Table 2

Estimated effects of demolitions on firearm violence with and without moderation

Estimated effectsModel 1 Treatment=
prior year’s demolitions
Model 2 Treatment=
cumulative demolitions
(95% CI)
P valueIRR
(95% CI)
P value
All units0.86 (0.83 to 0.90)<0.0010.98 (0.95 to 1.00)0.02
Stratified by cluster
 Cluster A0.85 (0.79 to 0.90)<0.0010.99 (0.97 to 1.01)0.40
 Cluster B1.04 (0.85 to 1.27)0.710.91 (0.85 to 0.98)0.02
 Cluster C0.78 (0.54 to 1.13)0.190.90 (0.84 to 0.97)0.01
 Cluster DNA – no demolitions conducted
 Cluster E
  • Note: models are negative binomial regressions with two-way fixed effects for spatial units and year, and controls for time-varying population count, poverty rate and housing occupancy rate (not displayed). Model coefficients are exponentiated and reported as incident rate ratios. SEs are clustered by unit, year and city quadrant.