In many ways, measurement is the core challenge to IDA. One must be able to identify or construct commensurate measures of predictors and outcomes for a pooled data analysis to be sensible. We have described a variety of measurement scenarios that might be encountered in IDA, including situations in which identical, highly similar, or truly distinct items are present in the original studies. Ideally, multiple items will be available from each study for a given scale, and a subset of these should be common items (i.e., either identical or harmonized items). Under these conditions, we have outlined and illustrated a series of steps and procedures that we have found useful in constructing commensurate measures. We have also described the use of bridging studies for enabling IDA when no common items are initially available.
