Methods: The rates of violent crime before and after the opening of the casino in the surrounding neighborhood’s census tracts were compared to rates at similar times in all other census tracts (year 2000) in the City of Pittsburgh. Violent crime data were acquired from the Pittsburgh Police for the years 2005 to 2015. Demographic, social and economic covariates were acquired from the American Community Survey for each census tract. Time-varying confounders accounted for changes within Pittsburgh over the study period and were entered into the difference-in-difference analysis. A difference-in-difference analysis allows for a community comparison group to separate the effect of the casino from other factors. A generalized linear multi-level model was fitted to allow adjustment for confounders.
Results: Violent crime declined by 12% in Pittsburgh, however there was no observable decline in violence in the census tracts surrounding the casino. Therefore, violent crime in the casino neighborhood was 19% higher than would be expected if the casino had not opened and operated. These results held after controlling for confounders including SES, family structure, structural factors, age distribution, home-ownership, sex distribution.
Conclusion: The opening and operation of a casino was associated with higher violent crime in that neighborhood than would be expected without the casino. Changing demographic, social and economic factors did not explain this effect. Evidence-based policy does not support placing casinos in distressed neighborhoods, as the casino may potentially harm instead of help the local community. Further research should examine if alternative types of developments can be used to help distressed neighborhoods.