Wednesday, June 20, 2012

letter to the Sun: liquor store and crime correlation

In which I once again link all of our city's problems to a lack of STEM education:


Dear Baltimore Sun,

The city health officials who plan to strip non-conforming liquor stores of their licenses because of the Johns Hopkins University study linking them to violent crime (“City Targets Liquor Stores”, June 17 [warning: pay wall]) are confusing correlation with causation.  It could be that the liquor stores themselves are not causing the crime, but rather that some other underlying cause is causing both the non-conforming liquor stores and the violent crime.  The distinction is important, because if the liquor stores are not causing the crime, then closing them would not actually be predicted to decrease violent crime, and could even exacerbate it.

As it turns out, property vacancy also correlates with violent crime (for example, Goodstein, Ryan and Lee, Yan Y., Do Foreclosures Increase Crime? (May 2010). Available at SSRN: http://ssrn.com/abstract=1670842 or http://dx.doi.org/10.2139/ssrn.1670842).  If a non-conforming liquor store loses its license, it seems at least as likely that the store will end up as a vacant property than that the owners will decide to open a healthy food store in its place, as the city officials hope.

Before taking away licenses and, as a consequence, closing small businesses (that would furthermore disproportionately impact a particular ethnic group), the city needs to carefully consider the evidence to determine whether the benefits of doing so really are predicted to outweigh the costs.  This story once again emphasizes the importance of having a basic mathematics education (for no student in the first week of an introductory statistics course would confuse correlation with causation), both for our city officials as well as for our journalists.

No comments: