Table 2.
Meta-regression analyses assessing links between study characteristics and recidivism risk
| β | SE | p value | |
|---|---|---|---|
| Year of publication: ≥1990 vs <1990 | −0·195 | 0·335 | 0·560 |
| Study location: USA vs elsewhere | 0·097 | 0·274 | 0·722 |
| Sample size (continuous) | 0·000 | 0·000 | 0·671 |
| Sex of participants: single sex vs both sexes | −0·404 | 0·371 | 0·276 |
| Mean age (continuous) | −0·016 | 0·018 | 0·372 |
| Age group: adolescents vs adults | −0·161 | 0·284 | 0·570 |
| Intervention type: cognitive behavioural therapy-based vs all other types | −0·217 | 0·270 | 0·422 |
| Comparator type: usual care vs waitlist or other | 0·396 | 0·301 | 0·189 |
| Follow-up time period (continuous) | 0·074 | 0·063 | 0·239 |
| Intervention format: individual vs group or combination | −0·055 | 0·348 | 0·875 |
| Intervention aimed at people in prison with a substance use disorder (dichotomous) | −0·283 | 0·256 | 0·269 |
| Risk of bias: high vs low or unclear | −0·146 | 0·266 | 0·583 |