TABLE 2—
Binary Logistic Regression of Client Engagement and Recidivism on Universal, Selective, and Indicated Interventions: Connecticut Building Bridges Community Reentry Initiative, 2005–2007
| Poor Engagement |
Recidivism |
|||
| Intervention Variable | OR (95% CI) | B or −2 Log Likelihood (χ2) | OR (95% CI) | B or −2 Log Likelihood (χ2) |
| Universal (n = 110) | 108.06* (12.24) | 96.74 (7.57) | ||
| Employment | 0.31* (0.11, 0.89) | −1.16 | 1.74 (0.62, 4.94) | 0.55 |
| Private residence (independent or with family) | 3.12* (1.20, 8.11) | 1.14 | 0.60 (0.21, 1.76) | -0.51 |
| Community supports | 1.15 (0.43, 3.08) | 0.14 | 0.26* (0.08, 0.88) | −1.35 |
| Access to health care | 0.47 (0.18, 1.26) | -0.75 | 1.03 (0.36, 2.95) | 0.03 |
| Selective (n = 94) | 97.76 (4.54) | 70.50 (1.30) | ||
| Financial assistance | 0.89 (0.32, 2.51) | -0.12 | 1.44 (0.42, 5.00) | 0.36 |
| Child support arrangements | 0.00 | −20.13 | 2.51 (0.44, 14.23) | 0.92 |
| Indicated (n = 100) | 103.50** (8.97) | 97.19 (0.05) | ||
| Any substance abuse treatment | 0.22** (0.07, 64) | −1.52 | 1.13 (0.41, 3.06) | 0.12 |
| Nontransitional housing treatment | 0.07* (0.01, 0.54) | −2.68 | 1.28 (0.41, 3.98) | 0.25 |
Note. CI = confidence interval; OR = odds ratio. Universal preventive interventions are directed at the entire population of interest, selective interventions are for a subgroup of the population that is at increased risk of difficulties, and indicated interventions are designed for those at greatest risk.
*P < .05; **P < .01.