Risk Factors for Violent Crime Arrest, by Gun Involvement in Crime
| Predictor | Type of Violent Crime | |||||
|---|---|---|---|---|---|---|
| Any | Not Gun-Involved | Gun-Involved | ||||
| Hazard Ratio | Standard Error (95% CI) | Hazard Ratio | Standard Error (95% CI) | Hazard Ratio | Standard Error (95% CI) | |
| Legal status | ||||||
| None | Reference | Reference | Reference | |||
| Short-term hold only | 1.84 | 0.0339 (1.72–1.96)d | 1.86 | 0.0360 (1.74–2)d | 1.55 | 0.1000 (1.27–1.88)d |
| Involuntary commitment | 1.25 | 0.0746 (1.08–1.45)b | 1.29 | 0.0777 (1.11–1.5)b | 0.86 | 0.2590 (0.52–1.43)e |
| Primary psychiatric diagnosis | ||||||
| Major depression | Reference | Reference | Reference | |||
| Bipolar disorder | 1.26 | 0.0334 (1.18–1.34)d | 1.28 | 0.0350 (1.2–1.37)d | 0.99 | 0.1132 (0.79–1.23)e |
| Schizophrenia spectrum | 1.13 | 0.0311 (1.07–1.2)d | 1.11 | 0.0330 (1.04–1.19)c | 1.23 | 0.0923 (1.03–1.48)a |
| Substance use disorder | ||||||
| No | Reference | Reference | Reference | |||
| Yes | 1.66 | 0.0304 (1.56–1.76)c | 1.67 | 0.0319 (1.56–1.77)d | 1.48 | 0.0968 (1.22–1.78)d |
| Age | 0.97 | 0.0011 (0.96–0.97)c | 0.97 | 0.0012 (0.96–0.97)d | 0.97 | 0.0034 (0.96–0.98)d |
| Sex | ||||||
| Female | Reference | Reference | Reference | |||
| Male | 1.85 | 0.0269 (1.76–1.95)d | 1.82 | 0.0285 (1.72–1.93)d | 2.01 | 0.0831 (1.71–2.37)d |
| Race/ethnicity | ||||||
| White | Reference | Reference | Reference | |||
| Black/African American | 1.57 | 0.0305 (1.48–1.66)d | 1.42 | 0.0324 (1.33–1.51)d | 3.40 | 0.0993 (2.8–4.13)d |
| Hispanic/Latino | 0.93 | 0.0334 (0.87–0.99)e | 0.86 | 0.0355 (0.8–0.92)d | 1.90 | 0.1070 (1.54–2.34)d |
| Other race/ethnicity | 0.93 | 0.1146 (0.74–1.16)e | 0.93 | 0.1183 (0.74–1.17)e | 0.88 | 0.4539 (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.