Modeling Repeat DUI Offender Probation Outcomes Using the Wisconsin Risk Need Assessment

Abstract

This multiyear study of felony driving while intoxicated (DWI) probationers explores the efficacy of the Wisconsin Risk Need Assessment tool along with sociodemographic factors as measures of probation outcomes. To date, few studies have explored the relationship between risk assessment data and technical violations as well as subsequent arrests of individuals on probation. The sample for this study consists of 596 chronic DWI offenders on community supervision in one county who either had been rearrested for a new offense, violated a technical condition of their probation, or committed no violations within the first 5 years of community supervision. The findings are that older defendants and those who had more dependents were more likely to have committed a technical violation compared with the other two groups. Those rearrested for a new offense were slightly younger compared with the other two groups, less likely to be employed and younger at the age of first adjudication of guilt. This study highlights the limited overall utility of the Wisconsin tool in determining probation outcomes and that static factors may be as important as dynamic factors when developing a supervision strategy for chronic DWI offenders.

Keywords

probation outcomes Wisconsin Risk Need DUI offender recidivism community supervision

Introduction

Drunk driving on our nation’s streets and highways continues to be a major social concern. Although the number of alcohol-related fatalities remains high, rates have decreased 7% since 2008 (National Highway Traffic Safety Administration [NHTSA], 2018), and the number of arrests for driving while intoxicated (DWI) has also declined. For instance, there were 806,369 DWI arrests in 2016 according to the FBI, a decrease of approximately 28% from 2014 (Driving While Intoxicated: Federal Bureau of Investigation, 2016). Numerous factors are likely to have contributed to this decline, such as greater social awareness from media campaigns such as “Drive Sober or Get Pulled Over” and “Drink-Drive-Go-To-Jail” (Buckley et al., 2016). Adding to the effects of greater public awareness of intoxicated driving is the fact that treatment for alcohol-dependent individuals has significantly improved with the expansion of evidence-based interventions (Drake, 2011; Taxman & Belenko, 2012). As a result, community corrections agencies have integrated a range of treatment options with court-ordered supervision programs to address chronic drunk driving (Miller et al., 2015; Nochajski et al., 2013).

Although the incidence of arrests for drunk driving has declined in recent years, the issue of chronic drunk driving continues to be a serious public safety concern to the nation. To address chronic drunk driving, state legislatures have responded with stiffer penalties for violators. Currently, 22 states make a third DWI conviction a felony offense, and in another 19 states, a fourth conviction is a felony (Mothers Against Drunk Driving [MADD], 2015). There are also time limits in some states that require convictions to occur within a specific time frame (e.g., 5, 6, or 10 years). Texas, like many other states, gives judges discretion to sentence a felony DWI offender to prison or community supervision, the latter of which can extend up to 10 years (Lee & Teske, 2015). The extended period of supervision means that probation departments must devote considerable resources to monitoring chronic DWI offenders, provide them with treatment, and respond to violations and implement sanctions when necessary. However, using rearrest as the sole measure of DWI probation success is somewhat problematic, given that DWI offenders are much less likely to be rearrested for a new offense than other drug users (National Drug Court Institute [NDCI], 2004). As is the case with other probationers, those on community supervision for DWI often incur technical violations leading to other negative outcomes such as revocation, jail time, or loss of employment (Wodahl et al., 2015). Data from the Council of State Governments showed that 45% of state prison admissions are the result of either a probation or parole violation, with technical violations accounting for 2.8 billion and new offense violations accounting for 6.5 billion annually (Council for State Governments, 2019). As such, probation staff must be able to identify the factors that have the greatest impact on both technical violations as well as new arrests. Findings from this research will help identify appropriate supervision strategies for repeat DWI offenders.

Prior research on chronic DWI has focused on the efficacy of treatment courts and similar programs on recidivism (Mitchell et al., 2012). While this research has produced valuable insights into how to respond and treat chronic DWI defendants, a large number of these individuals do not participate in specialized treatment programs (Belenko et al., 2011). Unlike other substance abuse offenders, DWI offenders frequently do not have extensive criminal histories and thus are more likely to have avoided court-mandated substance abuse treatment. Even among those individuals who are referred to drug treatment courts, other issues such as the lack of capacity and funding at the local level, eligibility and recruiting policies, and the tendency for defendants “opt-out” of these programs affect participation rates. Still yet, those convicted of felony DWI and placed on regular community supervision are subject to aggressive alcohol monitoring and longer terms of supervision compared with those on misdemeanor probation. Conditions of felony probation for DWI are also typically more stringent compared with misdemeanor probation, the former being more likely to involve completion of an inpatient or outpatient treatment program, randomized and frequent drug testing, and the suspension of one’s driver’s license (NHTSA, 2009). Furthermore, the consequences of failing on felony DWI probation are much more severe than misdemeanor DWI probation. A felony DWI revocation can lead to a term of confinement in a state correctional facility, whereas revocation of misdemeanor DWI probation results in a jail sentence of up to 1 year (Kenneally, 2016). Hence, the task of monitoring felony DWI offenders is more complex in terms of the conditions imposed on the offender, the duration of the supervision term, and the range of resources that may be used to effectively supervise the defendant.

Treatment for Chronic DWI Offenders

The principles of effective intervention, which are grounded in a General Personality and Cognitive Learning Perspective, are commonly known as risk-need-responsivity (RNR; Andrews & Bonta, 2010). These principles provide the theoretical foundation for managing substance abuse populations through a combination of treatment and supervision strategies, with the primary goal of reducing recidivism. Risk factors can be further delineated as static or dynamic, and are generally regarded to be predictive of recidivism (Andrews & Bonta, 2010). Static factors are considered those that do not change (i.e., prior arrest history), whereas dynamic factors (i.e., level of addiction, attitudes, quality of relationships, employment status), also referred to as “criminogenic needs,” can change over time. Generalized assessment tools, as well as specialized tools designed for alcohol dependency, have some combination of static and dynamic measures, with the latter being the focus of intervention strategies designed to curb alcohol and other drug use. Responsivity refers to the goal of aligning treatment and rehabilitation strategies with the relative risk level of the individual. The intensity and duration of treatment, along with the format in which it occurs, are all important factors affecting treatment responsivity (Wormith & Zidenberg, 2018).

Early research on the effects of treatment and rehabilitation for DWI offenders focused on education programs as the primary mode of treatment (Wells-Parker et al., 1995). However, these studies generally showed ineffective results in reducing alcohol abuse (Donovan et al., 1990). Reasons for the failure of early DWI programs appear to be related to their punitive nature and lack of treatment, with the primary strategies being license suspension and the use of ignition interlocks to deter drunk driving (Mullen et al., 2015). In a meta-analysis of studies examining the effects of remedial programs on DWI recidivism, Wells-Parker et al. (1995) found that DWI recidivism decreased by 7% to 9% compared with those who received treatment, and by 24% when participants received a combination of DWI education, treatment, and supervision.

The expansion of evidence-based practices (e.g., cognitive-behavioral therapy, medically-assisted therapy) has greatly improved the quality of treatment provided to DWI offenders (Lapham & McMillan, 2011; Ouimet et al., 2013). Treatment options for DWI offenders have benefited from evidence-based practices used to address other forms of substance abuse. Many treatment programs currently use a combination of psychiatric and cognitive-based therapies to address alcohol addiction due to the comorbid nature of the disease (Freeman et al., 2011). In a study exploring the treatment needs of DWI offenders from the community and a residential facility (n = 119), Mullen et al. (2015) found that most had complex needs beyond alcohol abuse. Of the entire sample, 26% had been previously diagnosed with a psychiatric disorder other than substance abuse disorder, and more than 60% were not receiving treatment for their conditions. Therefore, the efficacy of rehabilitation strategies to address DWI recidivism may be enhanced through some combination of psychiatric and cognitive-based treatment, depending on the needs of the offender.

Assessment Tools and DWI Offender Outcomes

As a matter of practice, probation agencies employ various risk assessment tools to assess the severity of a defendant’s alcohol addiction and overall risk level. These tools rely largely on either clinical data or self-reported information, or some combination thereof. Going back to the 1970s, researchers have attempted to develop risk assessment tools specifically to assess future DWI offending (for a review, see DeMichele et al., 2016).

As one of the first tools of its kind, the Mortimer-Filkins (MF) attempted to differentiate between the social drinker, the presumptive problem drinker, and problem drinker (Mortimer et al., 1971). One key limitation with this tool, and subsequent tools, was that it was not oriented toward probation. Subsequent research began to focus on traits that were thought to be more closely linked with excessive drinking, such as risk-taking, manipulativeness, and psychopathic deviance (Cavaiola et al., 2003). An instrument designed for this purpose was the Substance Abuse Screening Inventory (SASSI), yet subsequent research shows this tool does poorly in predicting substance use disorders (Feldstein & Miller, 2007).

Researchers then began to look for ways to expand the range of assessment items and domains to yield an instrument that better predicted chronic DWI behavior. The Driver Risk Inventory (DRI) and its newer version (DRI-II) was one such attempt and included five scales: truthfulness, alcohol, drugs, driver risk, and stress quotient. The Driver Risk Scale differentiated this instrument from prior ones because it contained separate measures that focused on driving infractions and other behaviors that do not have a specific connection with alcohol or drug use. Subsequent research examining the tool’s efficacy showed that factors such as adult imprisonment, prior revocations, previous moving violations, and accidents were predictive of subsequent DWIs, though the overall predictive validity of this tool was found to be weak (Chang et al., 2002). To summarize, the authors argue that assessment instruments designed to predict DWI recidivism suffer from “tunnel vision” because they focus too heavily on an individual’s drinking behavior. They further argue that a framework for understanding differences between single and multiple DWI offenders should incorporate a broader range of factors that have demonstrated a stronger connection to general offender recidivism (Farrington, 2003; Nagin, 2005). DeMichelle et al. (2016) reached a similar conclusion in their research, concluding that “the level of AOD is not the underlying characteristic shaping DWI patterns” (p. 1588).

More recent research has evaluated the efficacy of general assessment tools and more specialized tools using the same sample of DWI offenders. DeMichele et al. (2016) reviewed a sample of probationers and parolees (n = 4,022) to determine differences between chronic and first time DWI offenders. They used data from two assessment instruments (Level of Service Inventory-Revised [LSI-R] and Adult Substance Abuse Survey) along with demographic and social factors to assess relationships with DWI offender types. The analysis revealed that the factors distinguishing chronic and first-time DWI offenders bear less of a relationship with alcohol and other drug use than they do with emotional stability, a lack of willingness to change, and their history of drug treatment. In terms of demographics, the authors found that the only variable reaching significance in the models was age, with offenders between the ages of 30 to 44 years more likely to be in the chronic offending group.

The Wisconsin Risk Need Assessment

Although many generalized risk assessment tools have been developed to determine offender risk, research on their efficacy with high-risk DWI defendants is sparse. One widely used instrument by probation and parole agencies is the Wisconsin Risk Need Assessment (Jones et al., 1999). The appeal of the instrument to community corrections agencies is that it is much shorter and less time consuming to administer than other generalized assessment tools such as the LSI-R. The Wisconsin was developed by the Wisconsin Department of Corrections in the 1970s to determine the appropriate level of supervision, given the individual’s risk and needs profile (Wright et al., 1984). Specific measures were selected by researchers to determine their connection with different types of probation outcomes such as technical violations, rearrest, and absconded (Baird et al., 1979). The tool encompasses separate scales for risk and needs and assigns separate scores that supervision officers use to make judgments about the offender’s rehabilitative needs and the likelihood the offender will complete probation (Henderson & Miller, 2011). To assess risk, the instrument includes 11 items—two of which specifically deal with previous drug/alcohol use as well as several static factors such as the number of prior probation supervision periods, felony adjudications, and juvenile commitments. The needs scale includes 12 measures, ranging from alcohol and drug use, mental stability, quality of family and marital relationships, employment status, and education level.

Prior research on the predictive utility of the Wisconsin has largely produced unimpressive results (Henderson & Miller, 2011). In fact, most studies have produced correlation coefficients below 50% chance levels, which are of little use to community supervision officers who are routinely tasked with making decisions regarding appropriate supervision levels and treatment of offenders (Grommon et al., 2013; Steiner et al., 2012). However, of the studies that have been conducted, no research has examined the predictive validity of the Wisconsin on DWI offenders, specifically. This may be due to the assumption that DWI offenders are conceptually different from other offenders (Curtis et al., 1994) and that specialized instruments have been developed to assess this population. This assumption may not, however, hold for chronic DWI offenders who often possess more extensive criminal histories and are subject to longer periods of supervision than first-time DWI offenders.

Like other assessment instruments, a number of criminal history variables are included in the Wisconsin Risk Need tool. A person’s criminal history has long been established as a key indicator of future criminal behavior. The criminal history measures in the Wisconsin Risk Need tool include the age at first adjudication, number of prior periods of probation and parole supervision, number of probation/parole revocations, overall number of adult or juvenile adjudications, and the number of adult or juvenile adjudications for an assaultive offense. Although prior research has established that criminal history measures are correlated with future offending, how these measures affect all types of probation outcomes is less clear. In one study that examined measures of rearrest using a non-DWI sample of probationers, Henderson and Miller (2011) compared the predictive validity of the original Wisconsin with a reweighted version and found that four of the six criminal history (risk) measures in the instrument were significantly related to the likelihood of rearrest. Similarly, DeMichele et al. (2016) found that individuals without three or more prior offenses and those not arrested before the age of 16 years were 15.5% and 29% less likely to be chronic DWI offenders compared with those with such offense histories. However, their study involved offenders on parole as well as probation, and like the majority of prior work only used subsequent convictions as an outcome measure.

The current study seeks to determine the efficacy of the Wisconsin in distinguishing chronic DWI probationers who are likely to incur subsequent arrests, technical violations of probation conditions, or not to incur any violation during their supervision period (5 years). It is important to examine this tool, given its popularity and that the vast majority of individuals convicted of DWI are placed on community supervision (NHTSA, 2009). Given that alcohol and other drug use has not been consistently linked to DWI recidivism (DeMichelle et al., 2016), there remains some question regarding how specific risk and needs measures may help community supervision officers discern independent risk factors related to compliance of conditions of probation as well as future criminal behavior.

Method

Participants

The data for this study were taken from one county’s adult probation department in a large, southwestern state for individuals serving a probation sentence for felony DWI (n = 596) between the years 2006 to 2011. Felony DWI probation is reserved for individuals convicted of a third DWI offense (or more) and the time on supervision can range from 2 to 10 years. In reviewing the distribution of probation terms, the vast majority were given a sentence of 5 years (691/1291; 53.5%). Thus, cases with a term of 5 years of community supervision were selected as the target sample from the larger population. A list of defendants admitted to the county’s Felony Drug Court Program during this period was obtained and cross-referenced to the larger list. Probationers who had participated in the program (n = 49) and/or who had been sentenced to long-term residential treatment (4 months or greater; n = 27) were excluded from the sample to ensure the sample was relatively homogeneous in terms of the treatment and supervision received. The final sample consisted of 596 after dropping these cases and those with excessive missing data. Each record included information on social demographic characteristics, the number of violations, and length of time an ignition interlock device had been installed on their vehicle, as well as the individual criminal history items and composite scores from the Wisconsin Risk and Needs Assessment.

Analytical Strategy

A common technique used to examine offender recidivism is multinomial logistic regression (Lowenkamp & Latessa, 2002; Mertler & Vannatta, 2010). Multinomial logistic regression is an adaptation of the more well-known binary logistical regression and includes assessments of all paired outcomes. Rather, instead of conducting separate logistic regression models to compare odds ratios for those rearrested for a new offense, technical violators, and nonviolators, these comparisons all occur in one parsimonious model. Furthermore, multinomial logistic regression also includes estimates of goodness-of-fit and odds ratios of the coefficients across all groups. The odds ratios indicate the extent to which specific variables affect each two-way model (probation violators vs. subsequent offenders).

Generally, there are three groups of individuals that probation officers supervise—those who are compliant, those who commit technical violations, and those who are arrested for a subsequent offense. Although technical violations may result in a revocation of probation, it is the commission of a new offense that is generally viewed as a more serious offense. This analysis compared those individuals who incurred a probation violation (whether it resulted in revocation or not) and a subsequent offense with those who were compliant (e.g., incurred no violations) during a 5-year supervision period.

Results

Commensurate with the demographics of the community, just more than 68% were of Hispanic ethnicity and the majority (68.8%) had graduated from high school. About 79% were employed (which includes both full and part time), and the average age was roughly 46 years old (SD = 9.9). The average number of dependents was 1.41 (SD = 1.6). Risk and needs scores were collected at the time the offender began his or her probation term. The Wisconsin derives a risk score ranging from 0 to 15 or greater, with scores between 0 and 7 indicating low risk, 8 and 14 as moderate risk, and 15 or more as maximum risk. Likewise, needs scores can range from 0 to 30 or greater, with scores ranging from 0 to 14 considered minimum, 15 to 29 medium, and 30 or more as high. The average risk score (14.20; SD = 5.6) fell into the upper end of the moderate level and the average needs score (16.4, SD = 7.2) fell on the low end of the moderate level. Data on the number of months an individual was required to use an ignition interlock or other alcohol monitoring device in their vehicle were factored in, given the potential deterrent effect it may have on subsequent offending. Subsequent offenders had comparatively longer monitoring periods (M = 7.66, SD = 18.6) than did technical violators (M = 6.06; SD = 14.19), or those who did not have any violations (M = 5.25; SD = 14.6). In terms of probation outcomes, 477 (80.0%) incurred no violations, 98 (16.4%) had at least one violation, and 21 (3.5%) had been rearrested for a new offense.¹

Table 1 provides descriptive statistics for nonviolators, technical violators, and subsequent offenders. There are some notable differences in the criminal history measures of the Wisconsin across the three groups. Chronic DWI offenders are much more likely to have adjudicated guilty by the age of 19 years (70.1%) compared with technical violators (43.1%) and nonviolators (48.1%). In addition, subsequent offenders have a comparatively higher rate of prior probation/parole revocations (41.2%) than do nonviolators (32.4%), and they are somewhat more likely to have at least two prior felony adjudications of guilt (29.4%) than nonviolators (16.1%).

Table 1.

Summary of Study Variables of DWI Participants, by Probation Outcome (n = 596).

Variables	Coding	No violation (%; n = 477)	Technical violation (%; n = 98)	Subsequent arrest (%; n = 21)
Risk items
Address changes
None	0	299 (76.3)	44 (61.1)	8 (47.1)
One	2	77 (19.6)	24 (33.3)	8 (47.1)
Two or more	3	16 (4.1)	4 (5.6)	1 (5.9)
Employment
60% or more	0	274 (69.9)	55 (76.4)	7 (41.2)
40%–59%	1	55 (14)	8 (11.1)	6 (35.3)
Less than 40%	2	63 (16.1)	9 (12.5)	4 (23.5)
Alcohol usage
No relationship	0	22 (5.6)	6 (8.3)	1 (5.9)
Probable relationship	1	9 (2.3)	1 (1.4)	0
Close relationship	2	361 (92.1)	65 (90.3)	16 (94.1)
Drug usage
No relationship	0	290 (74)	55 (76.4)	8 (47.1)
Probable relationship	1	64 (16.3)	10 (13.9)	3 (17.6)
Close relationship	2	38 (9.7)	7 (9.7)	6 (3.5)
Attitude
Motivated to change	0	157 (40)	29 (40.3)	7 (41.2)
Somewhat motivated	3	183 (46.7)	33 (45.8)	7 (41.2)
Not motivated	5	52 (13.3)	10 (13.9)	3 (17.6)
Age at first adjudication (years)
19 or younger	0	190 (48.5)	31 (43.1)	12 (70.1)
20–23	2	119 (30.3)	23 (31.9)	2 (11.8)
24 or older	4	83 (21.2)	18 (25)	3 (17.6)
Prior periods of probation/parole
None	0	24 (6.1)	4 (5.6)	1 (5.9)
1 or more	4	368 (93.9)	68 (94.4)	16 (94.1)
Prior probation/parole revocations
None	0	265 (67.6)	45 (62.5)	10 (58.8)
1 or more	4	127 (32.4)	27 (37.5)	7 (41.2)
Prior felony adjudications of guilt
None	0	242 (61.7)	42 (58.3)	10 (58.8)
One	2	87 (22.2)	16 (22.2)	2 (11.8)
Two or more	4	63 (16.1)	14 (19.4)	5 (29.4)
Prior offenses
None	0	319 (81.4)	56 (77.8)	11 (64.7
Burglary/theft/autotheft/robbery	2	62 (15.8)	15 (20.8)	4 (23.5)
Worthless checks or forgery	4	11 (2.8)	1 (1.4)	2 (11.8)
Adjudications for assaultive offense ≥5 years
No	0	331 (84.4)	58 (80.6)	14 (82.4)
Yes	8	61 (15.6)	14 (19.4)	3 (17.6)
Needs items
Academic/vocational skills
High school or above	−1	134 (34.2)	28 (38.9)	4 (23.5)
Adequate skills	0	187 (47.7)	29 (40.3)	7 (41.2)
Less than adequate	2	59 (15.1)	13 (18.1)	6 (35.3)
Minimal skill level	4	12 (3.1)	2 (2.8)	0
Employment
Satisfactory over 1 year	−1	79 (20.2)	16 (22.2)	0
Retired/student and secure	0	194 (49.5)	40 (55.6)	85 (29.4)
Unsatisfactory with adequate job skills	2	92 (23.6)	14 (19.4)	11 (64.7)
Unemployed and lacking skills	6	27 (6.9)	2 (2.8)	1 (5.9)
Financial management
Favorable and stable	−1	9 (2.3)	1 (1.4)	0
Few difficulties	0	83(21.2)	13 (18.1)	1 (5.9)
Some difficulties	3	254(64.8)	50 (69.4)	13 (76.5)
Frequent problems	5	46 (11.7)	8 (11.1)	3 (17.6)
Family relations
Good support	−1	18 (4.6)	3 (4.2)	0
Relatively stable	0	168 (42.9)	34 (47.2)	7 (41.2)
Some disorganization	3	172 (43.9)	30 (41.7)	9 (52.9)
Major stress/disorganization	5	34 (8.7)	5 (6.9)	1 (5.9)
Companions
Good support	−1	11 (2.8)	1 (1.4)	1 (5.9)
No adverse relationships	0	178 (45.4)	26 (36.1)	6 (35.3)
Occasional problems	2	179 (45.7)	44 (61.1)	8 (11.1)
Mostly negative relationships	4	24 (6.1)	1 (1.4)	2 (2.8)
Emotional stability
Well-adjusted	−2	9 (2.3)	3 (4.2)	0
Fairly adjusted	0	329 (83.9)	55 (76.4)	14 (82.4)
Slightly unstable	4	47 (12)	13 (76)	2 (2.3)
Very unstable	7	7 (1.8)	1 (1.4)	1 (5.9)
Alcohol usage problems
None	0	20 (5.1)	2 (2.8)	1 (5.9)
Some abuse	3	96 (24.5)	16 (22.2)	1 (5.9)
Frequent abuse	6	276 (70.4)	54 (75)	15 (88.2)
Drug problems
None	0	299 (76.3)	57 (79.2)	10 (58.8)
Some abuse	3	77 (19.6)	14 (19.4)	4 (23.5)
Frequent abuse	6	16 (4.1)	1 (1.4)	3 (17.6)
Mental ability
Good functioning	0	373 (95.2)	68 (94.4)	17 (100)
Some instability	3	17 (4.3)	4 (5.6)	0
Severe disruption	6	2	0	0
Health
Good	0	343 (87.5)	62 (86.1)	15 (82.2)
Some interference	1	37 (9.4)	9 (12.5)	1 (5.9)
Chronic illness	2	12 (3.1)	1 (1.4)	1 (5.9)
Sexual behavior
No dysfunction	0	392 (99)	72	17
Minor problems	3	2	0	0
Chronic problems	5	—	—	—
P.O. impression of defendant’s needs
Well-adjusted	−1	3 (1)	1 (1.4)	0
No needs	0	26 (6.6)	4 (5.5)	1 (5.6)
Moderate needs	3	219 (5.6)	42 (58.3)	9 (52.9)
High needs	5	144 (36.7)	25 (34.7)	7 (41.2)
Controls
Male		418 (87.6)	89 (90.8)	18 (85.7)
Married		153 (32.4)	29 (29.6)	4 (19)
Hispanic		329 002,869.1)	64 (65.3)	13 (61.9)
HS graduate		329 (68.9)	67 (68.4)	14 (66.7)
Employed		385 (80.7)	76 (77.6)	13 (61.9)
M age (SD)		477 (9.78)	98 (10.65)	21 (6.95)
M dependents (SD)		1.35 (1.57)	1.69 (1.70)	1.62 (1.60)
M ignition (SD)		5.25 (14.61)	6.06 (14.19)	7.66 (18.64)

Note. Assaultive offense = assaultive offense or one in which involves the use of a weapon, physical force, or the threat of force. DWI = driving while intoxicated.

Multinomial Logistic Regression Results

The multivariate analysis focused on how the independent variables affected the two types of violators (e.g., those arrested for a new offense and technical violators) relative to those who incurred no violations during the follow-up period. Model 1 examines the relationship among the independent variables and technical violations (coded 1), with those offenders having no violations coded “0.” A number of demographical variables emerge as significant predictors of technical violations in the analysis. Not surprisingly, age was significant in the model, with every 1-year increase in age reducing the odds an offender would receive a technical violation by about 3%. The number of dependents was also found to be a significant factor, but not in the expected direction. Specifically, each additional dependent increased the odds the offender violated their probation by approximately 17%. Of the risk and needs measures, the only item that was significant in the model was the number of address changes in the last year. Furthermore, as the number increased from 0 to 1 or 2 or more, the odds of incurring a technical violation increased by about 44%. Table 2 reports the findings from this analysis.

Table 2.

Summary of Multinomial Regression on Probation Outcomes (n = 596).

Model 1: Technical violations	B	SE	Exp (B)
Age	−0.026*	0.016	0.974
Hispanic	−0.146	0.304	0.864
HS graduate	−0.383	0.346	0.682
Married	−0.390	0.330	0.677
Male	0.597	0.483	1.817
Employed	0.070	0.410	1.072
Dependents	0.153*	0.088	1.166
Ignition interlock months	0.012	0.008	1.012
Risk items
Address changes	0.363**	0.137	1.437
Alcohol usage	−0.202	0.286	0.817
Drug usage	0.106	0.250	1.11
Attitude	0.007	0.081	1.007
Age at first adjudication	0.007	0.092	1.007
Prior periods of probation/parole terms	0.003	0.153	1.003
Prior revocations	−0.016	0.077	0.984
Prior felony adjudications	0.044	0.102	1.045
Adult/juvenile adjudications	−0.011	0.131	0.989
Adjudications for assaultive offenses ≥5 years	−0.025	0.062	0.976
Needs
Academic/vocational skills	−0.006	0.138	0.994
Financial management	0.037	0.095	1.038
Family relations	−0.109	0.087	0.896
Companions	0.100	0.122	1.106
Emotional stability	0.130	0.092	1.139
Alcohol use problems	0.105	0.092	1.111
Other drug problems	−0.212	0.136	0.809
Mental ability	−0.031	0.205	0.97
Health	0.258	0.343	1.295
P.O. impression of needs	−0.065	0.103	0.937
Model 2: Subsequent arrest	B	SE	Exp (B)
Age	−0.084**	0.036	0.920
Hispanic	−0.103	0.664	0.902
HS graduate	−0.804	0.703	0.448
Married	−0.616	0.764	0.540
Male	0.830	1.021	2.294
Employed	−1.547**	0.705	0.213
Dependents	0.185	0.191	1.204
Ignition interlock months	0.009	0.018	1.01
Model 2: Subsequent arrest	B	SE	Exp (B)
Risk items
Address changes	0.526*	0.282	1.691
Alcohol usage	0.217	0.684	1.241
Drug usage	0.613	0.415	1.846
Attitude	0.076	0.167	1.079
Age at first adjudication	−0.654**	0.254	0.52
Prior periods of probation/parole terms	0.195	0.323	1.216
Prior revocations	0.008	0.162	1.008
Prior felony adjudications	−0.125	0.221	0.882
Adult/juvenile adjudications	0.562**	0.237	1.753
Adjudications for assaultive offenses ≥5 years	−0.021	0.118	0.979
Needs
Academic/vocational skills	0.153	0.256	1.166
Financial management	0.243	0.226	1.27
Family relations	−0.236	0.185	0.790
Companions	−0.130	0.274	0.878
Emotional stability	0.010	0.194	1.010
Alcohol use problems	0.231	0.215	1.260
Other drug problems	0.139	0.233	1.149
Health	0.478	0.686	1.612
P.O. impression of needs	−0.090	0.210	0.914
Needs
Intercept

Note. The reference category is no violations. SE = standard error.

Nagelkerke Pseudo R (model fit) = .185.

p < .10. **p < .05.

Model 2 estimated the effect of the same set of demographic and probation variables on the odds of a subsequent arrest versus no violation. Two control variables were found to be significant in the model. Specifically, the age of the offender at the beginning of the supervision period significantly affected the odds of rearrest. That is, for every 1-year reduction in age, the odds of rearrest decreased by 8%. Also, individuals who were employed full time were significantly less likely to be rearrested. Specifically, the model shows that full-time employment reduces the odds of rearrest by approximately 79% compared with those who are employed part time or unemployed. With respect to the risk and needs measures, three items emerged as significant predictors of rearrest. Similar to the previous model, the number of address changes significantly affected the odds of rearrest. Furthermore, the analysis showed that as the number of address changes increased from 0 to 1 or 2 or more in the last year, the odds of rearrest increased by roughly 69%. Two criminal history measures also appear significant in the model. Furthermore, as the offender’s age at first adjudication of guilt decreases (19≥; 20–23; 24≤), the offender’s likelihood of rearrest drops by 48%. Moreover, the number of overall adjudications as an adult or juvenile significantly affected the odds of rearrest, with every adjudication leading to a 75% increase in the likelihood of rearrest. Taken together, the models reveal limited, yet important, differences between the two groups of violators.

Discussion

The intent of this study was to better understand how the overall profile of a chronic DWI offender can be used to make assessments about probation outcomes. Although many community correction agencies have transitioned to specialized instruments to assess those convicted of DWI, generalized assessment tools continue to be widely used to assess this offender group. Despite its widespread popularity, few studies have examined the predictive validity of the Wisconsin Risk Need tool, or any substance abuse screening tool, in detecting technical violations as well as rearrests. The current study addresses this gap in literature and examines more closely a set of risk and need measures on probation outcomes in a sample of chronic DWI offenders. It highlights the limitations of the Wisconsin Risk and Needs Assessment tool in discerning outcomes for this population and reveals some insight into the factors that lead to a probationer’s success or failure on probation.

The results of this research shed light on a few areas that provide insight related to supervising chronic DWI offenders. First, the analysis revealed that criminal history variables play an important role in understanding recidivism among chronic DWI offenders. Criminal history variables have long been regarded as a reliable and valid predictor of future offending (Andrews & Bonta, 2010), yet the importance of prior exposure to the criminal justice system has not been identified as a salient factor predicting future arrests for chronic DWI offenders. This research reinforces early findings on problem drinking behavior indicating the importance of this variable as a risk factor to criminal recidivism (Barry et al., 2006). Given the magnitude of the odds ratios for age at first adjudication and overall adjudication, it is clear that contact with the justice system over the life course puts repeat DWI offenders at increased risk of future justice system involvement, despite the range of surveillance and treatment strategies used to supervise these offenders. Second, risk scores are normally considered when tailoring conditions of probation to the offender, yet this research underscores the potential for how such scores can be misused or misinterpreted when developing a supervision strategy for chronic DWI offenders. This is important considering the number of DWI offenders on community supervision and the potential for iatrogenic effects related to the misalignment of treatment strategies with individual risk and needs (Lowenkamp & Latessa, 2002). On the other hand, these results also suggest that early offending has no relationship to the offender’s ability to comply with other conditions of probation. Focusing on key indicators of risk may be a more effective and efficient method for supervising this population, given the large numbers of assessment items on tools such as the LSI-R or T-RAS (Texas Risk Assessment System).

Criminological theory also provides some perspective to understanding chronic DWI offending. Specifically, strain theory and stake in conformity may explain why those with a greater number of dependents or who are unemployed are at risk of failing on probation. General strain theory (GST; Agnew, 1992) posits that the loss of positively valued stimuli or confrontation with negative stimuli may lead to substance abuse as a coping strategy. Previous research has shown that those who experience stress related to family relationships are more at risk for alcohol use and abuse (Brezina, 1996; Gottfredson & Hirschi, 1990). Unlike general substance abuse offenders, chronic DWI offenders tend to have stronger social connections but also experience instability in personal relationships, which may make them prone to noncompliance with probation conditions. Findings from this research suggest that the additional responsibilities of caring for dependents can induce stress that in turn leads to continued alcohol use, financial problems, and other conflicts associated with meeting conditions of probation. Given that felony probation requirements are more stringent than misdemeanor requirements, it may be that treatment strategies should include a stronger emphasis on family therapy to address issues related to parenting and family relationships. Further support for strain theory is reflected in the finding related to the connection between employment and subsequent arrests. When examining the larger picture of the chronic DWI offender, it is clear that unemployment and early arrest history are significant risk factors affecting the likelihood of future arrests but that they do not appear to affect the ability to comply with technical conditions of probation.

The findings from this study may be useful to probation officers when making decisions about how to deal with specific types of violations, determining what types of sanctions or incentives are appropriate, and whether certain resources are necessary to assist the offender in their rehabilitation. A cursory examination of the data showed that in many cases, when a defendant violated a probation condition, the court ultimately reinstated the defendant’s probation. This suggests that minor violations are common, and that probation staff use the revocation process as a “carrot and stick approach” to managing their caseloads. Future research should explore probation officer responses to violations of probation conditions among chronic DWI offenders.

The results of this study support findings from previous work suggesting that alcohol and drug use is not a key factor influencing DWI recidivism. This may be an explanation for why static factors were found to be more salient predictors of DWI recidivism than dynamic factors. In their recent study, DeMichele et al. (2016) found that common risk factors included in the LSI-R and Adult Substance Use Inventory (ASUS) were unrelated to DWI recidivism. Given their results, the authors contend that DWI recidivism is a separate construct from general recidivism and that factors such as early arrest history and general deviance appear to bear a stronger connection to DWI outcomes. In the current study, none of the alcohol or substance abuse measures were related to DWI outcomes.

Contrary to previous research on the predictive validity of risk assessments, this study showed that static factors were more predictive of both technical violations and rearrests compared with dynamic factors. This finding may be explained by research indicating that community supervision officers do not always use risk assessment results to make supervision decisions (Schaefer & Williamson, 2018). Despite widespread support for the RNR model, several individual and institutional factors have been found to affect officer compliance with assessment tools (Andrews et al., 2011; Miller & Maloney, 2013; Viglione et al., 2018). One such factor involves the quality of training provided to raters. Lowenkamp et al. (2004) stress that formal training is required to reach an instrument’s fullest potential and that “bootleg” sessions conducted by in-house or inexperienced staff are inadequate (p. 53). The significance of training was further reinforced in a large scale meta-analysis of validation studies on the LSI risk assessment tool (Andrews et al., 2011). The authors stress the importance of fidelity to implementation protocols to ensure the reliability and validity of results. It is well documented that community corrections staff face challenges managing offender risk assessments, given their heavy caseloads and high-stress levels, which has the potential to undermine the reliability of dynamic measures (Cording et al., 2016). Furthermore, obtaining reliable and valid measures of risk would appear to be more difficult in a large probation department where each officer performs assessments for their respective caseload. It may also be the case that probation officers view chronic DWI offenders as lower risk, given their respective lower rearrest rates. Furthermore, the lack of fidelity to the risk assessment process could explain findings from previous work documenting the unmet treatment needs of DWI recidivists.

There are some limitations to the current study. First, information on the type of violation was not available through the agency’s computer records. It would be helpful to examine the relationship between the same set of variables and violations of different severity. A technical violation can range from failing to pay probation fees to absconding from probation. It may be the case that assessment data are more useful for predicting serious technical violations and that social and demographic factors are more closely linked to less serious violations. Second, the data used in this study did not contain information on the officer’s response to violations or a subsequent arrest. Probation officers have a range of options for dealing with probation violations, which range from filing a motion to revoke probation (MTR) to imposing additional sanctions on the defendant. How a probation officer deals with probation violations could lead to more severe violations or desistance.

An important finding of this research is that practitioners and researchers must understand the practical limits to risk assessment tools as they relate to the chronic DWI offender’s path on community supervision. It further raises the question of whether shorter instruments may be as effective as longer, more specialized instruments and how this information informs a supervision strategy. In terms of their probation outcomes, these data show felony DWI offenders are much more likely to commit a technical violation than be rearrested for a new offense. Identifying the salient factors (i.e., demographic, social, assessment) becomes a critical task from a supervision standpoint.

The challenge facing probation administrators and researchers is to develop a parsimonious model for managing chronic DWI defendants—one that assists probation officers in assessing risk of probation failure, whether that is due to subsequent criminal behavior or compliance with probation conditions (e.g., technical violations). Armed with a more efficient model for understanding chronic DWI offenders probation officers can more effectively respond to the needs of this population.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Durant Frantzen

Notes

References

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47–88.

Andrews

D. A.

Bonta

(2010). The psychology of criminal conduct (5th ed.). Matthew Bender & Company.

Andrews

D. A.

Bonta

Wormith

J. S.

Guzzo

Brews

Rettinger

Rowe

(2011). Sources of variability in estimates of predictive validity: A specification with Level of Service general risk and need. Criminal Justice and Behavior, 38, 413–432. https://doi.org/10.1177/0093854811401990

Baird

Heinz

R. C.

Bemus

B. J.

(1979). The Wisconsin case classification/staff deployment project: Two-year follow-up report. Wisconsin Division of Corrections.

Barry

A. E.

Misra

Dennis

(2006). Assessing driving while intoxicated (DWI) offender characteristics and drinking problems utilizing the numerical drinking profile (NDP). Journal of Alcohol & Drug Education, 50, 66–79.

Belenko

Fabrikant

Wolff

(2011). The long road to treatment: Models of screening and admission into drug courts. Criminal Justice and Behavior, 38, 1222–1243.

Brezina

(1996). Adapting to strain: An examination of delinquent coping responses. Criminology, 34, 39–60.

Buckley

Chapman

R. L. $.

Lewis

(2016). A systematic review of intervening to prevent driving while intoxicated: The problem of driving while intoxicated (DWI). Substance Use and Misuse, 51, 104–112.

Cavaiola

Strohmetz

Wolf

Lavender

(2003). Comparison of DWI offenders with non-DWI individuals on the MMPI-2 and the Michigan Alcoholism Screening Test. Addictive Behaviors, 28, 971–977.

10.

Chang

Gregory

Lapham

S. C.

(2002). Review of screening instruments and procedures for evaluating driving while intoxicated/impaired (DWI) offenders. AAA Foundation for Traffic Safety.

11.

Cording

J. R.

Beggs Christofferson

S. M.

Grace

R. C.

(2016). Challenges for the theory and application of dynamic risk factors. Psychology, Crime & Law, 22, 84–103. https://doi.org/10.1080/1068316x.2015.1111367

12.

Council for State Governments. (2019). Confined and costly: How supervision violations Are filling prisons and burdening budgets. https://csgjusticecenter.org/confinedandcostly/

13.

Curtis

Hoctor

Pennell

(1994). Intensive supervision for drug-involved probationers. In Fields

(Ed.), Innovative trends & specialized strategies in community-based corrections (pp. 87–119). Garland Publishing.

14.

DeMichele

Payne

Lowe

(2016). Identifying the characteristics of chronic drunk drivers using a criminological lens. Crime & Delinquency, 62, 1570–1596.

15.

Donovan

D. M.

Salzberg

P. M.

Chaney

E. F.

Queisser

H. R.

Marlatt

G. A.

(1990). Prevention skills for alcohol-involved drivers. Alcohol, Drugs and Driving, 6, 169–188.

16.

Drake

E. K.

(2011). What works in community supervision (Interim report, Document No. 11-12-1201). Washington State Institute for Public Policy.

17.

Farrington

D. P.

(2003). Developmental and life course criminology: Key theoretical and empirical issues: The 2002 Sutherland award address. Criminology, 41, 221–255. https://doi.org/10.1136/bmj.e4692

18.

Federal Bureau of Investigation. (2016). Crime in the United States, 2013. https://ucr.fbi.gov/crime-in-the-u.s/2016/crime-in-the-u.s.-2016/topic-pages/tables/table-20

19.

Feldstein

S. W.

Miller

W. R.

(2007). Does subtle screening for substance abuse work? A review of the Substance Abuse Subtle Screening Inventory (SASSI). Addiction, 102, 41–50.

20.

Freeman

Maxwell

J. C.

Davey

(2011). Unraveling the complexity of driving while intoxicated: A study into the prevalence of psychiatric and substance abuse comorbidity. Accident Analysis and Prevention, 43, 34–39.

21.

Gottfredson

M. R.

Hirschi

(1990). A general theory of crime. Stanford University Press.

22.

Grommon

Davidson

W. S.

Bynum

(2013). A randomized trial of a multimodal community-based prisoner reentry program emphasizing substance abuse treatment. Journal of Offender Rehabilitation, 52, 287–309.

23.

Henderson

Miller

(2011). The (twice) failure of the Wisconsin risk and need assessment in a sample of probationers. Criminal Justice Policy Review, 24, 199–221.

24.

Jones

Johnson

Latessa

Travis

(1999). Case classification in community corrections: Preliminary findings from a national survey. In Thigpen

(Ed.), Topics in community corrections: Classification & risk assessment (pp. 102–126). National Institute of Justice.

25.

Kenneally

(2016). Legislation to admit evidence of propensity when prosecuting DUI recidivists. Northern Illinois Law Review, 37, 126–136.

26.

Lapham

S. C.

McMillan

G. P.

(2011). Open-label pilot study of extended-release naltrexone to reduce drinking and driving among repeat offenders. Journal of Addiction Medicine, 5, 163–169.

27.

Lee

C. B.

Teske

R. H. C.

(2015). Specific deterrence, community context, and drunk driving: An event history analysis. International Journal of Offender Therapy and Comparative Criminology, 59, 230–258.

28.

Lowenkamp

Latessa

(2002). Evaluation of Ohio’s community based correctional facilities and halfway house programs [Technical Report]. University of Cincinnati.

29.

Lowenkamp

C. T.

Latessa

E. J.

Holsinger

A. M.

(2004). Empirical evidence on the importance of training and experience in using the Level Of Service Inventory-Revised (pp. 49–53). Topics in Community Corrections.

30.

Mertler

C. A.

Vannatta

R. A.

(2010). Advanced and multivariate statistical methods: Practical application and interpretation (4th ed.). Pyrczak.

31.

Miller

Maloney

(2013). Practitioner compliance with risk/needs assessment tools: A theoretical and empirical assessment. Crime & Delinquency, 60, 1235–1257.

32.

Miller

P. G.

Miller Curtis

Sønderlund

A. D.

Droste

(2015). Effectiveness of interventions for convicted DUI offenders in reducing recidivism: A systematic review of the peer-reviewed scientific literature. The American Journal of Drug and Alcohol Use, 41, 16–29.

33.

Mitchell

Wilson

D. B.

Eggers

MacKenzie

(2012). Assessing the effectiveness of drug courts on recidivism: A meta-analytic review of traditional and nontraditional drug courts. Journal of Criminal Justice, 40, 60–71.

34.

Mortimer

R. G.

Filkins

L. D.

Lower

J. S.

(1971). Court procedures for identifying problem drinkers: Phase II (Report No. HSR171-120, HUF-II). U.S. Department of Transportation, University of Michigan Highway Traffic Safety Research Institute.

35.

Mothers Against Drunk Driving. (2015). DUI felony laws. https://www.madd.org/wp-content/uploads/2017/08/DUI_Felony_Overview.pdf

36.

Mullen

Ryan

S. R.

Mathias

C. W.

Doughtery

D. M.

(2015). Treatment needs of driving while intoxicated offenders: The need for a multimodal approach to treatment. Traffic Injury Prevention, 16, 637–644.

37.

Nagin

D. S.

(2005). Group-based modeling of development. Harvard University Press.

38.

National Drug Court Institute. (2004). DWI/Drug courts: Defining a national strategy.

39.

National Highway Traffic Safety Administration. (2009). Guidelines for community supervision of DWI offenders. U.S. Department of Transportation.

40.

National Highway Traffic Safety Administration. (2018). Alcohol-impaired driving: Traffic safety facts, 2017 data. U.S. Department of Transportation.

41.

Nochajski

T. H.

Stasiewicz

P. R.

Patterson

D. A.

(2013). Depression, readiness for change, and treatment among court-mandated DUI offenders. Journal of Dual Diagnosis, 9, 139–148.

42.

Ouimet

M. C.

Dongier

Di Leo

Legault

Tremblay

Brown

T. G.

(2013). A randomized controlled trial of brief motivational interviewing in impaired driving recidivists: A 5-year follow-up of traffic offenses and crashes. Alcoholism, Clinical and Experimental Research, 37, 1979–1985.

43.

Schaefer

Williamson

(2018). Probation and parole officers’ compliance with case management tools: Professional discretion and override. International Journal of Offender Therapy and Comparative Criminology, 62, 4565–4584.

44.

Steiner

Makarios

Travis

L. F.

Meade

(2012). Examining the effects of community-based sanctions on offender recidivism. Justice Quarterly, 29, 229–257.

45.

Taxman

F. S.

Belenko

(2012). Implementing evidence-based practices in community corrections and addiction treatment. Springer.

46.

Viglione

J. V.

Blasko

Taxman

F. S.

(2018). Organizational factors and probation officer use of evidence-based practices: A multilevel examination. International Journal of Offender Therapy and Comparative Criminology, 62, 1648–1667.

47.

Wells-Parker

Bangert-Drowns

McMillen

Williams

(1995). Final results from a meta-analysis of remedial interventions with drink/drive offenders. Addiction, 90, 907–926.

48.

Wodahl

E. J.

Boman

Garland

B. E.

(2015). Responding to probation and parole violations: Are jail sanctions more effective than community-based graduated sanctions? Journal of Criminal Justice, 43, 242–250.

49.

Wormith

J. S.

Zidenberg

A. M.

(2018). The historical roots, current status, and future applications of the risk-need-responsivity model (RNR). In Jeglic

E. L.

Calkins

(Eds.), New Frontiers in offender treatment the translation of evidence-based practices to correctional settings (pp. 11–41). Springer. https://doi.org/10.1007/978-3-030-01030-0_2

50.

Wright

Clear

Dickson

(1984). Universal applicability of probation risk-assessment instruments. Criminology, 22, 113–134.