https://orcid.org/0000-0003-1183-3254
![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
ABSTRACT
Recently, interest has piqued on the effects of neighborhood context on parolee recidivism; however, examinations often neglect to model neighborhood institutions. One type of institution that may be important to neighborhood processes, and moreover, parolee reentry, are houses of worship. Based on the literature, we examine the effect of congregation-oriented (i.e., Evangelical Protestant) and community-oriented (i.e., Catholic and Mainline Protestant) churches on parolee reincarceration. Furthermore, reincarceration was disaggregated to test whether effects were more sensitive to technical parole violations (TPV), or varied by time served on parole. Parolee data were obtained from the Pennsylvania Department of Corrections (PA-DOC) and Pennsylvania Board of Probation and Parole (PBPP) for 3,077 parolees released to 209 neighborhoods across Philadelphia. Multilevel analyses indicated that community-oriented (i.e., Mainline Protestant, Catholic) churches were not associated with parolee outcomes. Conversely, Evangelical Protestant churches were associated with increased odds of reincarceration, and strongest on TPV reincarceration and after the initial six months following release.
Data for this study come from the Pennsylvania Department of Corrections, Pennsylvania Board of Probation and Parole, ReferenceUSA, and the U.S. Census Bureau’s Fact Finder webpage.
Damned from Release: The Effects of Neighborhood Churches on General and Technical Parole Violation Reincarcerations across Time
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1. We refer to houses of worship as “churches,” as the two measures of religious institutions used in this study are comprised of Catholic, Mainline Protestant, and Evangelical Protestant denominations, which are indeed “churches” under the Christian religion.
2. Additional variables were obtained (e.g., marital status, education level, number of prior arrests and sentences to incarceration); however, these variables failed to elicit an effect in the models. This was likely due to these variables being captured within the LSI-R.
3. Due to the complete absence of demographic and/or outcome data for these cases, it was not appropriate to impute these variables.
4. Target offenses are the offense for which the individual was incarcerated, and released onto parole.
5. Parolees with final addresses in a jail, CCC, or CCF are not necessarily revoked from parole. In some cases, this is an intermediate sanction.
6. Results from T-tests indicate sample parolees were more likely to be male and nonwhite, and less likely to have graduated high school. On average, sample parolees scored 1.5 points lower on the LSI-R than non-sample parolees. Sample block groups had more residents living below the poverty line, receiving food stamps/SNAP, without a high school degree, in female-headed households, unemployed, and nonwhite.
7. The average length of time served on parole was approximately 2.5 years (29.85 months), and ranged from 0.00 years (0.00 months) to 6.20 years (74.20 months), with 35.5% serving longer than 3 years. Consistent with the literature, this study is bounded to three years. All parolees not reincarcerated within three years following release are coded as having a successful parole term.
8. 1.8% (n = 59) of parolees died while under supervision. Because they were not reincarcerated these individuals were coded as “0.”
9. A measure for CPV was created; however, the null model indicated only 0.1% of the variation was explained by variables at the neighborhood-level. Because of the small amount of variation at the neighborhood-level, it was deemed inappropriate to further model the CPV reincarcerations.
10. A variable for parolees who were reincarcerated between the second and third year following release (i.e., Time 4) was created; however, null models revealed no variation at the neighborhood-level. Although Time 4 was not modeled separately, these outcomes are included in overall measures of dependent variables.
11. Because the buffer zones overlap, some churches are counted in more than one block group. Additionally, churches that may not be located within one of the sample block groups but are within the 500 foot buffer zone will be included in the count of churches within the buffer.
12. It should be noted that according to Steensland et al. (Citation2000), non-denominational churches and “general” Baptist churches are included in the Evangelical Protestant category.
13. Traditionally, researchers have included a race measure (e.g., % black), which generally loads on to the disadvantage factor. Correlation coefficients showed this measure was negatively correlated with other disadvantage measures, and it did not load in the theoretically expected direction. For this reason, this measure was not included in the factor analysis.
14. The sample scored extremely high on disadvantage measures, with the average block group having approximately 41% of residents living below the poverty line, 29% of households being female-headed households, 44% of household receiving food stamps, and 28% of residents reporting not earning a high school degree. Due to the overall high level of disadvantage within these neighborhoods, “extremely disadvantaged neighborhoods” refers to those neighborhoods score at or above the first deviation above the mean.
15. Non-violent/property offenses include: arson, burglary, driving while under the influence of alcohol or drugs, forgery, fraud, prison breach, theft, receiving stolen property, weapons charges, and other Part II offenses.
16. Drug offenses include: possession of drugs, manufacturing of drugs, and sales of drugs.
17. Person/violent offenses include: aggravated assault, general/other assaults, robbery, kidnapping, general rape, statutory rape, homicide by vehicle, voluntary manslaughter, involuntary manslaughter, murder I, murder II, murder III, other murder, and other categories of sex crimes.
18. In the PA DOC, LSI-R is part of the prison intake process, and takes place within the first two weeks of admission to prison; however, parolees are also reassessed throughout their incarceration (Hardyman, Austin, & Peyton, Citation2004). For parolees who did not have a LSI-R score listed for their current parole (n = 284) the previous score given was used. For those without a previous score (n = 9), multiple imputations, using all other variables, were used to calculate a predicted LSI-R score.