> If you actually want to show something, do a longitudinal study: first of all, pick cohorts that match up in most other respects, then compare their relative

> If you actually want to show something, do a longitudinal study: first of all, pick cohorts that match up in most other respects, then compare their relative performance within the cohort after increasing their DEI numbers. Or not doing so. I agree these study parameters would yield much better data. The construction of such a study would probably prove challenging.