Correction to: The effects of the Summer All Out Foot Patrol Initiative in New York City: a difference-in-differences approach

Abstract

The original version of the article unfortunately contained mistakes. In the Table 2 footnote, the number of precinct-month observations were incorrect, and the terms for the treatment effect for individual exposure months in parentheses in the second to last sentence were doubled and formatted incorrectly. The Table 2 footnote should have appeared as shown below.Notes: Model 3 reports results for DiD models comparing shootings in months preand post-deployment of uniformed foot patrols. The 2014 panel includes 70 precincts observed over 9 months for a total of 630 (n*t) precinct-month observations. Because the second iteration commenced 1 month earlier, the 2015 panel includes 70 precincts observed over 8 months for a total of 560 (n*t) precinct-month observations. Column 1 is the base specification ignoring the cluster structure. Column 2 incorporates the exposure variable. Column 3 adjusts for pre-trends. Individual post-exposure effects are reported beneath the main interaction. For example, the "after" periods in 2014 represent separate estimates of the treatment effect for individual exposure months (e.g., ̂J ul ; ̂A ug ; ̂S ep ). Standard errors are clustered on precinct and reported in parentheses *p < .05; **p < .01; ***p < .001 (two-tailed)The original article has been corrected.