Third is about reviewing. Is the claim sized right for the evidence? What else could explain this finding? Does the measure actually capture the construct? Connect the claims, interpretations with evidence based on strength. (minimal interaction)