I have to disagree with both of you. The original criticism was incorrect, because this study does control for income and other socioeconomic stuff. But "correlation isn't causation" is still a valid criticism of this study. To prove causation you would have to have some random effect controlling where people live.

Just because it's peer-reviewed doesn't mean the article proves what it claims to. Unless you have a natural experiment it's hard to prove causation in situations like these. So you get a lot of published research that just proves correlation.