Table 2

Risk Factors for Violent Crime Arrest, by Gun Involvement in Crime

PredictorType of Violent Crime
AnyNot Gun-InvolvedGun-Involved
Hazard RatioStandard Error (95% CI)Hazard RatioStandard Error (95% CI)Hazard RatioStandard Error (95% CI)
Legal status
 Short-term hold only1.840.0339 (1.72–1.96)d1.860.0360 (1.74–2)d1.550.1000 (1.27–1.88)d
 Involuntary commitment1.250.0746 (1.08–1.45)b1.290.0777 (1.11–1.5)b0.860.2590 (0.52–1.43)e
Primary psychiatric diagnosis
 Major depressionReferenceReferenceReference
 Bipolar disorder1.260.0334 (1.18–1.34)d1.280.0350 (1.2–1.37)d0.990.1132 (0.79–1.23)e
 Schizophrenia spectrum1.130.0311 (1.07–1.2)d1.110.0330 (1.04–1.19)c1.230.0923 (1.03–1.48)a
Substance use disorder
 Yes1.660.0304 (1.56–1.76)c1.670.0319 (1.56–1.77)d1.480.0968 (1.22–1.78)d
Age0.970.0011 (0.96–0.97)c0.970.0012 (0.96–0.97)d0.970.0034 (0.96–0.98)d
 Male1.850.0269 (1.76–1.95)d1.820.0285 (1.72–1.93)d2.010.0831 (1.71–2.37)d
 Black/African American1.570.0305 (1.48–1.66)d1.420.0324 (1.33–1.51)d3.400.0993 (2.8–4.13)d
 Hispanic/Latino0.930.0334 (0.87–0.99)e0.860.0355 (0.8–0.92)d1.900.1070 (1.54–2.34)d
 Other race/ethnicity0.930.1146 (0.74–1.16)e0.930.1183 (0.74–1.17)e0.880.4539 (0.36–2.13)e
  • Observations were censored following any gun-disqualifying felony conviction or death. Estimates are based on 709 gun-related arrests and 5,747 non–gun-related arrests among 77,048 individuals over 168 months of observations from 1998 to 2011. Legal status was treated as time-varying, while the values of other displayed covariates were fixed. The models were also controlled for community tenure, a time-varying measure of the proportion of days residing in the community each month (range 0–100%). The hazard ratio for age refers to the change associated with each year of increasing age, measured as a continuous variable. The underlying proportional hazard assumption cannot be rejected with the Kolgomorov-type supremum test.

  • a p < .05.

  • b p < .01.

  • c p < .001.

  • d p < .0001.

  • e Not significant.