of between-study heterogeneity in order to create commensurate measures for and test hypotheses in a pooled data analysis. For this reason, the statistical techniques used for hypothesis testing in IDA may draw from many traditions, but they should share the capacity to account for study differences at all plausible points in the modeling sequence. As a result, hypothesis testing in IDA may be challenging due to necessary model complexity, particularly in the context of longitudinal study designs. However, accounting for such between-sample heterogeneity is essential to valid inference testing.
