To me, the "compared to census tracts that were scheduled to receive a Starbucks but did not do so..." is a weak défense against the correlation argument. Like

To me, the "compared to census tracts that were scheduled to receive a Starbucks but did not do so..." is a weak défense against the correlation argument. Like you said, without knowing the why of the change in plans, there is no way to judge causality confidently. Anecdotally, I love these types of cafes (not "Starbucks" but this model), hope we get more, and see huge benefit both for myself and the community.