FOR SURAHSIGN ONLY Due wed.

Journal of Criminal Justice 43 (2015) 242–250

Contents lists available at ScienceDirect

Journal of Criminal Justice

Responding to probation and parole violations: Are jail sanctions more effective than community-based graduated sanctions?

Eric J. Wodahl a,⁎, John H. Boman IV a, Brett E. Garland b

a University of Wyoming b Missouri State University

⁎ Corresponding author at: University of Wyoming, D Dept. 3197, 1000 E. University Ave. Laramie, WY, 82071.

E-mail address: [email protected] (E.J. Wodahl).

http://dx.doi.org/10.1016/j.jcrimjus.2015.04.010 0047-2352/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Available online 25 May 2015

Purpose: In response to escalating revocation rates in community supervision, many jurisdictions have adopted graduated sanction policies. Research on graduated sanctions has shown promising results. However, most stud- ies focus exclusively on jail sanctions and have largely ignored the possibility that community-based graduated

sanctions such as written assignments, increased treatment participation, or community service hours may be as effective, or more effective, than jail sanctions. Extending this research, the current study examines whether community-based sanctions are as effective in increasing offender compliance as spending time in jail. Methods: Using data from over 800 violations committed by a random sample of probationers and parolees on in- tensive supervision probation, multilevel models are estimated that examine whether jail sanctions are more effec- tive than community sanctions in 1) extending time to the offender’s next violation event, 2) reducing the number of future violations, and 3) successfully completing the probation program. Results: Results consistently indicate that jail sanctions do not outperform community-based sanctions. Conclusion: Due to the financial, social, and potentially criminogenic effects of jail, the lack of significant differences between jail sanctions and community-based sanctions calls into question the use of jail as a means of punishing persons on community supervision.

© 2015 Elsevier Ltd. All rights reserved.

Introduction

A primary appeal of community-based corrections since their incep- tion has been the belief that these forms of punishment reduce our reli- ance on incarceration, which is not only costly but also potentially criminogenic (Wodahl & Garland, 2009). Recently, however, the capac- ity of community supervision to ease prison populations has been questioned, due in large part to burgeoning revocation rates. Since the 1980s, revocation rates among offenders on probation and parole have increased substantially (Wodahl, Ogle & Heck, 2011b). Community corrections failures have contributed noticeably to prison growth and crowding, and strained state and federal economic resources (Adams, 2013; Criminal Justice Policy Council, 2002; Grattet, Petersilia, Lin, & Beckman, 2009; Travis & Lawrence, 2002). In addition, the incarceration accompanying revocation has been shown to have a detrimental impact on offenders, their families, and the communities in which they reside (Clear & Rose, 2003; Petersilia, 2003; Rollo, 1988; Sabol & Lynch, 2003).

The use of graduated sanctions has gained traction in recent years as a mechanism to mitigate the effects of rising revocation rates. Graduat- ed sanctions, sometimes referred to as administrative responses, refer

epartment of Criminal Justice,

to the imposition of swift, certain, and proportionate punishments for offenders who violate the conditions of their community supervision (Taxman, Soule, & Gelb, 1999). While graduated sanctions are official responses to noncompliance, they do not involve the formal revocation of community supervision, which often results in long-term imprison- ment. Normally imposed by the supervising officer or judge, these sanctions allow the offender to avoid revocation and remain under community supervision. For example, a probation officer may impose 10 hours of community service on a probationer as a sanction for miss- ing a probation appointment, or a parolee may receive a 4 day jail sanc- tion as a response to a positive drug test.

A growing body of research suggests that graduated sanctions can be an effective tool to increase offender compliance with community supervision conditions and reduce revocation rates1 (Grommon, Cox, Davidson, & Bynum, 2013; Hawken & Kleiman, 2009; Kilmer, Nicosia, Heaton, & Midgette, 2013; Steiner, Makarios, Travis, & Meade, 2012; Wodahl, Garland, Culhane & McCarty, 2011a). While there is consider- able evidence to support the efficacy of graduated sanctions in improv- ing community supervision outcomes, little is known about how these sanctions can be implemented to achieve the best results. One particu- larly important gap in the research is the salience of sanction type. The vast majority of research on graduated sanctions has focused exclu- sively on one type of sanction – jail sanctions (Grommon et al., 2013; Hawken & Kleiman, 2009; Kilmer et al., 2013; Steiner et al., 2012).

243E.J. Wodahl et al. / Journal of Criminal Justice 43 (2015) 242–250

Aside from sending offenders to jail, there are a number of alternative, community-based graduated sanctions that jurisdictions can impose on recalcitrant supervisees, such as electronic monitoring, written assign- ments, or increased treatment participation. To date, however, we know very little about the capacity of these noncustodial sanctions to improve community supervision outcomes. The current study seeks to address this limitation by examining whether community-based sanctions are as effective in increasing offender compliance as spending time in jail.

Graduated Sanctions in Community Corrections

Graduated sanctions first gained wide-spread attention for their use in drug court programs (NADCP, 1997). Likely motivated by the suc- cess of drug court programs, a number of jurisdictions adopted the use of graduated sanctions in more traditional probation and parole case- loads. Perhaps the most well-known example is the Hawaii HOPE program, which focuses on probationers who are at a high risk to experi- ence a probation violation. Hawken and Kleiman’s (2009) well-known evaluation of HOPE program found that the imposition of short jail sentences for offender noncompliance was associated with a number of positive outcomes, including reduced positive drug tests, fewer missed appointments, and lower revocation rates. Similarly, Grommon et al. (2013) studied graduated sanctions in the supervision of parolees in a Midwestern state, finding that parolees who were subject to frequent random drug testing and swift and certain jail sanctions had lower rates of relapse and recidivism than parolees who were not subject to these in- terventions. Other recent examples of graduate sanction implementation in jurisdictions include the Probation Operation Management (POM) program implemented in 2004 in the state of the Georgia, the Wyoming Department of Corrections’ Intensive Supervision Program, and South Dakota’s 24/7 Sobriety Program (APPA, 2013).

Research on the efficacy of community-based graduated sanctions is important for several reasons. First, not all agencies have the capacity to impose jail sanctions on noncompliant offenders. While many jurisdic- tions have changed policies and passed legislation authorizing the use of jail sanctions, others have relied exclusively on noncustodial responses (APPA, 2013). Furthermore, many agencies are likely contemplating the implementation of a graduated sanctioning program but may be discouraged from acting because they lack the ability to include jail in their repertoire of sanctions.

It must also be recognized that utilizing jail sanctions can be resource intensive. The cost of incarcerating a probationer or parolee in jail even for a short period of time can be substantial, with estimated daily costs per jail inmate exceeding $100 in many locales (Piquero, 2010; Santora, 2013). Further exacerbating the cost issue, jail sanctions can also be time and labor intensive for correctional and court personnel. In many programs, jail sanctions must be imposed by the judge. This requires the offender to be returned to court, which subsequently creates addition- al work and time commitments for those involved in the process (Kleiman, 2011). Community-based graduated sanctions by contrast can often be imposed directly by the supervision agent without the need to return the offender to court for a hearing. The costs in terms of both time and money can certainly be justified given the benefits of jail sanc- tions on offender outcomes; however, if community-based responses are equally effective, the use of jail sanctions becomes more difficult to justify.

Jail versus Community-Based Graduated Sanctions

Guided by theory and prior research, two possible outcomes regarding the effectiveness of jail versus community-based graduated sanctions are proposed. The first is guided by deterrence theory and suggests that jail will outperform community-based graduated sanctions because of the punitive nature of spending time in jail. The second is that community-based graduated sanctions will outperform jail sanctions due to the deleterious effects of incarceration and the jail environment.

The following paragraphs explore these competing proposals in more detail.

Graduated Sanctions and Deterrence

The use of graduated sanctions to improve offender compliance with supervision conditions is guided primarily by deterrence theory (Braga & Weisburd, 2012; Duriez, Cullen, & Manchak, 2014). Punishments or sanctions as applied in deterrence theory are designed to reduce the like- lihood of an undesired behavior by increasing the perceived costs or negative consequences associated with the action (Pogarsky, 2009). This deterrent effect is theorized to operate on two levels. On one hand, general deterrence asserts that the knowledge of punishments is sufficient to dissuade the act, meaning that individuals do not need to experience the effects of punishment first-hand in order to be discouraged from offending (Gibbs, 1975). On the other hand, specific deterrence operates when individuals who have previously been caught and sanctioned for engaging in deviant behavior cease or curtail their involvement in these activities because they are unwilling to risk future punishment (Gibbs, 1975). While it is likely that the deterrent effect of graduated sanctions operates at both the general and specific levels, our inquiry is limited to the latter – the specific deterrent effect of graduated sanctions. Deterrence theory further posits that the capacity of sanctions to reduce criminal behavior is dependent upon three interrelated factors – the severity, cer- tainty, and celerity of the punishment (Gibbs, 1975; Pogarsky, 2009). Given our focus on the effects of jail versus community-based sanctions, it is the severity aspect with which we are most concerned.

On its face, deterrence theory has little to say about the efficacy of specific types of punishments (i.e. jail versus community-based sanc- tions); it does, however, assert an inverse relationship between punish- ment severity and the likelihood of future transgressions (Paternoster, 2010). This suggests that the effectiveness of community-based gradu- ated sanctions compared to jail will be largely determined by the degree to which offenders differentiate the two in terms of their austerity. As noted by Nagin, Cullen, and Jonson (2009), “if a custodial sanction is perceived to be more costly than a noncustodial sanction, the imprison- ment sanction will exert a greater deterrent effect” (p. 124).

Research demonstrates that offenders view spending time in jail as a particularly punitive sanction, especially when compared to community- based alternatives (May, Applegate, Ruddell, & Wood, 2014; May, Wood, Mooney, & Minor, 2005; Wood & Grasmick, 1999). For example, May et al. (2005) found that Kentucky probationers viewed county jail as more punitive than a variety of community-based punishments including elec- tronic monitoring, day reporting, and community service hours. May et al. (2014) offer several potential explanations for why offenders view jail in such punitive terms. They note that jails are often dangerous environ- ments that produce high rates of violent and sexual victimization. Jails also house a broad range of offenders, many of which are mentally ill and/or under the influence of drugs and alcohol, which exacerbates the disruptive environment (May et al., 2014). Additionally, jails typically provide few education, work, or treatment opportunities for inmates, meaning that much of the inmates’ time is idle and passes slowly (May et al., 2014). These findings suggest that graduated sanctions which involve jail time will be more effective deterrents than community- based responses due to the punitive nature of spending time in jail.

The Deleterious Effects of Incarceration and the Jail Environment

A competing hypothesis is that community-based graduated sanc- tions will outperform jail sanctions due to the deleterious effects of incarceration and the jail environment. Removing a probationer or parol- ee from the community, even for a short duration, may produce a number of negative consequences that heighten the risk of supervision failure. Employment, for example, is strongly related to both reoffending and community supervision outcomes (Morgan, 1994; Wilson, Gallagher, & Mackenzie, 2000). Jail terms can adversely affect one’s employability by

Table 1 List of ISP Sanctions

Sanctions

Verbal reprimand Written assignment Community service hours Modify curfew hours Community service hours Restrict visitation Program extension or regression Electronic monitoring Inpatient or outpatient treatment County jail time

244 E.J. Wodahl et al. / Journal of Criminal Justice 43 (2015) 242–250

creating employment gaps on résumés and discouraging offenders from seeking employment for fear a job will be lost once a jail sanction is received (Western, Kling, & Weiman, 2001). Jail time also means tempo- rary losses in wages or even loss of a job if an offender is currently employed (Grogger, 1995).

Family support and prosocial peer associations, which are also associ- ated with improved supervision outcomes (Latessa, Listwan, & Koetzle, 2015), can also be jeopardized through incarceration. Family disruption, which is particularly damaging, is likely to occur from relationship strain and financial loss from the absence of a breadwinner (Clear, Rose, & Ryder, 2001; Freudenberg, 2001). Furthermore, the stigma attached to the offender as a result of the incarceration is likely to erode prosocial peer associations (Nagin et al., 2009).

In addition to the negative effects of removing the individual from the community, other adverse consequences are likely to result from sanc- tioning an offender to jail. As noted above, jails can be crowded, violent, exploitive, and stressful places (Cornelius, 2008; May et al., 2014). Research focusing on prison inmates reveals that reactions to these envi- ronments can include hypervigilance, psychological distancing, dimin- ished feelings of self-worth, post-traumatic stress, depression, and even suicide (Haney, 2001; Wooldredge, 1999). It is likely that jail incarcera- tion exerts similar reactions, which might inhibit an individual’s capacity to comply with conditions of supervision upon their release. Incarceration also provides exposure to criminal motives, value systems, and tech- niques which have the potential to encourage behavior contradictory to successful community supervision outcomes such as drug use and future offending (Akers, 1997; Cullen & Gilbert, 1982; Nagin et al., 2009). These issues are compounded by the fact that substance abuse, mental health, and other treatment needs are often insufficiently met in jail due to limited resources, thereby impeding the delivery of quality rehabilitative services (Solomon, Osborne, LoBuglio, Mellow, & Mukamal, 2008). In sum, there is good reason to believe that the deleterious effects associated with removing the offender from the community and the jail environ- ment will increase the risk of supervision failure. Community-based grad- uated sanctions, by contrast, avoid these negative consequences by allowing the offender to remain in the community.

Current Study

The current study seeks to address an important gap in the literature surrounding the use of graduated sanctions by exploring the effect of sanction type on offender compliance with release conditions. More spe- cifically, the study examines whether the imposition of community-based sanctions such as community service hours, electronic monitoring, or increased treatment attendance are more or less effective in increasing offender compliance with supervision conditions than spending time in jail. The following three research questions guided this analysis:

1. Does the imposition of a jail sanction have a greater or lesser influence on time to next violation, compared to community-based sanctions?

2. Does the imposition of a jail sanction as opposed to a community- based sanction influence the number of subsequent violations?

3. Does the imposition of a jail sanction as opposed to a community- based sanction influence the likelihood of successfully completing supervision?

Methods

Data

The data for this study were collected from case files and agency records on adult, felony offenders2 supervised in the Wyoming Depart- ment of Corrections’ Intensive Supervision Program (ISP). The Wyoming ISP program is a one year program reserved for high risk and/or high needs felony offenders. While on the program, participants are subjected to intensive levels of supervision, including frequent drug tests and home

visits. In addition to close monitoring, the program emphasizes treatment to address criminogenic needs to include substance abuse counseling, cognitive behavioral treatment, and sex offender programming.

A key component of the Wyoming ISP program is the use of graduated sanctions. As part of their ISP supervision, offenders are subject to a vari- ety of graduated sanctions in response to violations of the conditions of their supervision. The application of graduated sanctions is determined by both department policy and officer discretion. Department policy requires officers to promptly impose sanctions on all violations; however, the type and intensity of the sanction is left to the discretion of the super- vising officer, often in consultation with his or her supervisor.

A list of available graduated sanctions is provided in Table 1. In this ISP program, probation officers have the ability to impose both community- based and jail-based sanctions. In addition to imposing community- based sanctions such as community service hours and electronic monitor- ing, department policy and state law authorizes the imposition of jail sanctions for up to 30 days for offenders who violate supervision conditions.

The study sample consists of a random selection of 283 offenders who participated in ISP between 2000 and 2003. These 283 individuals, who committed a total of 861 violations, represent approximately 20% of the eligible probation and parole ISP cases in the state. Disproportionate strat- ified sampling procedures were utilized to oversample on two dimen- sions, gender and supervision type. This was done to ensure that sufficient numbers of females and parolees were included in the study sample so that meaningful comparisons could be made across these attri- butes. Thus, while females made up 20% of the ISP population, they accounted for 31.5% of the study sample. Parolees made up 17.5% of the ISP population, but accounted for 32% of the study sample. Aside from gender and release type, the study sample was representative of the ISP population on key attributes, including age, educational attainment, race/ethnicity, and criminal history.3

A variety of descriptive statistics which reflect the characteristics of the offenders in the sample are presented in Table 2. Approximately 69% of the offenders were male and most (82%) were white. The average age of the offenders was 31 and 22% were married. The vast majority (79%) have received their high school diploma or GED. Only 35% of the sample had a prior felony conviction, slightly less than half (47%) have been revoked from community supervision in the past, and about two thirds of individuals were on probation (69%) instead of parole. Property offenders were the most common type of program participant, compris- ing 37% of the sample, followed by drug offenders (30%), violent offenders (15%), and sex offenders (14%). While not reported in Table 2, statistics reveal that over 80% of the sample was sanctioned for committing at least one violation while on ISP. The average number of violations com- mitted per offender was 3.1 and approximately 68% of the study sample successfully completed ISP supervision.

Before proceeding to the discussion of the specific variables, it is important to highlight two relevant factors about these data and subse- quent analyses. First, while the study sample consists of 283 offenders, the unit of analysis is not individual offenders. Rather, the unit of analysis is each violation and sanction event experienced by these 283 offenders.

Table 2 Descriptive characteristics of offenders in survey sample (n = 283 offenders)

n / 283 total %

Age Mean 30 Median 28 Range 17 – 78

Gender Male 194 68.6% Female 89 31.5%

Race White 233 82.3% Non-white 50 17.7%

Marital status Married 64 22.6% Not Married 219 77.4%

Property offender Yes 106 37.5% No 177 62.5%

Prior felony conviction Yes 102 36.0% No 181 64.0%

Prior revocation Yes 139 49.1% No 144 50.9%

Arrested as juvenile Yes 138 48.8% No 145 51.2%

High school diploma / GED Yes 224 79.2% No 59 20.9%

Supervision type Probation 192 68.7% Parole 91 32.2%

Type of ISP discharge Successful 180 63.6% Revoked 103 36.4%

Table 3 Means, standard deviations, observed minimums, and observed maximums of variables used in analyses (N = 861 total violations)

Mean SD Obs. Min. Obs. Max.

Dependent Variables Days until next violation 45.076 48.389 0 397 # of subsequent violations 2.240 2.199 0 10 Successful ISP program completion .677 .469 0 1

Independent Variables based on Sanction Events Jail sanction .160 .367 0 1 Total # of violations while on ISP 3.128 2.713 0 11 ISP completion % at time of violation 42.494 27.194 0 100

Independent Variables based on Personal Characteristics Property offender .372 .484 0 1 Prior felony conviction .346 .476 0 1 Prior revocation .470 .499 0 1 Arrested as juvenile .492 .500 0 1 Parolee .327 .469 0 1 High school diploma / GED .789 .408 0 1 Male .688 .463 0 1 Age 30.936 10.669 17 78 Non-white .180 .385 0 1 Unmarried .778 .416 0 1

245E.J. Wodahl et al. / Journal of Criminal Justice 43 (2015) 242–250

As such, the dataset is structured in a long format where each line of data represents one violation and sanction event (N = 861 total violations).

Second, it is important to highlight that previous research utilizing these data has revealed that the use of graduated sanctions is effective in improving offender outcomes (Wodahl et al., 2011b). In this pre- vious analysis, graduated sanctions were not differentiated between community-based and jail sanctions. Thus, it is important to keep in mind while the current study examines whether the imposition of com- munity-based sanctions are more or less effective than spending time in jail, the efficacy of sanctions in this data has already been established.

Dependent Variables

Days until next violation Our first set of analyses investigates factors that influence the number

of days until the next violation (defined as a new or technical violation). At the most rudimentary level, this measure represents the number of days in between each violation. Calculation for this measure began after the person’s first violation, meaning that the number of days between the start of ISP and the first violation is excluded. Necessarily, this implies that the offender must have had two or more violations while on ISP to qualify for this model.

On ISP, offenders run the risk of a jail sanction following a violation. Those who receive such a sanction, however, cannot commit a new ISP violation during their time in jail. Because not correcting for this would create one or more days where it would be impossible for offenders to vi- olate, this measure is compensated for the number of days spent in jail in

between each violation. Mathematically, the formula for calculation of this measure is (|days between violations| - |days spent in jail between violations|). The descriptive statistics of event-level variables used in anal- yses are presented in Table 3.

Number of subsequent violations Following the commission of each violation, a dependent variable was

created that captured the number of future violations (new and technical violations combined) a respondent committed. To qualify, a person must have had at least one violation while on ISP. Since the highest number of violations in these data was eleven, this measure has a range from zero to ten.

Successful ISP program completion The final dependent variable is a binary measure of whether the of-

fender was revoked (coded ‘0’) or successfully completed the ISP pro- gram (coded ‘1’). Overall, 67.7% of those who started the ISP program completed it successfully.

Independent Variables

Jail sanction A central theme to this study involves the comparison of outcomes

among those who receive jail sanctions versus those who receive commu- nity sanctions. To make these comparisons, a binary variable was created that distinguished whether the respondent received a community sanction immediately following a violation (coded ‘0’) or a jail sanction immediately following a violation (coded ‘1’). This first variation of the ‘jail’ variable is used in our first two sets of analyses, which investigate whether a jail sanction as opposed to a community-based sanction influ- ences the number of days until the next violation and the number of sub- sequent violations, respectively. From the first ‘jail’ variable, we created a second binary variable where those who received only community-based sanctions (coded ‘0’) were compared to those who received at least one jail sentence (coded ‘1’). The latter measure is the one used in our final analysis, which investigates if receiving a jail sanction impacts successful ISP program completion.4

As a point of clarification, the descriptive statistics of this variable indicate that 16% of violations resulted in a jail sentence. A total of 36% of offenders in the sample went to jail while they were enrolled in the ISP program.

246 E.J. Wodahl et al. / Journal of Criminal Justice 43 (2015) 242–250

Control Variables

Total number of violations while on ISP Offenders who violate frequently run a higher risk of revocation

(e.g., Jones, 1995). As such, we control for the total number of violations that each person had while on ISP.

ISP completion percentage at time of violation Offenders who have just begun ISP will likely have more subsequent

violations than those who are nearly finished. Thus, we control for the percentage completion of the ISP program in models which utilize the number of subsequent violations as a dependent variable. This variable was constructed by taking the total number of days each person was enrolled in ISP, dividing by the date which each violation occurred, and multiplying by 100.

Property offender Property offenders tend to have more difficulties completing com-

munity supervision than non-property offenders (Gray, Fields, & Maxwell, 2001; Jones, 1995; Morgan, 1994). Accordingly, we control for whether the offender was classified by the agency as being a proper- ty offender (coded ‘1’) or a non-property offender (coded ‘0’).

Prior felony conviction Prior felony convictions are also a risk factor which may influence

supervision outcomes (Jones, 1995; Morgan, 1994). We control for whether the respondent had been previously convicted of a felony (coded ‘1’) or not convicted of a felony (coded ‘0’).

Prior revocation Being previously revoked while on community supervision is a risk

factor for future revocation (Jones, 1995; Morgan, 1994). Thus, we con- trol whether offenders had ever been previously revoked (coded ‘1’) or not (coded ‘0’).

Arrested as a juvenile Arrests at early ages are risk factors for later offending (Gray et al.,

2001; Jones, 1995; Morgan, 1994). Accordingly, we control for whether the offender was arrested as a juvenile (coded ‘1’) or not (coded ‘0’).

Supervision type Statistics reveal that success rates of probationers and parolees differ

substantially, with parolees experiencing higher rates of supervision failure (Herberman & Bonczar, 2014; Wodahl, Garland et al., 2011a, Wodahl et al., 2011b). As such, we control for whether the ISP partici- pant was on probation (coded ‘0’) or parole (coded ‘1’).5

High school diploma / GED Since research suggests that education level may be related to success

in ISP programs (Gray et al., 2001; Jones, 1995; Morgan, 1994), we control for respondent education level. Specifically, we measure whether the respondent had received a high school diploma or received a GED (coded ‘1’) or had not received either (coded ‘0’).

Demographics To provide additional protection against spuriousness, we include four

variables that measure gender, age, race, and marital status. To measure gender, females (coded ‘0’) are compared to males (coded ‘1’). The age of the respondent was captured at the time that the offender began ISP. Finally, both race and marital status are captured through binary variables where whites and married offenders (both coded ‘0’) are com- pared to non-white and unmarried offenders (both coded ‘1’), respectively.6

Analytical Strategy

To evaluate our three research questions, we employ the use of two- level mixed and mixed logistic regression models. The multilevel approach is preferential in this case over a one-level regression because the event-level data are nested, and failing to account for this could result in a violation of the assumption of local independence. The first level of these models represents the target equation as would be found in a single level regression. The second level in all models is to account for nesting within the data by grouping the level one equation around the county in which the probationer’s ISP office was located. Because we are using a long file data structure instead of person-level data, the standard errors of all models are clustered around persons.

Due to somewhat complex models and natural nesting within the data, we employ the use of a non-parametric bootstrap replication meth- od. Designed to mimic the data’s collection procedure, the bootstrap technique iteratively creates many subsamples within the data and esti- mates the regression model for each subsample. This in turn aids in the accuracy of the estimation of the standard error. Because researchers have concluded that very large numbers of bootstraps are necessary to gain the most accurate estimates of standard errors (Rodgers, 1999), we use 1,000 random bootstrap draws from 100,000 random subsamples of the dataset.

Being that the data was recorded directly from completed ISP case files, there were extremely minor amounts of missing data. Variables had no missing data, with the exception of missing violation dates for two probationers. These cases were dropped, providing a final sample size of N = 859 violation and sanction events. Models were estimated in Stata v. 11.2 and re-estimated in Mplus v. 6.1 for confirmation of findings.

Results

Our first research question examines whether the imposition of a jail sanction has a greater or lesser influence on time to next violation, com- pared to a community-based sanction. Findings relevant to this question are presented in a two-level mixed model presented in Table 4. Recall that the dependent variable, days until next violation, is compensated for days spent in jail in between offenses since persons on ISP cannot re-violate during time they spend in jail.7

Although the multilevel model reaches significance (Wald χ2 = 26.22, p ≤ .01), our primary variable of interest – jail sanctions – is not signifi- cantly related to the number of days until the next violation. This suggests that receiving a jail sanction as opposed to a community-based sanction for a violation on ISP does not significantly increase or decrease the num- ber of days until a person’s next violation. Instead, the only variable in the level-one fixed effects equation that reaches statistical significance is the measure tapping the total number of violations a person commits while on ISP (b = -3.314, SE = .884, z = -3.75, p ≤ .001). The direction of this measure indicates that more violations are related to less days until a person’s next violation. Additionally, a significant sigma-squared statistic at level-two suggests that there is significant residual variance around the county grouping variable (b = 47.227, SE = 2.947, z = 16.03, p ≤ .001).8

Our second research question inquires if the imposition of a jail sanc- tion as opposed to a community-based sanction influences the number of subsequent violations.9 Table 5 reports results from two multilevel mixed models which regress the number of subsequent violations onto predictors. In the first model, the jail sanctions measure again fails to reach levels of statistical significance, suggesting that jail sanc- tions as opposed to community-based sanctions do not significantly increase or decrease the number of subsequent violations an offender commits.

Several control measures in the fixed effects equation of model one do reach significance, however. The percent ISP completion measure reaches significance (b = -.046, SE = .002, z = -18.72, p ≤ .001), and the direction of this relationship demonstrates that offenders who are

Table 4 Two-level mixed models investigating if jail sanctions as opposed to community-based sanctions influence the number of days until next violation (bootstrapped standard errors; n = 543)

b SE

Fixed Effects (Level 1) Independent Variables based on Sanction Events Jail sanction 12.474 7.131 Total # of violations while on ISP −3.314 .884***

Independent Variables based on Personal Characteristics Property offender −5.801 4.632 Prior felony conviction 3.519 5.183 Prior revocation 5.015 4.979 Arrested as juvenile 1.089 5.684 Parolee −5.004 6.372 High school diploma/GED 2.093 4.763 Male .304 6.006 Age .384 .343 Non-white 1.199 5.589 Unmarried −4.841 5.560 Intercept 54.218 13.979***

Random Effects (Level 2) τ 0 9.311 20.374 σ 2 47.227 2.947***

Model Statistic Wald χ2 26.22**

* p ≤ .05 ** p ≤ .01 *** p ≤ .001

247E.J. Wodahl et al. / Journal of Criminal Justice 43 (2015) 242–250

relatively new to the ISP program commit significantly more subse- quent violations than those who have been in the program longer. Additionally, being previously revoked while on probation (b = .530, SE = .140, z = 3.80, p ≤ .001), being a parolee (b = .407, SE = .186, z = 2.19, p ≤ .05), not having a high school diploma or GED equivalent (b = -.825, SE = .162, z = -5.10, p ≤ .001), and being non-white (b = .394, SE = .174, z = 2.26, p ≤ .05) are all related to having more subse- quent violations while on ISP. Finally, the residual variance at level two suggests that differences in the number of subsequent violations an

Table 5 Two-level mixed models investigating if jail sanctions as opposed to community-based sanctions influence the number of subsequent violations (bootstrapped standard errors; n = 708)

Model 1 Model 2

B SE b SE

Fixed Effects (Level 1) Independent Variables based on Sanction Events Jail sanction -.130 .184 -.116 .177 ISP completion % at violation -.046 .002*** -.046 .002*** Jail * ISP completion % at violation -- -- .010 .007

Independent Variables based on Personal Characteristics Property offender .230 .136 .229 .137 Prior felony conviction .019 .159 .030 .160 Prior revocation .530 .140*** .525 .137*** Arrested as juvenile .220 .154 .235 .154 Parolee .407 .186* .425 .185* High school diploma/GED -.825 .162*** -.803 .163*** Male -.060 .171 -.073 .174 Age -.021 .008* -.021 .009* Non-white .394 .174* .387 .173* Unmarried -.180 .167 -.184 .165 Intercept 4.746 .381*** 4.743 .393***

Random Effects (Level 2) τ 0 .156 .703 .152 .677 σ 2 1.749 .056*** 1.747 .055***

Model Statistic Wald χ2 413.00*** 396.88***

* p ≤ .05 ** p ≤ .01 *** p ≤ .001

offender has is partially attributable to the county in which he/she is en- rolled in ISP (b = 1.749, SE = .056, z = 31.07, p ≤ .001).

While it is important to know that the main effect of the jail sanction variable is not significant, this does not answer an important related question: Does the timing of a jail sanction influence the number of subsequent violations for an offender? An interaction term between the binary jail sanction variable and the percent ISP completion variable pre- sented in the second model in Table 5 investigates this issue.10 The inter- action term is not significant, indicating that there are no differences in the number of subsequent violations an offender commits based on the timing of a jail sanction. Overall, the results from model two closely mimic the results from the first model, although the model statistic in model two (Wald χ2 = 396.88, p ≤ .001) decreases from the model statis- tic in the first model (Wald χ2 = 413.00, p ≤ .001) due to a change in the degrees of freedom.

Our final research question inquires if receiving a jail sanction as opposed to community-based sanctions impacts successful completion of the ISP program. This issue is tested with a two-level mixed logit regression presented in Table 6. As in the previous models, the jail sanc- tion variable does not approach levels of statistical significance, suggest- ing that receiving a jail sanction as opposed to a community sanction does not influence whether the probationer or parolee successfully completes the ISP program. Only one control variable, the dichotomous property offender measure, reaches significance (b = -.774, SE = .191, z = -4.05, p ≤ .001). The direction of this relationship shows that non- property offenders are more successful in completing ISP than property offenders. Finally, a significant level of variance at level two (b = .494, SE = .110, z = 4.51, p ≤ .001) suggests that there are significant differ- ences in successful ISP completion rates between counties.

Supplementary Analyses

In considering the importance of reported results, three points of clar- ification are necessary. First, recall that the models in Tables 4 through 6 report results from a dichotomous measure that indicates whether the person ever went to jail while on ISP or not. We also estimated two addi- tional sets of models (unreported) using two other operationalizations of the jail variable. The first measure was a sum of the total number of times the probationer went to jail while on ISP, and the second was a sum of the total number of days the probationer spent in jail while on ISP. In no case

Table 6 Two-level mixed logit model investigating if jail sanctions as opposed to community- based sanctions influence the likelihood of successfully completing the ISP program (bootstrapped standard errors; n = 737)

b SE

Fixed Effects (Level 1) Independent Variables based on Sanction Events Jail sanction .051 .237 Total # of violations while on ISP -.075 .036*

Independent Variables based on Personal Characteristics Property offender -.774 .191*** Prior felony conviction .058 .194 Prior revocation -.053 .194 Arrested as juvenile -.253 .207 Parolee -.218 .248 High school diploma/GED -.237 .205 Male -.154 .221 Age .000 .011 Non-white -.395 .220 Unmarried .395 .218 Intercept 1.245 .536*

Random Effects (Level 2) τ 0 .494 .110***

Model Statistic Wald χ2 37.47***

* p ≤ .05 ** p ≤ .01 *** p ≤ .001

248 E.J. Wodahl et al. / Journal of Criminal Justice 43 (2015) 242–250

did either of these alternative operationalizations approach levels of statistical significance. Further, the inclusion of these variables did not substantively change any findings which were observed in the models we reported. As such, it very much appears that jail sanctions perform no better or worse than community-based sanctions in these data.

Second, Tables 4 through 6 reported results using both low risk and high risk violations, raising another point which is worth emphasizing. High risk violations generally carry a higher level of harm as well as the potential for more severe repercussions from the probation agency. Because of this, the reported models (and those with the two additional operationalizations of the jail measure) were re-estimated using high risk violations only. Even when considering only the most severe viola- tions, in no case did jail sanctions perform differently than community- based sanctions. The substantive results in the high risk only models also closely mirrored the results reported in the earlier analyses. Thus, there does not appear to be any suppressed relationships between high risk violations, jail sanctions, and community-based sanctions that may have been masked in our reported results.

Third and finally, the analyses in this paper have analyzed the data at the sanction-event-level. To cross-validate results, we estimated similar models that analyzed offenders instead of violation events. The results of these models were substantively similar to the results of the models we reported,11 further speaking to the robustness of the reported find- ings in these data.

Discussion and Conclusions

There is no evidence from the current study to suggest that jail sanc- tions are any more or less effective than community-based graduated sanctions in bringing about offender compliance with release conditions. The imposition of a jail sanction for offender noncompliance as opposed to a community-based sanction did not affect the number of days until the next violation, the number of subsequent violations, or the overall likelihood of completing supervision. Furthermore, the number of times an offender went to jail, the number of days spent in jail, or the timing of the jail sanction did not influence offender outcomes.

There are a number of potential explanations for these null findings. A first possibility is that jail and community-based graduated sanctions pos- sess similar specific deterrent qualities. While the public and criminal justice practitioners often view community-based punishments as being substantially less onerous than incarceration, research has shown that offenders do not necessarily hold the same views, especially when the duration of incarceration is short (May & Wood, 2010; Wodahl, Ogle, Kadleck, & Gerow, 2013). For instance, Wodahl et al. (2013) found that offenders perceived a two day jail sanction as having the same punitive effect as completing six hours of community service or seven days of supervision on electronic monitoring. These findings indicate that offenders do not view short increments of jail as being highly punitive.

These findings might also reflect the relative inconsequentiality of punishment severity in the deterrence framework. Research has consis- tently found that certainty and celerity of punishment are more salient than severity in deterring future misbehavior (Paternoster, 1987). As such, as long as punishments are being meted out in a swift and certain manner, the type of sanction imposed may be of little importance. Finally, the findings might reflect a dilution effect brought on by the criminogenic effects of incarceration. The potential collateral consequences of spending time in jail such as loss of employment or disruption of family relation- ships may be diluting the deterrent benefits of jail sanctions.12

From a policy standpoint, the findings raise serious questions about the solitary use of jail sanctions to respond to offender noncompliance. As discussed earlier, jail is a resource intensive sanction, often costing over $100 per day to incarcerate an offender (Piquero, 2010; Santora, 2013). While jail sanctions are typically limited to a few days, these costs can add up when considering the volume of offenders under com- munity supervision. Added to these costs are the additional time and workload strains created for correctional and court personnel who must

assure offenders’ due process rights are upheld before incarceration is imposed (Kleiman, 2011). Community-based graduated sanctions, by contrast, can be imposed without a need for a formal hearing and with lit- tle or no cost to the agency. In addition to the cost savings, community- based graduated sanctions avoid the collateral consequences often associ- ated with spending time in jail such as exposure to criminal values, loss of employment, and disruption of positive peer networks (Clear et al., 2001; Cullen & Gilbert, 1982; Grogger, 1995). Thus, given that jail sanctions appear to offer no added benefits over community-based sanctions, agen- cies would be wise to consider expanding their repertoire of community- based sanctions and limiting the use of jail sanctions to instances where removing the offender from the community is paramount to the needs of the offender or the safety of the community.

Correctional administrators may be hesitant in reducing their reliance on jail sanctions in favor of community-based alternatives out of concern that the public will view these alternatives as soft measures. However, there is ample evidence to suggest that a large number of citizens favor the use of community sanctions. A Massachusetts study, for example, found that residents were in favor of community-based options over incarceration when parolees would do such things as miss a scheduled meeting, fail routine drug tests, and get caught shoplifting (Roberts, Doble, Clawson, Selton, & Briker, 2005). Similarly, Ohio residents were amenable to community options even for acts of burglary and robbery, with 43% favoring community sentences for a burglary where $250 was stolen and 48% for robbery with no injury (Turner, Cullen, Sundt, & Applegate, 1997). Given that the typical violation for an offender under community supervision is either a technical violation or a relatively low-level criminal offense, public opposition to community-based sanc- tions is unlikely to stem from perceptions of leniency. Furthermore, public support for community sanctions could be increased by illustrating the benefits of these approaches toward reducing recidivism as support for a policy increases when people are knowledgeable of accompanying benefits (Mears & Mancini, 2006; Nagin, Piquero, Scott, & Steinberg, 2006; Piquero & Steinberg, 2010). As such, policymakers should be encouraged to prepare materials for public distribution illustrating the positive effects of community-based sanctions and showing public advo- cacy of these practices.

Although our research findings raise serious questions regarding the use of jail sanctions, replication is necessary to confirm that the results hold in different regions with more diverse populations. The current study is limited to a single state with a relatively homogenous population. For example, while over 80 percent of the study sample is white, national level statistics reveal that whites account for just over half (54%) of all probationers and only 43% of parolees (Herberman & Bonczar, 2014). Research also suggests that race/ethnicity influences how offenders per- ceive a variety of punishments. May and Wood (2010), for example, found that African Americans viewed incarceration as being less punitive than their white counterparts. While the degree to which differences in perceptions might lead to differences in the effectiveness of jail versus community-based graduated sanctions is unclear, efforts should be made to replicate these findings with more racially and ethnically diverse populations.

Future research should also explore whether the effectiveness of jail sanctions versus community-based sanctions varies among jails. Jails vary by organizational design (e.g., linear versus podular designs), staff quality, population density, housing arrangements (e.g., single versus double bunking), type of offenders (e.g, gang versus nongang members, violent versus nonviolent offenders), and recreational and program opportunities. Jails with stricter disciplinary environments and fewer amenities might enhance the punitive aspects of incarcerated experience and lend greater force to its sanctioning impact. On the other hand, jails with many inmate confrontations and large proportions of gang members could enhance the criminogenic effect of the environ- ment to such an extent that the experience is counterproductive.

All things considered, community-based sanctions such as increased treatment participation, electronic monitoring, and written assignments

249E.J. Wodahl et al. / Journal of Criminal Justice 43 (2015) 242–250

are equally as effective as spending time in jail. Since jail sanctions are expensive and can have deleterious effects on both offenders and their families, does this mean that probation agencies should simply stop using jail as a means of punishing offenders? Based on our results, some may clearly see the answer to this question being ‘yes.’ However, we cau- tion against such a ‘black and white’ interpretation. Offenders often need to be removed from the community for their own and the community’s benefit, and jail certainly serves a useful role in these cases. With this im- portant caveat kept in mind, practitioners are urged to err on the side of keeping probationers and parolees out of jail, particularly when there is no urgent necessity to remove the offender from the community. While this may pose no added benefit to successful probation completion, doing so would keep families intact, help offenders stay employed, and could potentially save local governments a considerable amount of money.

Notes

1 While research on the use of graduated sanctions has generally revealed positive findings, other studies have failed to replicate these results, leading some to question the efficacy of these deterrence-based interventions (Duriez et al., 2014).

2 The ISP program supervises both probationers and parolees; thus, the term offender here refers to both groups.

3 Full descriptive statistics on the ISP population from which the study sample is drawn are available from the lead author upon request.

4 We additionally constructed measures of the number of times each respondent had been to jail as well as the total number of days the respondent spent in jail. Analyses using these measures were extremely similar to the forthcoming results. We comment on this at the end of the Results section.

5 We would like to thank an anonymous reviewer for recommending this control variable.

6 The minute differences in descriptive statistics between the offenders in the person- level file (N = 283) in Table 2 and the offense-level file (N = 861) in Table 3 are due the transformation of the data to the offense-level file.

7 While one would want jail time to be isolated from a deterrence standpoint, proba- tion and parole officers could choose to use jail sanctions in an attempt to increase an of- fender’s time until next violation (an incapacitation effect). As such, we re-estimated this model using a dependent variable of an uncompensated number of days until the next vi- olation. Substantively, the results were similar.

8 The dependent variable of days until next violation could also be viewed as time to failure, which can be analyzed through survival analysis techniques. As such, this model was also estimated using a continuous, semi-parametric Cox Regression. No substantive differences were observed between the models. To avoid adding considerable analytical complexity without strengthening our findings, we chose to report the standard multilev- el model.

9 Instead of being normally distributed, the dependent variable in this model was sig- nificantly skewed right (p ≤ .001). To ensure that our results were not being affected by this, we re-estimated this model 1) with a logged dependent variable and 2) with a Poisson mixed model. The results were substantively similar. As such, we report a non- logged, standard mixed model.

10 The jail and percent ISP completion variables were grand mean centered prior to creating the interaction term.

11 We also estimated the days until next violation models with a measure of compen- sated (i.e., | number of days until next violation – number of prior days spent in jail |) and uncompensated (raw number of days in between violations, including jail time). Relevant to the model presented in Table 4, individual-level results demonstrated that offenders who experienced a jail sanction at some point during probation were more likely to have a higher number of days until their next probation violation with the uncompensated, but not the compensated measure. Similar to our event-level findings, this once again suggests that jail sanctions do little to stop violations when an offender is not in confinement.

12 Some scholars might be surprised that age did not influence each of the outcome measures considering the prevalence of the age-crime relationship in criminological re- search. In this study, all participants are high risk offenders and a number of criminal jus- tice background variables are controlled, all of which likely diminish the impact of age in this all-adult sample.

References

Adams, B. (2013, September 17). 67% of new Utah prisoners there for probation, parole violation. The Salt Lake Tribune (http://archive.sltrib.com/printfriendly.php?id= 56883340&itype=cmsid).

Akers, R. L. (1997). Criminological theories: Introduction and evaluation. Los Angeles, CA: Roxbury.

American Probation and Parole Association (APPA) (2013). Effective responses to offender behavior: Lessons learned for probation and parole supervision. (Retrieved October 19, 2014, from http://www.appa-net.org/eWeb/docs/APPA/pubs/EROBLLPPS-Report.pdf).

Braga, A. A., & Weisburd, D. L. (2012). The effects of focused deterrence strategies on crime: A systematic review and meta-analysis of the empirical evidence. Journal of Research in Crime and Delinquency, 49, 323–358.

Clear, T. R., & Rose, D. R. (2003). Individual sentencing practices and aggregate social problems. In D. F. Hawkins, S. L. MyersJr., & R. N. Stone (Eds.), Crime control and social justice: The delicate balance (pp. 27–52). Westport, CT: Greenwood Press.

Clear, T. R., Rose, D. R., & Ryder, J. A. (2001). Incarceration and the community: The problem of removing and returning offenders. Crime and Delinquency, 47, 335–351.

Cornelius, G. F. (2008). The American jail: Cornerstone of modern corrections. Upper Saddle River, NJ: Pearson.

Cullen, F. T., & Gilbert, K. E. (1982). Reaffirming rehabilitation. Cincinnati, OH: Anderson. Duriez, S. A., Cullen, F. T., & Manchak, S. M. (2014). Is Project Hope creating a false sense of

hope? A case study in correctional popularity. Federal Probation, 78(2), 57–70. Freudenberg, N. (2001). Jails, prisons, and the health of urban populations: A review of

the impact of the correctional system on community health. Journal of Urban Health, 78, 214–235.

Gibbs, J. P. (1975). Crime, punishment, and deterrence. New York, NY: Elsevier Scientific Publishing Company.

Grattet, R., Petersilia, J., Lin, J., & Beckman, M. (2009). Parole violations and revocations in California: Analysis and suggestions for action. Federal Probation, 73, 2–11.

Gray, M. K., Fields, M., & Maxwell, S. R. (2001). Examining probation violations: Who, what, and when. Crime & Delinquency, 47, 537–557.

Grogger, J. (1995). The effect of arrests on the employment and earnings of young men. Quarterly Journal of Economics, 110, 51–71.

Grommon, E., Cox, S. M., Davidson, W. S., & Bynum, T. S. (2013). Alternative models of instant drug testing: evidence from an experimental trial. Journal of Experimental Criminology, 9, 145–168.

Haney, C. (2001). The psychological impact of incarceration: Implications for post-prison adjustment. http://aspe.hhs.gov/hsp/prison2home02/haney.pdf

Hawken, A., & Kleiman, M. (2009). Managing drug involved probationers with swift and certain sanctions: Evaluating Hawaii’s HOPE. Washington, D.C.: U.S. Department of Justice.

Herberman, E. J., & Bonczar, T. P. (2014). Probation and parole in the United States, 2013. Washington, D.C.: U.S. Department of Justice.

Jones, M. (1995). Predictors of success and failure on intensive probation supervision. American Journal of Criminal Justice, 19, 239–254.

Justice Policy Council, Criminal (2002). Trends, profile and policy issues related to felony probation revocations in Texas. (http://www.lbb.state.tx.us/Public_Safety_Criminal_ Justice/Reports/felpro2.pdf).

Kilmer, B., Nicosia, N., Heaton, P., & Midgette, G. (2013). Efficacy of frequent monitoring with swift, certain, and modest sanctions for violations: Insights from South Dakota’s 24/7 Sobriety Project. American Journal of Public Health, 103, e37–e43.

Kleiman, M. A. R. (2011). Justice reinvestment in community supervision. Criminology & Public Policy, 10, 651–659.

Latessa, E. J., Listwan, S. J., & Koetzle, D. (2015). What works (and doesn’t) in reducing recidivism. New York, NY: Routledge.

May, D. C., Applegate, B. K., Ruddell, R., & Wood, P. B. (2014). Going to jail sucks (and it really doesn’t matter who you ask). American Journal of Criminal Justice, 39, 250–266.

May, D. C., & Wood, P. B. (2010). Ranking correctional punishments: Views from offenders, practitioners, and the public. Durham, NC: Carolina Academic Press.

May, D. C., Wood, P. B., Mooney, J. L., & Minor, K. I. (2005). Predicting offender-generated exchange rates: implications for a theory of sentence severity. Crime & Delinquency, 51, 373–399.

Mears, D. P., & Mancini, C. (2006). The 2006 public opinion survey of Florida citizens concerning the Florida Department of Corrections. Center for Criminology and Public Policy. Research: Florida State University (Retrieved from http://www.criminologycenter.fsu.edu/p/pdf/ 2006_DOC_Survey_Rpt.pdf).

Morgan, K. D. (1994). Factors associated with probation outcome. Journal of Criminal Justice, 22, 341–353.

Nagin, D. S., Cullen, F. T., & Jonson, C. L. (2009). Imprisonment and reoffending. Crime and Justice, 38, 115–200.

Nagin, D. S., Piquero, A. R., Scott, E. S., & Steinberg, L. (2006). Public preferences for rehabilitation versus incarceration of juvenile offenders: Evidence from a contingent valuation survey. Criminology and Public Policy, 5, 627–652.

National Association of Drug Court Professionals (NADCP) (1997). Defining drug courts: The key components. Washington, DC: U.S. Department of Justice.

Paternoster, R. (1987). The deterrent effect of the perceived certainty and severity of punishment: A review of the evidence and issues. Justice Quarterly, 4, 173–217.

Paternoster, R. (2010). How much do we really know about criminal deterrence? The Journal of Criminal Law & Criminology, 100, 765–824.

Petersilia, J. (2003). When prisoners come home: Parole and prisoner reentry. New York, NY: Oxford University Press.

Piquero, A. (2010). Cost-benefit analysis for jail alternatives and jail. Retrieved November 8, 2014, from http://criminology.fsu.edu/wp-content/uploads/Cost-Benefit-Analysis- for-Jail-Alternatives-and-Jail.pdf.

Piquero, A. R., & Steinberg, L. (2010). Public preferences of rehabilitation versus incarcer- ation of juvenile offenders. Journal of Criminal Justice, 38, 1–6.

Pogarsky, G. (2009). Deterrence and decision making: Research questions and theoretical refinements. In M. D. Krohn (Eds.), Handbook on Crime and Deviance (pp. 241–258). New York, NY: Springer.

Roberts, C. A., Doble, J., Clawson, E., Selton, C., & Briker, A. (2005). Rethinking justice in Massachusetts: Public attitudes toward crime and punishment. http://b.3cdn.net/ crjustice/d52ab7c9b58c890adc_81m6i8x2z.pdf.

250 E.J. Wodahl et al. / Journal of Criminal Justice 43 (2015) 242–250

Rodgers, J. L. (1999). The bootstrap, the jackknife, and the randomization test: A sampling taxonomy. Multivariate Behavioral Research, 34, 441–456.

Rollo, V. N., Jr. (1988). 99 days and a get up. Dallas, TX: Open, Inc. Sabol, W. J., & Lynch, J. P. (2003). Assessing the longer-run consequences of incarceration:

Effects on families and employment. In D. F. Hawkins, S. L. MyersJr., & R. N. Stone (Eds.), Crime control and social justice: The delicate balance (pp. 3–26). Westport, CT: Greenwood Press.

Santora, M. (2013, August 23). City’s annual cost per inmate is $168,000, study finds. The New York Times. http://www.nytimes.com/2013/08/24/nyregion/citys-annual-cost- per-inmate-is-nearly-168000-study-says.html?_r = 0.

Solomon, A. L., Osborne, J. W. L., LoBuglio, S. F., Mellow, J., & Mukamal, D. A. (2008). Life after lockup: Improving reentry from jail to the community. Washington, DC: Urban Institute.

Steiner, B., Makarios, M. D., Travis, L. F., & Meade, B. (2012). Examining the effects of community-based sanctions on offender recidivism. Justice Quarterly, 29, 229–257.

Taxman, F. S., Soule, D., & Gelb, A. (1999). Graduated sanctions: Stepping into accountable systems and offenders. The Prison Journal, 79, 182–204.

Travis, J., & Lawrence, S. (2002). Beyond the prison gates: The state of parole in America. Washington, D.C.: The Urban Institute.

Turner, M. G., Cullen, F. T., Sundt, J. L., & Applegate, B. K. (1997). Public tolerance for community-based sanctions. The Prison Journal, 77, 6–26.

Western, B., Kling, J. R., & Weiman, D. F. (2001). The labor market: Consequences of incar- ceration. Crime & Delinquency, 47, 410–427.

Wilson, D., Gallagher, C., & Mackenzie, D. (2000). A meta-analysis of corrections-based education, vocation, and work programs for adult offenders. Journal of Research in Crime and Delinquency, 37, 347–368.

Wodahl, E. J., & Garland, B. (2009). The evolution of community corrections: The enduring influence of the prison. The Prison Journal, 89, 81S–104S.

Wodahl, E. J., Garland, B., Culhane, S. E., & McCarty, W. P. (2011a). Utilizing behavioral interventions to improve supervision outcomes in community-based corrections. Criminal Justice and Behavior, 38, 386–405.

Wodahl, E. J., Ogle, R., & Heck, C. (2011b). Revocation trends: A threat to the legitimacy of community-based corrections. The Prison Journal, 91, 207–226.

Wodahl, E. J., Ogle, R., Kadleck, C., & Gerow, K. (2013). Offender perceptions of graduated sanctions. Crime & Delinquency, 59, 1185–1210.

Wood, P. B., & Grasmick, H. G. (1999). Toward the development of punishment equivalen- cies: Male and female inmates rate the severity of alternative sanctions compared to prison. Justice Quarterly, 16, 19–50.

Wooldredge, J. D. (1999). Inmate experiences and psychological well-being. Criminal Justice and Behavior, 26, 235–250.