Table 4:
Overall Sample (N = 3,264) |
Male Subgroup (n = 2,951) |
Female Subgroup (n = 313) |
|||||||
---|---|---|---|---|---|---|---|---|---|
Recidivistic Outcomes | AUC (SE) | 95% CI | p | AUC (SE) | 95% CI | p | AUC (SE) | 95% CI | p |
General | |||||||||
Entire follow-up | .64 (.01) | [.62, .66] | <.001 | .65 (.01) | .63, .67 | <.001 | .59 (.04) | .53, .66 | .008 |
1-year follow-up | .64 (.01) | [.62, .67] | <.001 | .64 (.01) | .62, .67 | <.001 | .66 (.04) | .58, .74 | .001 |
3-year follow-up | .65 (.01) | [.63, .67] | <.001 | .66 (.01) | .64, .68 | <.001 | .62 (.04) | .55, .70 | .002 |
5-year follow-up | .65 (.01) | [.62, .68] | <.001 | .66 (.02) | .63, .69 | <.001 | .65 (.04) | .57, .74 | .001 |
Nonviolent | |||||||||
Entire follow-up | .63 (.01) | [.61, .65] | <.001 | .64 (.01) | .62, .66 | <.001 | .59 (.04) | .52,.66 | .010 |
1-year follow-up | .64 (.01) | [.61, .67] | <.001 | .64 (.02) | .61, .67 | <.001 | .65 (.04) | .57, .74 | .001 |
3-year follow-up | .64 (.01) | [.61, .66] | <.001 | .64 (.01) | .62, .66 | <.001 | .62 (.04) | .54, .69 | .004 |
5-year follow-up | .65 (.01) | [.62, .67] | <.001 | .65 (.02) | .62, .68 | <.001 | .64 (.04) | .56, .73 | .002 |
Violent | |||||||||
Entire follow-up | .61 (.02) | [.58, .64] | <.001 | .62 (.02) | .59, .65 | <.001 | |||
1-year follow-up | .61 (.03) | [.56, .66] | <.001 | .62 (.03) | .57, .67 | <.001 | |||
3-year follow-up | .62 (.02) | [.58, .65] | <.001 | .63 (.02) | .59, .66 | <.001 | |||
5-year follow-up | .63 (.02) | [.58, .67] | <.001 | .65 (.02) | .60, .69 | <.001 |
Note. All the p values in bold remained statistically significant even after FDR corrections. AUC values for violent recidivism were not reported for the female subgroup given that only 3 (1.0%) females committed violent offenses during the follow-up period. The AUC values did not differ significantly across the follow-up periods for the various recidivistic outcomes. AUC = area under curve; FDR = False Discovery Rate.