A final consideration in planning IDA within primary data collection is in assessing control variables for hypothesis testing. To the extent that variables can be directly modeled and controlled in hypothesis testing, we can differentiate their impact on study outcomes and predicted associations from the influence of study membership. For this reason, some initial planning regarding important ways in which individuals within and across the contributing samples may differ from one another can identify variables that should be included in the common battery across studies. For example, individual studies may overlap in the range of SES sampled for their participants, but some studies may cluster on the high-end whereas others cluster on the low-end. To the extent that SES is correlated, though not wholly confounded, with study membership, we can unpack the influence of study membership and SES on outcomes and predicted associations if we have a commensurate measure of SES across studies to include in our models that test our hypotheses.
