Hacker News Reader

The study basically compared neighborhoods that received a Starbucks versus neighborhoods that did not. What if Starbucks decided to open in those neighborhoods because of some confounding factor that suggested an increase in economic activity?

View on HN

> We find that, when compared to census tracts that were scheduled to receive a Starbucks but did not do so

There still might be a confounders, but they at least tried to control for that.

Except how do we know that the reason they didn’t do so was precisely because of that confounding factor? It’s a neat idea but leaves one wanting for better data

Makes me wonder why Starbucks planned to open there, but didn't... Maybe a weak economy in the area

IIRC (I read this study a few weeks ago), they compared areas where a Starbucks opened; areas where Starbucks planned an expansion but were prevented by things like city planning snafus; and areas where Starbucks opened in non-traditional (eg low income) areas through their partnership with Magic Johnson.

Reply on news.ycombinator.com