How can I prove that the attrition bias affects or does not affect my OLS estimator?
Can I do it with a balance table between the group of which I have information and with the one of which I lost information? How would you interpret that balance table? (I have complete data of both groups in other variables)