Study
Duwe and Clark (2011) conducted a quasi-experimental design to examine the effects of prison visitation on recidivism among 16,420 inmates released from Minnesota prisons between 2003 and 2007. Of the 16,420 inmates, 61 percent were visited at least once during their confinement. The treatment group (n = 10,053) included inmates who were visited at least once during their confinement; the comparison group (n = 6,367) included inmates who were not visited during their confinement.
The study sample was majority male (90 percent), and the average age at release from prison was 33.8 years. More than half the study sample (53 percent) were white, and 47 percent were minority (the specific minority races/ethnicities information was not indicated). Of those inmates visited at least once over their entire incarceration period, the average number of visits per inmate was 36 (about 2 visits a month); inmates were on average visited by three individuals. Nearly half (47 percent) of the visited inmates were visited by a friend, nearly one third were visited by their mothers, and more than one fourth were visited by a sibling. There were no statistically significant differences between groups at baseline.
The primary outcome of interest was recidivism, which was measured in two different ways: 1) a reconviction for a felony-level offense and 2) a revocation for a technical violation. Recidivism data was collected through June 30, 2010, thus the follow-up period ranges from 2.5 to 6.5 years (the average was 4.5 years). Data on felony reconvictions was collected from the Minnesota Bureau of Criminal Apprehension (BCA), and technical revocation data was collected from the Minnesota Department of Correction’s Correctional Operation Management System (COMS) database. To estimate the effects of any visitation, a dichotomous variable (any visit) was created, in which visited inmates received a value of 1 and those not visited were given a value of 0.
Data was analyzed using a Cox regression model, which uses both “time” and “status” variables in estimating the impact of the independent variable on recidivism. Subgroup analyses were conducted to look at the effects of visitation according to the relationship between the visitor and inmate.
Study
Duwe and Johnson (2016) conducted a follow-up study to Duwe and Clark (2011), using a retrospective quasi-experimental design, to look at the specific effects of visits from community volunteers (CVs), including clergy and mentors, on recidivism of inmates who are visited during their confinement. Of the 10,053 inmates visited at least once, 2.5 percent (418 inmates) were visited by a CV.
To address potential selection bias issues, propensity score matching was used to match the 418 inmates who received a CV visit (treatment group) with 418 inmates who were not visited by a CV (comparison group). Twenty-seven relevant covariates were included in the propensity score model (such as prior convictions, religious affiliation, and age). Inmates in the treatment group were then matched to comparison group inmates who had the closest propensity score. After propensity score matching, the matched sample was 78.7 percent male and 40.4 percent minority (the specific minority races/ethnicities were not indicated), and the average age was 36.9 years. Both CV–visited and comparison-group inmates received on average about 150 visits during confinement. For inmates visited by CVs, the average number of visits per month was 3.5, while the average number of visits for the comparison group was 3.8. There were no statistically significant differences between groups at baseline.
The primary outcome of interest was recidivism, measured as 1) rearrest, 2) reconviction, 3) reincarceration for a new sentence, and 4) revocation for a technical violation. As in the 2011 study, recidivism data was collected through June 30, 2010, from the Minnesota Department of Correction’s COMS database (for reincarceration and revocation data) and the Minnesota BCA (for arrest and conviction data). The follow-up period ranged from 2.5 to 6.5 years (the average was 4.5 years). To evaluate the relationship between CV visits and recidivism, inmates who received a CV visit were assigned a value of 1 and inmates in the comparison group were assigned a value of 0.
For all four recidivism measures, data was analyzed using a Cox regression model, which uses both “time” and “status” variables in estimating the impact of the independent variable on recidivism.
