/’1/\/\(/’\ (7. ’l, «I, “ ||l|||l||||llIIIHIIHHIIIIIWHUI!IIIHIIIIHIIHIHII 31293 02060 4017 LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE ~4we+3fimn LUV! golfiflz DEC 1 0 2003 11/00 chiRC/DateDuepSS—p.“ EXAMINING THE DETERRENT EFFECTS OF MULTIPLE CONDITIONS TO PROBATION ON RECIDIVISM IN MICHIGAN BY Omara Rivera — Vézquez A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTERS OF SCIENCE School of Criminal Justice — Urban Affairs 1999 ABSTRACT EXAMINING THE DETERRENT EFFECTS OF MULTIPLE CONDITIONS TO PROBATION ON RECIDIVISM IN MICHIGAN BY Omara Rivera-Vézquez Despite its rapid growth and popularity, probation stills needs to circumvent several difficulties mostly dealing with recidivism. The present study examines the effects of multiple conditions of probation and its overall impact on recidivism by probationers. Based on the theory that sanctions can be used to modify certain behaviors, this study examines the effects of multiple conditions of probation on the behaviors of probationers. Following the deterrence framework, the more conditions the offender is presented with, the higher should be their compliance and the less their probation recidivism. This research also examines the relationship of offenders’ criminal histories and socio-demographic characteristics on probationers recidivism. Using logistic regression, two models were created to assess the effects of multiple conditions of probation, criminal histories and demographics on the dependent measure of discharge type. The results suggest a negative relationship between multiple conditions to probation and discharge type. Examination of demographic and criminal history measures demonstrated that they influence the performance of offenders placed on probation. ACKNOWLEDMENTS I would like to thank a number of people that in one form or another helped me completing this thesis. First, I thank my thesis advisor and committee members for their tremendous patience and support through this year. I especially thank Dr. Sheila Maxwell for her patience in going through several drafts of this document and her wisdom when making comments and/or suggestions. In addition, I thank Dr. Maxwell for allowing me access to the probation recidivism project database and for taking time to answer my questions about the data. I thank Dr. John Schweitzer for his valuable suggestions and for literally sitting down with me to explain to me some statistical techniques. I am also grateful to Dr. Dorothy Harper Jones in the Graduate School, for all the times that she served as my counselor when l was so worried about not finishing my thesis. I thank my mother Adroberta Vézquez and my sister Mercedes A. Rivera for their strong support and invaluable prayers. To Marni and Chedy for their love and dedication, my thanks to you are never enough, and to you, I dedicate this thesis. iii TABLE OF CONTENTS Abstract Acknowledgments Table of Contents List of Tables Table 5.1. Demographics and criminal History Measures by Discharge Type Table 5.2. Conditions of Probation by Discharge Type Table 5.3. Predictors of Discharge Type Using Logistic Regression List of Figures Figure 1. Effects of Probation Conditions on Successful Discharges I. Introduction Purpose Relevance II. Review of Related Literature Deterrence Theory Sanctions, Specific Deterrence, and Recidivism Individual Characteristics and Recidivism Conditions of Probation and Probation Recidivism Probation in the State of Michigan Ill. Assessing Hypotheses Research Questions Hypotheses IV. Methodology Sample Data Collection Instrument Description of Measures Demographic Variables Criminal History Variables Multiple Conditions of Probation Probation Recidivism iv ii iii iv-v 38 40 54 61 1-4 4-6 6-7 8-28 8- 13 14- 18 18-21 22-27 27-28 29-3 1 29 29-3 1 32-36 32 32-33 33-34 34 34 35 35 36 V. Results Probationer Characteristics Demographic ad Criminal History Characteristics Conditions of Probation Discharge Type Bivariate Analysis Demographic Characteristics by Discharge Type Relationships Criminal Histories by Discharge Type Relationships Conditions of Probation by Discharge Type Relationships Regression Models 48-50 Effects of Demographic Measures on Discharge Type Effects of Probationers’ Criminal Histories on Discharge Type Effects of Conditions to Probation on Discharge Type VI. Discussion Implications for Future Research Limitations of the Study Appendices 37-53 37 37 39 41 42-47 42-43 43-44 52-53 5 0-5 1 5 1 -52 52-53 55-5 7 5 7-58 58-59 Appendix A. Effects of Probation Conditions on Successful Discharges 61 References 62-68 CHAPTERI INTRODUCTION EXAMINING THE DETERRENT EFFECTS OF MULTIPLE CONDITIONS TO PROBATION ON RECIDIVISM IN MICHIGAN Probation has long been one of the most popular forms of correctional disposition. It is the most widely used form of corrections (McCarthy and McCarthy, 1997, Petersilia, 1997). In 1995, the Bureau of Justice Statistics (BJS) reported that almost two-thirds of all convicted offenders served their sentences on probation. As a sentencing alternative, probation is expected to control the offender in the community while complying with the sanctions imposed by the court. Probation is intended to control offenders in the community by placing them under supervision and providing treatment in order to promote law abiding behavior (McCarthy and McCarthy, 1997). Similar to other forms of correctional alternatives, in the last decade, probation has experienced a tremendous increase in the number of offenders sentenced (BJS, 1995). Moreover, those sentenced to probation changed from primarily misdemeanor offenders to felony offenders (Whitehead. 1991; Vito, 1986; BJS, 1995). Probation is a judicially imposed criminal sanction permitting court supervision of the offender within the community (American Bar Association, 1994). Several years ago its primary purpose was the reintegration of the offender into the community. Currently, community protection has become probation’s principal objective (Tonry and Hamilton; 1994, Petersilia, 1997; McCarthy and McCarthy, 1997). In order to reach the goal of community protection, probation focuses on supervision by monitoring probationers’ behavior and removing them from the community when they violate the conditions of their probation. y More than three million adult offenders are currently being supervised in the community on probation and similar types of supervised release (Taxman and Cherkos, 1995). Presently, mostly due to overcrowding in jails and prisons- } community sanctions have been pronounced as the “best remedy" to an ongoing . crisis (Tonry and Hamilton, 1995). Consequently, this movement to alleviate? institutional crowding has resulted in the sentencing of more offenders to? probation, which is now considered the primary alternative to incarceration Z i (Petersilia, 1997). Probation has experienced explosive growth in the last; decade (McCarthy and McCarthy, 1997; Petersilia, 1997). Since 1985, the Nation’s probation population has grown an average of 3 percent per year. Remarkably, probationers accounted for the largest portion of adults under correctional supervision (58%), including persons held in jails and prisons and those on parole. Despite this rapid growth and its given popularity, there still are many difficulties associated with probation, mostly dealing with recidivism. Recidivism as an outcome measure has provided particular criticisms regarding probation. Specifically, it has been argued that probation recidivism may have the unexpected effect on prison overcrowding (Petersilia, 1997; McCarthy and McCarthy, 1997). Recent figures tend to confirm that instead of reducing overcrowding in prison, probationers who recidivate are exacerbating the problem. In 1991 the Bureau of Justice Statistics published a survey of \ I I. I r probation and parole violators in state prison (BJS, 1991). This report clearly highlighted that in 1991, 45 percent of State prisoners were individuals who, at the time they committed their offense, were under community supervision-either on probation or on parole. Moreover, thirty-five percent of State prison inmates in 1991 were convicted of a new offense that they had committed while they were on probation or parole from a previous sentence; 10 percent of the inmates had been returned to prison for technically violating the conditions of their probation or parole terms. Current literature indicates that researchers exploring probation recidivism have considered @fissues when studying this issue. Existing literature mainly refer to offenders’ individual characteristics and the effects of conditions of probation. Studies focusing on individual characteristics have identified a number of variables that correlate well with success or failure on probation. These variables include age (younger offenders have greater failure rates); employment (those employed and financially stable do better); marital status (married individuals are more likely to succeed); and type of offense (those on probation for theft-related crimes have higher recidivism rates) (Morgan, 1994). One important and many times overlooked issue when studying probation recidivism is the increased used of a series of conditions that are more and more often imposed on probationers. There are not many studies on probation recidivism that have devoted particular attention to these conditions attached to probation orders. Interestingly, studies focusing on such conditions to probation have ascertained that their imposition may lead to higher recidivism rates among probationers. Over the past decade, the number of offenders who had special conditions attached to probation has increased (Taxman and Cherkos, 1995, Petersilia, 1997). Probation conditions often include drug and alcohol testing, curfews, house arrest, and mandatory employment and treatments (Petersilia, 1997). The increased use of these conditions is not actually supported by research outcomes, but on the logic that they will help reduce probation recidivism. In other words, it is expected that an increase in the number of conditions imposed on probationers will have a specific deterrent effect on recidivism. The analysis of this argument is the gist of this thesis. Using the deterrence theory as a framework, the researcher aims to test the deterrent effects of “multiple conditions” and how they relate to recidivism. Purgose of the Study There is an obvious similarity in the use of sanctions to deter criminal conduct, as well as the use of conditions to probation, as analyzed here, to enforce compliance with probation orders. In both cases, sanctions are used in order to help individuals exercise desirable forms of behavior, or to prevent the commission of undesirable conducts. This classic theory of prevention is often characterized as deterrence (Maxwell, 1994). Orland (1973) described deterrence as the inhibiting effect that punishment, either actual or threatened, will have on the actions of those who are otherwise disposed to commit crimes or are disposed to violate a legally prescribed conduct. Consequently, the present study examines the conditions that are given to offenders who were sentenced to probation in Michigan. The key features of probation orders, and those examined in this study, often include any combination of the following: curfews with or without electronic monitoring, special conditions established by the judge (i.e., employment, and counseling), drug and alcohol monitoring, community service, probation fees, split sentences, and restitution. These conditions are hereafter referred as “multiple sanctions”. This study examines multiple conditions of probation and the overall impact on recidivism by probationers. Therefore, the study examines the theory that sanctions can be used to modify certain types of behavior, and the effects of multiple of sanctions on these behaviors. It is expected that this research will have important implications for criminal justice policy as it relates to community corrections. Especially given the results of previous studies that showed positive links between apparent severity of sanctions and reduced recidivism. The findings that more severe sanctions increases offenders’ compliance gives support to the widely held assumption that behavior can indeed be modified or influenced by sanctions. It is expected that the harsher the sanction the greater the degree of compliance. Given the utilitarian support for the use of harsher sanctions in enforcing probationers compliance, this study analyzes several variables previously referenced in deterrence research and probation recidivism studies as important factors in understanding the relative effects of sanctions on behavior. In this study, “multiple sanctions” are analyzed by counting the total number of conditions given to probationers. Following the deterrence theoretical framework, the more conditions the offender is presented with, the more compliance and the less the recidivism. The study also examines the relationship between successful discharges and the offenders’ criminal histories including prior offenses/convictions, juvenile records, and type of offenses. In addition, the variable of prior drug use is included for analysis. These variables are generally found as predictors in recidivism literature (Morgan, 1994; Petersilia, 1997 & 1986) and are, therefore, included in this analysis. Other variables examined are the basic social and demographic characteristics of the probationers, such as age and gender. The relationships among these variables and multiplicity of sanctions is examined as well as their effects on future criminal behavior. Relevance Government policies to reduce prison overcrowding in the United States have recently looked towards implementing sanctions that provide a range of community-based alternatives that focus on the surveillance and control of offenders (Tonry and Hamilton, 1994; Morris and Tonry, 1990, Byme, Lurigio, and Petersilia, 1992; Abadinski, 1991). Although numerous evaluative studies have been published concerning probation, we still need more evidence regarding the increasing use of conditions of probation and their effects on offender recidivism. By clarifying this issue, the present study seeks to build on these theoretical aspects as well as to inform policy makers on specific links between sanctions and behavior that affect criminal justice policies in the United States, specifically those regarding probation. Furthermore, this study provides an opportunity for added empirical examination on the theory of deterrence. Whether or not multiple conditions to probation are met and probationers’ refrain from committing further criminal acts is achieved, is an issue that offers a unique perspective to analyzing the deterrence theory. CHAPTER II LITERATURE REVIEW Theoretical and empirical bases for this study and the rationales for selection of variables are provided in this chapter. Deterrence Theog The punishment of offenders in the United States is based on several ideological perspectives including justice, retribution and deterrence (Brooker, 1971). In this study, the focus will be on the utilitarian aspects of punishment, more generally known as deterrence. There are two broad categories of deterrence; these include general and specific deterrence. This distinction between general and specific deterrence is an important one since it recognizes the difference between two categories of individuals: those who have suffered a punishment for having committed a crime and those who have not (Gibbs, 1975). Ball (1955) stated that deterrence is usually defined as the preventive effect which actual or threatened punishment of offenders has upon potential offenders. On the other hand, specific deterrence has been defined as “the omission or curtailment of some type of criminal activity by an individual throughout a period of time because in whole or part he or she has been accused of a crime for which someone was punished, and he or she is therefore unwilling to risk someone being punished again” (Gibbs, 1975). Simply stated specific deterrence refers to the preventive effect that punishment has on an individual who is being punished (Pontell, 1984). The classical school of criminal law taught that deviant behavior can be deterred by punishment or the fear of punishment (Roshier, 1987; Boostrom, 1995; Ball 1955; Claster, 1967, Abadinski, 1994). The positivist school restricts the scope of this position by rejecting the classical postulate that deterrence is equally applicable to all individuals (Degler, 1991; Boostrom, 1995; Ball 1956; Claster, 1967, Abadinski, 1994). Focusing on differing tendencies toward criminal behavior, positivism considers deterrence in view of variable factors, factors that make some people less susceptible to threatened sanctions than others (Claster, 1967) From these perspectives we then find numerous studies dealing with the nature and characteristics of the deterrence doctrine. The main focus of these studies has been the proposition that when certain, swift, and severe punishments are administered, the costs of committing crime are increased, which should lead to a subsequent reduction in criminal behavior (Pontell, 1984). There is a vast quantity of literature dealing with the effectiveness of punishment as a deterrent (Gibbs, 1968: Chambliss, 1966; Tittle, 1969, 1980, Fagan, 1973, 1994; Logan, 1972). The deterrence theory mainly focuses on three aspects: celerity, certainty, and severity. The theory emphasizes that fast, certain, and harsh punishment discourages punished individuals from violating again (specific deterrence) and dissuades the general public, which learns that there is punishment from committing crimes (general deterrence) (Liska, 1987). Basically, the deterrence theory simply asserts that people fear punishment (Yu, 1994) The deterrence theory considers an inverse relationship between the certainty of punishment of a type of crime and the crime rate (Gibbs, 1975; Logan, 1972). On the other hand, a deterrent efficacy of punishments depends on their severity, while celerity considers suppositions that immediate punishment is more dreaded than delayed punishment (Gibbs, 1975). Existing analyses of the deterrence theory have not yet clearly determined support or non-support for this theory. Some studies report inverse relationships between sanctions and crime commission (Gibbs, 1968, Tittle, 1969), while others report weak to non-existent relationships (Fagan, 1973). Fagan (1994) conducted a study of criminal sanctions for drug offenders in the state of New York and found no relationship between sanction severity and recidivism. Furthermore, Tittle and Logan (1973) were only able to conclude that: “sanctions have some deterrent effect under some circumstances.” Even when evidence on the deterrent effect of sanctions on criminal behavior is inconclusive, the tendency has been to continue legislation for harsher and heavier penalties (Maxwell. 1994). The fundamental principle here is consistent with the deterrence theory, a utilitarian view that points that offenders calculate the risks and rewards of their actions (Maxwell, 1994, van den Haag, 1975). Despite the well-known evidence that deterrence requires some combination of certainty, celerity, and severity, significant differences exist in the way they have been operationalized (Decker and Kohfeld, 1990). Additionally, 10 as Decker and Kohfeld (1990) stated, several different approaches are utilized in attempting to assess the effects of deterrent measures. The first approach to measuring deterrence studies the effects of sanctions on offenses. From this position, when an increment in sanctions is experienced and reductions in measures of crime result, a deterrent effect is said to have taken place. Hence it is assumed that the activities of the criminal justice system, usually apprehension or sentencing, produce a corresponding crime reduction (Decker and Kohfeld, 1990). A second approach analyzes the impact of resources on crime levels. These examine the effect of additional resources either personal or monetary on the number of offenses. A third perspective to measuring deterrence has been to test the effect of hypothetical changes in risk and severity of punishments on perceptions. This approach has considered the levels at which individuals perceive risks associated with offending to be important. The focus on the present study follows the first approach; thus, the literature review is centered on issues relevant to that focal point. Certainty and severity have been measured in several ways. Certainty has constantly been defined as the probability of the application of some sanction (Decker and Kohfeld, 1990). Many studies have examined this issue in order to find out if police activities produced a deterrent effect on offending rates. Such efforts often analyzed the ratio of clearances to the number of reported offenses for given jurisdictions. On the other side, many studies have used imprisonment as a measure of certainty. This is usually conceptualized as the probability of imprisonment to arrests or reported offenses. Measures of severity have traditionally focused on the length of the sentence imposed for any particular offense. In each of these approaches, an inverse relationship is considered evidence of a deterrent effect. The results of all these investigations have generally been consistent. Numerous studies have demonstrated that measures of certainty are inversely related to the levels of offenses (Decker and Kohfeld, 1990). These studies often reveal that the effects of certainty and severity of punishment on crime rates is weak, although significant. Thus there is a negative association between certainty of punishment and crime rates, and no association between severity of punishment and crime rates except for homicide (Chambliss, 1966; Gibbs, 1968; Jensen, 1969; Tittle, 1969; Chiricos and Waldo, 1972; Tittle and Rowe, 1974). Furthermore, studies have analyzed the interaction between the degree of certainty or severity of punishment and the deterrent effect of the threat of punishment. For example, Tittle and Rowe (1974) demonstrated that the deterrent effect of the certainty of arrest is a function of the degree of certainty itself, meaning that there is a critical level for certainty of punishment (30% clearance by arrest) and the implied perceived threat of punishment as indicated by a threshold level below which arrest is relatively unlikely to occur. Tittle’s earlier study (1969) reported that the impact of certainty of punishment is independent of severity, while severity of punishment is related to lower crime rates for varying levels of certainty of punishment (Silberrnan, 1976). Studies on severity have produced the most striking results. Contrary to the deterrence perspective, it has been found that increases in severity of 12 punishment do not necessarily lead to less crime commissions. Gibbs (1968) found that the deterrent effect of severity was minimal. Moreover, a study by Chiricos and Waldo (1970) found little support for the hypothesis that the length of prison sentences is negatively correlated with crime rates. The authors concluded that there was no empirical evidence to support the theory of deterrence. Furthermore, Logan (1972) found no significant correlation between severity of prison sentences and crime rates, and Schwartz (1968) analysis of rape data in Philadelphia before and after increased penalties, provided no basis for concluding that increased severity of sanctions significantly affected the amount of rape. There are many studies documenting the increasing rates of crime; similarly, there is considerable evidence documenting the high rates of recidivism. This outlook serves as a stimulus to analyze the deterrent measures that are currently in practice and to provide empirical evidence for potential systems of deterrence for the future. The present study focuses on probation outcomes. Specifically, it examines the effects of the number of probation conditions and their effects on recidivism if any. Evidence that sanctions may backfire and lead offenders to more serious or frequent offending (Bridges and Stone, 1986; Farrington, Ohlin, and Wilson,1986; Petersillia and Turner, 1986; Sherman, Gartin, Doig, and Miller, 1986) has important implications for understanding the impact of sanctions on probation recidivism. 13 SanctionsI Sgcific DeterrenceI and Recidivism Specific deterrence refers to the effect of the imposition of sanctions on the subsequent behavior of the individual punished. When punished, offenders will desist from offending, commit less serious offenses, or offend at lower rates because of the fear of some future sanction (Paternoster & Piquero, 1995). Traditionally, specific deterrence has been considered to be relevant for individuals who have committed and been punished for offenses. Therefore, specific deterrence concems personal and/or direct experiences with punishment. Among those with lengthier criminal histories, deterrence is primarily due to their personal experiences with punishment and punishment avoidance (specific deterrence). Specific deterrence will then occur directly through perceptions about one’s risk, and directly from one’s perceptions about others’ risk of being punished. Such personal experiences as being closely supervised or monitored by others may convince individuals that they are unable to commit crimes without detection. If this belief prevents further criminal activity by strengthening the individual’s perceptions of risk, then this would be noted as a part of specific deterrence (Paternoster and Piquero, 1995). In their reconceptualization of deterrence, Stafford and Warr (1993) emphasized that the direct and indirect experiences of punishment for offending are expected to decrease the individuals’ inclination to offend by increasing the perceived risk of future sanctions. However, Paternoster and Piquero (1995) were unable to prove this true for substance abuse offenders. It was expected that the experience of being apprehended and sanctioned would have a negative 14 effect on subsequent alcohol and marihuana use. They found, instead, a positive effect. Those who reported being sanctioned for their acts at some point in the previous year were more likely to drink liquor and use marihuana in the following year than those not punished. Moreover, after controlling for several variables, these researchers considered this to be a “genuine positive effect of punishment on future substance abuse” (Paternoster and Piquero, 1995). Recent evidence tends to suggest that specific deterrence may not be as strong as hypothetically expected. This finding is consistent with the defiance effect of sanctions (Sherman, 1993). For Sherman, when sanctions are imposed on those who think they have been unfairly imposed, the reaction may be a defiant rather than a repentant one. The perceived unfairness may be due to substantive as well as procedural elements of the imposed punishment. The effect is that the individual being punished experiences shame and anger and reacts to the imposed sanction by defiance, meaning further law breaking. Widely varying results across sanction studies raise important questions regarding the condition(s) by which sanctions reduce, increase, or have no effect on future crime commission. Answering that question is one of the main objectives of the present study. An important fact to state at this point is that similar criminal sanctions have opposite or different effects in different social settings, on different type of offenders and offenses, and at different levels of analysis (Sherman, 1993; Gibbs, 1975, Wilson, 1983; Severy and Whitaker, 1982; Anderson, 1978; van Andel, 1989; Tyler, 1990; Kinsey, 1992; Klein, 1986; Gold and Williams, 1970; Smith and Gartin, 1989; Scheneider, 1990, Wolfgang, 15 Figlio, and Sellin, 1972; Farrington, 1977; Falkowski, 1984; Tyler, 1990; Braithwaite, 1985). When explaining differences in sanctioning effects the following issues have showed some effects: individual differences, social settings, and offense types (Sherman, 1993). Much of the differences observed lack any specific pattern, but there are two main arguments through these pattems. The first pattern is that violations in expectations of fairness as a predictor that sanctions will increase crime rates, both individually and generally. The other pattern implies that sanctions are more likely to increase or fail to deter crime among out-groups even while they deter in-groups, at both individual and general levels (Sherman. 1993). Studies of recidivism using individual-level data have been used to examine the specific deterrent effects of strategies for reducing recidivism rates of juvenile or adult offenders. In these studies, punishment-oriented programs are contrasted with treatment or diversion approaches, using recidivism rates or other indicators of program success. Direct comparisons of recidivism rates of individuals who have been incarcerated and those that were placed on probation or diverted are limited and usually characterized by ample differences in the seriousness of the offenders in the different programs. Since incarcerated populations are generally more serious offenders, they can commonly be expected to have higher recidivism rates after release (Schneider, 1990). The most cited review of probation and other rehabilitation and/or treatment-oriented programs, conducted by Martinson 16 (1974) concluded that “nothing works”. Moreover, a study reported by Palmer (1978) found that the less punishment-oriented programs were more effective for first-time offenders than the more coercive programs. Furthermore, another analysis showed that naive offenders were more influenced by prior experiences with punishment, while experienced offenders were not (Bridges and Stone, 1986). In fact, those who had more experiences with punishment tended to have more tendencies to commit crimes, which in turn was associated with lower perceptions of the threat of punishment. Bridges and Stone (1986) concluded that for experienced offenders, punishment may be expected to have little or no deterrent effect on criminal recidivism since it has limited impact on the perceived threat of punishment. They also found that among offenders with long crime histories, apprehension and punishment might be viewed as hazards that are inconsequential when compared with the potential rewards of crime. Schneider (1990) reported that deterrence variables are unrelated to reduced reoffending in the expected direction. Certainty of punishment was negatively related to subsequent crime. Those who believed that they were more likely to be caught committed more subsequent offenses (Schneider, 1990). Perceived severity of punishment was also related to recidivism in the inverse direction. Higher perceptions of severity were related to greater offending. Clear and Harris (1992) found no indication that the application of severe sanctions (excluding revocation of probation) would be more effective at discouraging chronic violators from committing further serious criminal acts than would be the application of some lenient sanctions. In this study, probationers committing second violations were not less likely to commit major violent offenses following a high severity response from an officer than they were when the officer had undertaken a lenient response. Additional analysis of this study indicated that sanction severity had an effect on the prevalence of new violations than it may had on violation type. Revocation was associated with the lowest increase in new violations, however the authors stated that this finding might be due to the fact that some individuals detained, awaiting revocation, were never returned to the community. Individual Characteristics and Probation Recidivism The Bureau of Justice Statistics (1995) reported that on December 31, 1996, State and local probation agencies supervised more than 3 million adults. Studies on probation have mainly focused on the characteristics of offenders who recidivate from probation. In general, these studies have found that there is significant evidence about the relationship between recidivism and the following offender characteristics: previous criminal history, youth, status other than married, unemployment, low income, education below fourth grade, abuse of alcohol or drugs, and property offenders. Previous criminal history was most frequently found to be a significant factor influencing recidivism. This finding supports previous reports that persons with several prior felony convictions were less successful on probation than other offenders. 18 Clarke et al. (1988) demonstrated that there were five factors that significantly increased recidivism rates among felony probationers in North Carolina; youth, minority status, male, drug problems, and having more previous fingerprinted arrests than their counterparts. Being "divorced or separated" compared to "never married" was also marginally significant. During the 36- month follow up period between 1982 and 1985, 33 percent of the 21,789 felony probationers were rearrested at least once for a new crime (Clarke et al. 1988). Several researchers have investigated risk factors most predictive of adult offender recidivism. For instance, Andrews and Bonta (1994) identified two categories of risk factors: static and dynamic. Static factors (i.e., age, previous convictions) are aspects of the offender’s past that are predictive of recidivism but cannot be changed. Dynamic risk factors, commonly refer to as criminogenic needs (e.g., antisocial cognition, values, and behaviors), are changeable and thus serve as the appropriate targets for treatment (Andrews, Bonta, and Hoge, 1990) Additionally, various studies have examined probationers’ backgrounds and criminal histories in order to identify those characteristics that are associated with recidivism (Petersillia et al., 1985. Petersillia and Turner, 1993; Langan, 1994). There is no disagreement in the criminological literature about some of the predictors of offender recidivism, such as age, gender, past criminal history, early family factors, and criminal associates. Morgan (1994) summarized previous findings as follows: type of crime conviction and extent of prior criminal record: offenders with more previous convictions and property offenders had 19 higher rates of recidivism; income at arrest: higher unemployment/lower income are linked to higher recidivism; household composition: individuals living with spouse, children, or both have lower recidivism; age: younger offenders have higher recidivism rates than older offenders; drug use: probationers who used heroine had higher recidivism rates (Petersillia, 1997). Various studies showed that age at first offense is directly linked to recidivism. In these studies, younger offenders were found to have higher rates of recidivism (Dembo, Washbum, Wish, Schmeidler, Getreu, Berry, Williams, and Blount, 1991; Ganzer & Sarason, 1973; Hanson, Henggeler, Haefele, and Rodick, 1984; Wierson & Forehand, 1995). Contradictory results were reported by Niarhos and Routh (1992) who indicated that recidivism was not predicted by age at first arrest. As Myner (1998) explained, this discrepancy can be resolved in part by the length of the follow-up period. Niarhos and Routh (1992) followed their participants for only one year, whereas those who found age to be a significant predictor followed participants for considerably longer periods, ranging from 20 months to 5 years. Furthermore, Tolanand and Lorion (1988) also found age to be significant. As in Niarhos and Routh’s(1992) study, these researchers used a 1-year follow- up period. However, they used self-report data rather than actual arrest records to measure reoffending. Therefore, the methodology of studies in which age was significant allowed for a greater number of offenses to be committed, due to the longer follow-up period, or more offenses reported through self-reports. Such 20 measures provide a greater count of recidivism rates, which increases the chance of achieving a significant relationship between variables. Recidivism literature has also reported contradictory results regarding substance abuse. Several studies showed that abusing alcohol and/or other drugs predicts greater reoffending (Dembo et al., 1987; Loeber & Dishion, 1983; Niarhos& Routh, 1992), while on the contrary others indicated that substance abusers are less likely to reoffend (Wierson & Forehand, 1995; Wooldredge, Hartman, Latessa, & Holmes, 1994). Substance abusers in the sample studied by Wierson and Forehand (1995) received drug and alcohol treatment while in prison. This treatment targeted and reduced substance abuse, a primary contributor to criminal behavior, and consequently recidivism rates. Therefore, inconsistencies among the research are most likely a function of whether substance abusers in the samples received treatments. Crime related variables as well have been linked to recidivism in previous research. For instance, type of offense was shown to be a strong predictor of reoffending. Recidivists are more likely to have committed property offenses than non-recidivists (Craig & Budd, 1967; Loeber & Dishion, 1983; Moore, Chamberlain, & Mukai, 1979). Despite this finding, Ganzer and Sarason in 1973 indicated that type of offense committed does not differentiate between recidivists and nonrecidivists. A goal of the present study is to investigate the relationships among the previously discussed individual-level variables that have been found to be predictors of probation recidivism. 21 Conditions of Probation and Probation Recidivism Probation is being used as the principal alternative to incarceration in the United States. For offenders on probation, the court decides what conditions will be included in the probation contract between the offenders and the court. Typically, when sentencing an offender to probation, judges often combine the probation term with a split sentence, under which the judge sentences a defendant to prison or jail and then suspends the sentence in favor of probation (Petersillia, 1997). It is assumed that offenders will comply with the conditions of probation by knowing what to expect if they fail to do so. Petersillia (1997) outlined three realms in which these conditions may fall. These realms are referred to as standard conditions or those imposed in all probationers (i.e. reporting to the probation office, notifying the agency of any address changes, remain employed, and not leave the jurisdiction without permission); punitive conditions. usually reflect the seriousness of the offense and increase the intrusiveness and burden of probation (i.e. fines, community service, house arrest, and drug and/or alcohol testing); and treatment conditions which are imposed to force probationers to face with their problems and needs, (i.e. substance abuse, and vocational training). The conditions to probation establish a legal contract between the offender and the court. If defendants violate a probation condition at any time prior to the expiration of the term, the court may, after a hearing, continue the probationer on the program, with or without extending the term or modifying the 22 conditions, or revoke probation and impose other sanctions such as jail and prison terms (McCarthy and McCarthy, 1997; Petersilia, 1997). The constant crowding of prisons continues to have correctional administrators searching for alternatives to incarceration. However, in spite of the widespread use of alternative to incarceration programs, prison population have continued to rise. Intermediate sanctions focusing on supervision have actually exacerbated prison crowding due to increased returns to incarceration for violations of conditions of release (Tonry, 1990; Pearson 1988; Erwin 1986). Over the years, the proportion of probationers subject to special conditions has increased (Clear, 1994). The public’s more punitive attitude contribute to this trend. In accordance with the deterrence theory, both the public and the criminal justice system perceive that harsher penalties will produce higher compliance levels for offenders serving probation terms (Petersillia, 1997; Clear, 1994; Byme, Lurigio and Petersillia, 1992). However, existing data on adult probationers in the United States have shown that research on probation has provided mixed results. The actual trend illustrates that more severe conditions increase the chances of failure (Petersillia and Turner, 1993). According to BJS, the number of probationers successfully completing their probation terms is falling. In 1986, 74% probationers finished their probation terms successfully, in 1992 the figure dropped to 67%, and in 1994 it had dropped to 60% (Langan, 1996). In 1992, Langan and Cuniff studied felons on probation and found that 55% of the offenders had some special condition added to their probation terms, 23 the most common being drug testing. Further analysis of this data by Langan (1994) showed that many probationers failed to satisfy their probation-ordered conditions (Petersillia, 1997). This study found that the majority of probationers simply did not comply with the terms of their probation, and only half of known violators went to jail or prison for their non-compliance. Langan (1994) then concluded that “sanctions are not vigorously enforced.” Petersilia et al. (1985) reported that approximately two thirds of a sample of felony probationers in California were rearrested during a 40-month follow-up period. In 1986, the Bureau of Justice Statistics found that 43 percent of a large sample of felons sentenced to probation were rearrested within three years on a new felony charge. Approximately one-half of these arrests were for violent crimes or serious drug offenses (BJS, 1986). The number of offenders who have special conditions attached to their probation orders has increased. Judges are using split sentences more often and are setting more conditions to probation orders including drug testing, curfews, house arrest, and mandatory employment and treatment (Taxman and Cherkos, 1995). The change in the nature of correctional services has affected the performance of offenders in the system. Actually, several studies ascertain that more severe conditions increase the chances for failure (Taxman and Cherkos, 1995). According to the Bureau of Justice Statistics, a lower percentage of offenders are successfully completing their probation orders. Additionally, numerous offenders who are discharged as “successful” completions are 24 committing serious technical violations during their probation terms. The results of a two-year follow-up study by BJS revealed that out of 3,000 convicted felons placed on probation 38% made no progress or failed to satisfy their court order. Of those individuals required to be tested for drugs, 31% made no progress or failed to satisfy their court order. In additions, a high proportion of those required to perform community service, 37% made no progress or failed to satisfy their court order; and of those receiving fines, 26% did not pay or paid slightly under one-third of their total fees (BJS, 1990). Another study by The National Association of Criminal Justice Planners found that thirty-one percent of probationers who completed their sentence did not fully satisfy the conditions of their probation. Of those 21% did not satisfy their conditions, and 10% only partially satisfied their conditions. Studies indicate that more than 50% of the offenders are not in compliance with their court orders during probation. This study also found that 62% of the offenders were involved in at least one disciplinary hearing for another felony arrest or a technical violation. Furthermore, a study by Taxman and Byme found that offenders were typically reinstated to probation even when they have absconded or been arrested again for criminal activity. Landis, Mercer, and Wolff (1969) reported that jail as a condition of probation appeared to increase the likelihood of failure, though on this variable the differences are small. However, this changes between the successes and failures when restitution is ordered as a condition of probation. Twice as many 25 failures had restitution ordered as part of their probation terms compared to those who did not have restitution as a condition in their probation orders. With fewer probationers successfully completing their sentences and with the violation patterns for those who do complete their sentences, it is estimated that around 25 to 50 percent of probationers will engage in behavior that requires the court’ s attention (Taxman and Cherkos, 1995). In addition, as probationers fail to comply with their conditions, the more violators are sent to prison and jails, which will then contribute to the actual overcrowding problem. Further evaluation findings on felony probation show that the more stringently programs enforce their punitive conditions, the more likely they are to exacerbate prison crowding (Tonry and Hamilton, 1995). Moreover, another study by Landis et al. (1969) concluded that requiring conditions of probation might not be having the desired effects of reducing recidivism rates and decreasing prison crowding. Other studies on probation, however, have shown different results. Vito, (1986) followed a sample of felony probationers in Kentucky for 36 months and found that only 22% were rearrested. Federal probation reported that in 1992, 20,856 probationers were terminated from probation and of these, 81 percent completed probation successfully. Fourteen percent were terminated because of violations of probation (3.5 percent for new crimes and 10.6 percent for new crimes and 10.6 percent for technical violations) (BJS, 1994). In 1979, the National Institute of Law Enforcement and Criminal Justice (NILECJ) reported the results of a review of the probation evaluation literature 26 (Allen, et. al., 1979). This assessment produced general conclusions about probation effectiveness because the studies examined diverse groups of offenders and employed varying definitions of success and follow-up periods. It appeared that a failure rate of 30 percent or less was generally viewed as demonstrating the effectiveness of probation. Probation in the State of Michigan The data for this study comes from a collaborative project by the School of Criminal Justice at Michigan State University and the Michigan Department of Corrections. The study was initiated as an effort to assess probation recidivism in this state. Since 1913 probation has been the primary alternative to prison for most individuals convicted of a felony offense in the state of Michigan. In Michigan, probation may be imposed for all misdemeanors and felonies except murder, treason, armed robbery, criminal sexual conduct in the first and third degree, certain controlled substance offenses and for convictions where a firearm was used in the commission of a felony. In 1995, about 39% of all persons convicted of felonies in Michigan were sentenced to probation (Michigan Department of Corrections, 1997). Currently, there are more than 60,000 adult felony probationers in Michigan under supervision (Michigan Department of Corrections, 1997). The number of probationers under supervision has increased around 12.3% from 45,000 to over 50,5000. With the increase of community supervision, the number of probation violators has also increased and has 27 contributed to the prison-overcrowding problem. During, 1997 around 3,084 adult probationers in Michigan violated their conditions of supervision and were re-sentenced to prison. This figure implies that in 1997, one out of four probation violators received a prison sentence from the courts. Furthermore, almost 86% or 9 out of 10 probation violators sentenced to prison were convicted of or involved in a new felony. The remaining 14% had an average of 4 violations and more than a third included absconding (Michigan Department of Corrections, 1997) While, violators of probation represent 32% of the total prison intake, the Michigan department of Corrections currently has limited data on probation recidivism rates. The Department needs to know how recidivism relates to a probationer’s criminal history, conviction offense, personal characteristics and program participation may identify commonalties in offenders that violate their probation. 28 CHAPTER III ASSESSING HYPOTHESES This study uses quantitative research techniques, using a sample of probationers in Michigan. Following is a discussion of the research questions and related hypotheses. Research Questions Based on the theory and research evidence discussed above, this study seeks to answer the following research questions: Do multiple conditions of probation deter individuals from violating the conditions of their probation terms? This question will be examined using many control variables related to the literature previously discussed. The hypotheses bellow outline the variables that will be examined. All of these hypotheses will be tested at the P<.05 level. Hypotheses Hypothesis 1: Increase in probation conditions decreases risks of failing probation. As previously noted in the literature review, the most recent pattern in probation has been to increase of conditions associated with probation terms. It is then expected that increased surveillance may reduce further violations. However, results had been mixed. This study examines the multiplicity of multiple probation conditions and their effects on recidivism. Multiple conditions is an additive measure of all conditions imposed on probationers for their probation terms. It is expected that the more conditions to probation there are in the probation order, the more compliance by probationers. Currently, the database includes all these conditions to probation orders in a detailed manner. The research team coded these into conditions ordered and completed. The analysis will include the following conditions ordered: alcohol testing, drug testing, mental health treatment, educational requirements, vocational requirements, restitution (fines), community service, other treatment, and supervision fees. Hypothesis 2: Probationers with lengthier criminal histories are more likely to recidivate. It has been stated that certain offenses are more likely to be deterred by activities of the legal system than others. Despite some discussions regarding this matter, there has been little empirical examination of the relative deterrent effects by offense type. This study analyzes the relationships for type of current offense, previous juvenile records/offenses, and drug abuse. In addition, the study examines the relationship between prior misdemeanor and felony convictions and probationers’ discharge type. These aspects have been selected for analysis 30 since supporting literature have demonstrated that they are key variables in understanding and explaining probation recidivism and recidivism in general. Hypothesis 3: Socio-demographic variables influence the performance of offenders placed on probation. Literature has indicated that several socio-demographic variables are consistently associated with probation outcomes. In the present study we will be able to examine several variables identified in previous research. Specifically, in this study, we will consider gender, race, age, employment status, marital status, and highest grade completed. 31 CHAPTER IV METHODOLOGY Chapter II and Ill provided pertinent literature and rationales for selection of variables for the study. In this chapter, information on the sample and data collection, and description of measures are discussed. m A sample of approximately 1500 probationers was randomly chosen from a database of offenders sentenced to probation from January to June, 1996. From this database, probationers who were sentenced from February to March, 1996 were sampled. The year 1996 was chosen to allow three years of available records to conduct follow-up with sufficient time to assess recidivism rates of probationers. The sample size will provide an adequate number of cases to conduct analysis, allowing for attrition or missing records. Data collection The original study was conducted by Sheila Maxwell and Timothy Bynum from the School of Criminal Justice at Michigan State University. The Department of Corrections in Michigan provided access to its field supervision files and databases to identify information about each probationer included in the sample. The field supervision files Contain vast and detailed information on probationers such as the probation officer assessments, conditions of probation, employment history. number and types of violations, different interventions tried 32 before re-sentencing, programs participation and completions, drug and/or alcohol use, among other personal characteristics. lnforrnation obtained from the Department’s database was matched with the field supervision files on key variables like social security number, and State Police ID. Characteristics that are theoretically relevant to recidivism were drawn from all this information. Case files of probationers across Michigan were reviewed to record crime characteristics and violations of probation order. Code sheets were brought to the sites and filled-in using the case files of probationers. Code sheets were keyed by case number. Existing Department of Corrections’ database includes identifiers including the names of probationers, but the names were not recorded in the codesheets and were not used in the analysis. Instrument The data collected for this project was assessed by means of a coding instrument developed by the project’s main researchers. The instrument included variables based on previous surveys and probation studies dealing with recidivism. Included are variables that are supported by literature as established predictors of success and failure in probation. The code sheet was put together after a pre-test looking at five different counties across Michigan. A sample of around 50 case files of probationers was obtained. These files were then thoroughly examined to check for consistency of 33 information across files. The main researchers of the project then put together the code sheet, which entailed pertinent information found in the files. In order to ensure reliability, these cases were also used to train all individuals working on data coding as to how to code the information and where to find it in the probationers’ case files. In addition, a section was provided were all coders decided on the best way to code the information this was done using the interrater reliability method. Descrigtion of Measures This section provides detailed descriptions of the data elements used in this research. Demographic Variables Measures of age, gender, ethnicity, marital status, education level, and employment history are used. Age was computed using the dates of birth available. Ethnicity distinguishes among Caucasian, African-American, Hispanics, and Asians. Due to the overall low frequencies of Asians and Hispanics among the probation clientele in the state of Michigan, these ethnic categories were recoded into a new category referred to as “Other races”. For information on education, the highest grade completed was coded including categories for no formal education, GED, some college, college degree, and post college degree. This variable was recoded into a new variable distinguishing between those probationers who completed up to 12th grade/GED and those with some college. Employment history is obtained from a 34 dichotomous variable (yes-no), which indicated the employment status of the offender at the time of arrest. Criminal Historv Enables Criminal history variables are routinely collected by the probation officers, and are available for study. Information on offenders’ prior offenses is obtained and coded as prior felony convictions, prior misdemeanor convictions, and previous juvenile records. Type of current offense is a nominal measure originally coded by the Michigan Department of Corrections in three categories including NonAssaultive, Drug, and Assaultive offenses. Multiple conditions of probation Multiple conditions of probation are analyzed counting all the conditions ordered in the probation contract. These conditions include punitive conditions such as supervision fees, split jail sentences, alcohol and drug testing, electronic monitoring, fines, and community service. Treatment conditions are also included these comprise conditions such as alcohol and substance abuse treatments, educational training. mental health treatment, vocational/employment requirements, and other treatment required. These conditions to probation are all coded into dichotomous (yes-no) variables. 35 Probation Recinivism In this study, probation recidivism was analyzed using probationers’ discharge type. Therefore, recidivism was conceived as the probationers’ failure to successfully complete their probation terms. Recidivism took place when one of the following discharge types was present: jail pending violation, dead, revocation (not specified), discharge without improvement, and jail. It is necessary to explain that all discharges were coded as they appeared in the files. In all cases, the probation agent is the one who decided the type of discharge. Probationers’ discharge type was recoded by this researcher into a dichotomous variable where the only two possible outcomes for this dependent variable were “failures” (0) and “successful” (1). This recoding allowed proceeding with the statistical analysis using logistic regression. 36 CHAPTER V RESULTS A discussion of probationers’ characteristics is covered in this chapter. Additionally, this chapter includes evidence on bivariate relationships and logistic ngI'BSSIOI'I outcomes. Probationer Characteristics This section provides baseline information on the probationers’ characteristics, including their demographic attributes, criminal history, and discharge type. Demographic and CriminalHfiistorv Characteristics Most of the probationers were men, and a large proportion is Caucasian and African-American (Refer to table 5.1 for a listing of probationer's basic demographic and criminal history characteristics). The clients were also quite young, with the mean age at twenty-nine. The majority of the probationers’ sampled were single and reported no drug abuse. In addition, most of them have an educational level up to 12th grade or GED and almost half of them were employed at the time of the offense. Criminal history measures demonstrated that probationers in Michigan tended to have higher numbers of previous misdemeanor convictions as compared to felony convictions. In addition, a high proportion of probationers in Michigan did not have a juvenile record before entering probation and the majority of them have committed non-assaultive offenses. 37 Table 5.1- Demographics and Criminal History Measures by Dlscharge Type Discharge T13 Demograghics and Criminal Histog Failures Successful Degndent Measure Total % N % N 3/3 Discharge Type 1158 100 548 47 610 53 lndegndent Measures Gender Women 230 20 90 49 1 40 61 Men 928 80 458 39 470 51 Ethnicity Caucasian 606 52 233 38 373 62 African American-Asian 552 48 315 57 62 43 & Hispanics Age Mean-29 Median-27 Range 16-76 1158 100 Highest Grade Completed To 12th grade 998 86 493 49 505 51 Some college 139 12 46 33 93 67 Employment Status Not employed 569 49 325 58 231 42 Employed 556 48 204 36 365 64 Marital Status Single 800 69 412 52 388 48 Married 1 54 1 3 38 24 1 1 6 75 Divorced-Separated-Widow 186 16 91 49 95 51 Drug Abuse No 638 55 251 39 387 61 “Yes 501 43 289 58 212 42 Prior Misdemeanor Convictions None 616 53 255 41 361 59 1 or more 542 47 293 54 249 46 Prior Felony Convictions None ‘ 832 72 352 42 480 58 One or more 326 28 196 60 130 40 Previous Juvenlle Record No 923 17 401 43 522 57 Yes 199 80 127 64 72 36 Type of Offense ' Non Assaultive 647 57 313 48 334 52 Drug 326 28 1 57 48 1 69 52 Assaultive 185 1.7 78 42 107 58 *** *ti iii ii". *t* *t* iii iii *** Notes: * =p<.05 ** $2.01 *** =p<.001 38 Conditions of probation The most frequently imposed conditions of probation were supervision fees and vocational/employment requirements (Table 5.1 gives full details on these conditions). There were also a substantial number of cases with drug and alcohol testing requirements. These were followed by other conditions such as jail, restitution, mental treatment, alcohol and drug treatments, and educational requirements. The data indicated that probationers could be imposed a total of thirteen conditions. The total count of conditions ordered showed that most probationers were ordered to comply with five or more conditions to their probation terms. In order to assess the effects of conditions to probation in a more effective manner, the conditions were also recoded into two different variables. Six conditions were more related to treatment approaches, these were counted together as a single variable. This new variable included such conditions as drug treatment, alcohol treatment, other treatment ordered, mental health treatment, vocational/employment. requirements, and educational requirements. The total count of treatment conditions indicated that around fifty-one percent of probationers in the sample were ordered 2-6 conditions while the rest were included in the category of zero or one treatment conditions. A total of seven conditions to probation were distinguished as punitive conditions. The count of these punishment-oriented conditions showed that most probationers were ordered to comply with zero to four of these conditions as compared to those with 4 or more conditions. Table 5.2- Conditions of Probation by Discharge Type n Itlon of Pro ation Alcohol Treatment No Yes Alcohol Testing No Yes Community Service No Yes Drug Treatment No Yes Drug Testing No Yes Educational Requirements No Yes Mental Health Treatment No Yes Restitution No Yes Supervision Fees No Yes Electronic Monitoring No Yes Other Treatment No Yes Vocational/Employment Requirements No Yes Total Count of Conditions Ordered Mean- 4.65 556 602 0-4 Conditions 5 or more conditions Punitive Conditions 04 Punitive conditions 5 or more punitive conditions Treatment Conditions None One or more Tgtgl 918 224 601 542 762 380 826 317 470 673 823 318 992 151 759 383 234 909 997 1 46 896 203 416 727 Mean- 2.98 1404 170 Mean- 1.68 61 1 963 Dlgcggrge Tygg Ififlflflfifi §flfl¥flflfifiul Zn 1! 22 fl 9.6 79 421 46 497 54 20 113 50 111 50 52 241 40 360 60 47 294 54 248 46 66 360 47 402 53 33 175 46 205 54 71 347 42 479 58 27 188 59 129 41 41 165 35 305 65 58 370 55 303 45 71 346 42 477 58 28 189 59 129 41 86 461 46 531 54 13 74 49 77 51 66 351 46 408 54 33 184 48 199 52 20 104 44 130 56 79 433 48 476 52 86 455 46 542 54 13 80 55 66 45 77 408 46 488 54 18 102 50 101 50 36 181 44 235 57 63 354 49 373 51 Median- 5 Mode- 5 Range- 0-12 48 204 37 352 63 52 344 57 258 43 Median- 3 Mode- 3 Range- 0-7 89 449 45 539 55 11 99 58 71 42 Median- 2 Mode- 1 Range- 0-6 39 74 38 121 62 61 474 49 489 51 Ct. Notes: ' =p<.05, " =p<.01, “’ =p<.001 40 Discharge Type Approximately, fifty-three percent of all probationers completed their probation terms successfully. On the other hand, forty-seven percent of the discharges from probation were considered failures. Note that these failures comprised the categories originally coded as jail pending violation hearing, prison, absconded, revocation (not specified), discharged without improvement, and jail. In addition, twenty-four percent of the probationers were still on probation. It is important to mention that although the original sample included a significant number of probationers who were still on probation, these were excluded from the analysis. Individuals reported as dead were also excluded from the analysis. Since the final measure of recidivism in this study is discharge type, those who have not yet completed their probation terms could not be included for analysis. 41 Bivariate Discussion This section offers the results at the bivariate level. These include analyses where all the independent measures are examined against the dependent measure of discharge type. Demographic Characteristics and Discharge Elbe Relationships The contingency tables for the independent variables of age, gender, ethnicity, highest grade completed, employment status, and marital status were significantly related to probationers’ discharge type. These relationships were all significant at p<.05 level. The independent measure of age indicated that older probationers have a greater tendency to successfully complete their probation terms than younger offenders do. Additionally, there is a greater tendency among females to successfully complete their probation terms as compared to their male counterparts in the sample. It was demonstrated that sixty-one percent of female probationers successfully completed their probation terms versus fifty- one percent of male probationers successfully completing probation. The contingency table for ethnicity showed that Caucasians are more likely to be discharged from probation as successful while African Americans as well as other ethnic groups are more likely to be discharged as failures. This relationship was clearly marked by significant differences, with sixty-two percent of Caucasian probationers as successfully discharged while only forty-three percent of success rates for all the other ethnic groups. 42 Regarding probationers’ education levels, probationers with some college experience had higher success rates. These comprised seventy percent of those who completed their terms successfully. Only fifty-one percent of offenders who reported an education level of 12th grade or GED successfully completed their probation terms. Employment status also resulted significantly related to discharge type. In this case, sixty-four percent of those probationers who have been employed at the time that they committed their offense were more likely to successfully complete their terms. Moreover, married probationers were also more likely to be discharged as successful as compared to those probationers who were single, divorced, separated, and/or widowed. In this instance, around fifty-one percent of the failures reported being single, divorced, etc. while seventy-five percent of the married probationers were discharged as successful. Probationers’ Criminal Histories and Disctfime Tvpe Relationships Criminal history variables included measures of probationers’ prior juvenile records, misdemeanor and felony convictions, prior dug abuse and type of offense. When examining probationers’ juvenile records, it was found that those offenders with previous juvenile records were more likely to fail probation than those with no reported juvenile records. A majority of sixty-four percent of those probationers with prior juvenile records tended to fail their probation terms while fifty-seven percent of those with no prior records completed their terms successfully. 43 It was also found that drug abuse was related to probationers’ discharge type. It was demonstrated that a majority of fifty-eight percent of those individuals who reported drug abuse were more likely to fail probation while sixty- one percent of those with no reported drug abuse were more likely to complete their terms successfully. The contingency tables for prior misdemeanor convictions illustrated that those individuals with one or more reported misdemeanor convictions were more likely to fail probation. Specifically, it was found that fifty-four percent of probationers with one or more misdemeanor convictions had a greater tendency to fail probation. On the other hand, those offenders with no reported previous misdemeanor convictions were more likely to complete their probation terms successfully. The same relationship is true for prior felony convictions, in this case sixty percent of probationers with one or more previous felony convictions were more likely to fail probation as compared to forty-two percent failures of those with no prior felony convictions. The relationships for both prior misdemeanor and felony convictions were significant at P<.05. The contingency tables for type of offense differed from the previously discussed relationships. The observed tendency was for probationers charged with assaultive offenses to be less likely to succeed in their probation terms. However, for this variable, there is no statistically significant relationship when examined against the dependent variable of discharge type. 44 Conditions to Prob_ation and DischmTvpe Re_lationship§ The bivariate analysis of the conditions to probation demonstrated that the count of total number of conditions imposed was related to discharge type. Probationers with five or more conditions had higher failure rates. Specifically, it was found that sixty-three percent of those probationers ordered with zero to four conditions were successful while only forty-three percent of those ordered five or more conditions successfully completed their terms. Indeed, a higher proportion of failures or fifty-seven percent of failures was attributed to those with more than five conditions attached to their probation orders. This tendency demonstrated that these cases with more conditions were more likely to fail probation. Furthermore, analysis of the count of punitive conditions reflected that these conditions are significantly related to discharge type. For probationers with more than five punishment-oriented conditions, the tendency was to fail probation. In this case, fifty-five percent of those individuals with 0—4 punitive conditions were more likely to successfully complete probation terms while fifty- eight percent of those with four or more of these conditions were more likely to fail. The analysis therefore implied that the less punitive conditions probationers are given, the more successful they are in probation. The analysis of the count of treatment conditions was also significantly associated to probationers’ discharge type. Offenders ordered with zero treatment conditions were more likely to successfully complete probation while a vast majority of those ordered with one or more of these conditions were more likely to fail their probation terms. This contingency table demonstrates that the 45 lower the number of treatment conditions the more likely the probationer will be discharged as successful. In effect, sixty-two percent of those individuals with none treatment conditions were more likely to be successfully discharged from probation as compared to only forty-nine percent success rate of those ordered with one or more of such conditions. Additionally, all thirteen conditions to probation were individually examined against the dependent measure of discharge type. As previously mentioned, these conditions include: alcohol treatment, alcohol testing, drug treatment, drug testing, educational requirements, mental health treatment, restitution requirement, supervision fees, electronic monitoring, other treatment ordered, vocationaVemployment requirements, community service, and probation jail split. From individual analyses of these conditions, it was found that only six of them were statistically significant when examined against the dependent variable. The condition of alcohol testing was significantly related to probationers’ discharge type. In this case, those individuals with alcohol testing as a term of probation were more likely to fail. Approximately fifty-four percent of those ordered alcohol treatment failed their probation terms, while sixty percent of the others who were not ordered this condition were more successful. Drug treatment as a term of probation was also significant at the P<.05 level. The majority of probationers who were ordered this condition showed a higher tendency to fail. Failures comprised sixty percent of those ordered with drug treatment. Another condition that exhibited significance when examined against discharge type was drug testing. This condition was associated with 46 more failures from probationers. Around fifty-five percent of those offenders with drug testing as a term of probation were discharged as failures. On the other hand, a huge majority of sixty-five percent of those who were not ordered this condition successfully completed their probation terms. Furthermore, individuals ordered educational requirements were significantly more likely to fail probation. Close to sixty percent of those offenders required to complete educational requirements failed to successfully complete their probation terms. Similarly, offenders who were under electronic monitoring were also more likely to be discharged as failures. A majority of fifty-five percent of those under such condition failed probation while those not on tether were more likely to succeed. Having a probation jail split type of sentence was also associated with more failures from probation. Around fifty-one percent of those probationers with split sentences were more likely to fail their probation terms while those who did not served split sentences were more likely to successfully complete their probation terms. The analysis of the other seven conditions including alcohol treatment, community services, other treatment, mental health treatment, vocational/employment requirements, restitution, and supervision fees, did not reach significance at the p<.05 probability level. 47 Regression Models The following section provides the results of the regression models used in the analysis. The logistic regression technique was used to examine the relationships outlined above. This technique is the most reasonable statistical tool given the binary nature of the dependent variable used in the study, which is probationers’ discharge type (Bachman and Paternoster, 1997). As shown in table 5.3 two models were used in order to examine these relationships. The first model included a count of all probation conditions and the second model included counts of punitive and treatment conditions. The model including all conditions to probation reached significant model chi-square and a -2 log likelihood of 1319.78. On the other hand, the model including punitive and treatment conditions also obtained significant model chi- square and a slightly lower -2log likelihood of 1319.59. The -2|og likelihood is a measure of how well the model fits the data, the smaller the value, the better fit. The slight difference between the —2log likelihood among models indicate that the model including the counts of all probation conditions provides a better fit for the data. These models have high explanatory power. Both models perform well, predicting about sixty-nine percent of all cases correctly, a twenty-three percent improvement over a null model. The null model is based on the modal category of the dependent variable. 48 Since the two models exhibited similar significant variables, other than the differences in counts of conditions to probation, those similar results are discussed together in the following subsections. Also, it is important to mention that all socio-demographic variables that were not continuous as well as criminal history variables and the dependent measure of discharge type were dummy variables with the only possible outcomes of 0 and 1. The variable of type of offense has three categories including assaultive, drug, and non-assaultive offenses. In addition, all the different counts of conditions to probation are continuous variables. In order to facilitate the interpretation of the odds-ratios and predicted probabilities it is necessary to explicate their meanings. Odds ratios help describe the strength of the effect of a variable. The odds ratio can be described as a ratio of the odds at two different values of X (Hamilton, 1992). If the X variable is not related to the value of Y variable, then the coefficient of X will be zero and the odds ratio will be one. The stronger the relationship between the two variables the more distant the odds ratio will be from one (Hamilton, 1992). For a dummy variable this ratio is the odds of the events with and without the dummy variable included in the model. On the other hand, predicted probabilities were used to indicate the effects of probation conditions on successful discharges while controlling for all other measures included in both regression models. For the purposes of this study, the conditions of probation in the models were varied. This was performed in order to assess their effects on probationers’ discharge type. 49 Predicted probabilities of success were obtained by increasing the number of conditions in the model one by one while holding all other variables constant. This produced specific probabilities of success when probationers were ordered one, two, or more conditions. ELffects of socio-d_emographic measures on discharge gpe The independent measure of sex did not reach significance at the p<.05 in any of the two models. Although non-significant it is important to mention that males were more likely to fail their probation terms whereas females were more likely to succeed. Probationers’ age was also non-significant in the two models and the tendency was for younger probationers to be discharged as failures while older individuals were more likely to successfully complete their probation terms. Contrary to what was expected and supported by previous studies, offenders’ sex and age were not related to more recidivism or in this specific case, to more failure to successfully complete probation terms. However, the patterns from females and older probationers as more likely to be discharged as successful, as well as younger males to have higher recidivism rates give some support to previous findings. The measure of highest grade completed was significant in both models. Probationers with lower levels of education were more likely to fail probation than those with higher education levels. Approximately ten percent of those 50 probationers with higher education levels were more likely to successfully complete their terms. Another variable that was also significantly and highly related to discharge type was ethnicity. Caucasian probationers were more likely to successfully complete their terms while African Americans, Asian and Hispanics were more likely to fall. In this case, African American and other minorities were fifty-four percent less likely to complete their probation terms successfully. Probationers’ employment status at the time of offense was statistically significant in both models. Offenders who were employed at the time of their current offenses were more likely to be successfully discharged. The odds ratio in table 5.3 shows that employment has a large effect. Approximately ninety- three percent of employed probationers were more likely to be discharged as successful. Moreover, married probationers are more likely to be successfully discharged. In this case, the odds ratio also showed a large effect where eighty- one percent of married probationers were more likely to succeed as compared to those who were single, divorced, separated or widowed. This relationship was significant in both models at the p<.001 level. Effects of probationers’ criminal historig: on discharge gpe Probationers who reported drug abuse were significantly (43%) less likely to be discharged as successful. However, among those offenders charge with drug offenses, there is a greater tendency to successfully complete their probation terms. This finding is intriguing since one would expect that if individuals who report drug use are more likely to fail their probation terms, then those with drug offenses would also be less likely to succeed in probation. Nevertheless, this is not the case; the tendency is for a majority (38%) of drug offenders to be less likely to fail. The odds ratio for offenders with previous juvenile records/offenses indicate that they were thirty-two percent less likely to successfully complete probation. This relationship is significant at p<.05. In addition, probationers with prior misdemeanors and felony convictions were also less likely to successfully complete their terms. Specifically, those individuals with higher counts of prior misdemeanor and felony convictions were more likely to fail probation. These relationships were statistically significant and consistent with what has been found in the literature where offenders with lengthier criminal histories were more likely to recidivate. Effects of conditions to probation on discharge gpe In the first model including a count of all probation conditions, probationers with higher counts of conditions were significantly less likely to successfully complete their probation terms when controlling for the variables previously mentioned. This regression model indicated that probationers with higher counts of conditions were more likely to recidivate in this case, meaning, more failures from probation. As previously mentioned. model B included counts of punitive and treatment conditions. In this model the count of punitive conditions was significantly related to a lesser likelihood of successfully completing probation. Specifically, fourteen percent of those with higher counts of punitive conditions were more likely to fail their probation terms. On the other hand, the count ‘of treatment conditions was not significantly related to discharge type. Although non-significant, it is important to mention, that in this situation, again, there was a greater tendency to fail by those probationers with higher counts of conditions. Figure 1 (see Appendix A) shows the effects of conditions to probation, while holding all other variables constant at their means and all dummy variables at their modal category. Together these lines indicate the effect of the conditions to probation on discharge type. Notice that having more conditions of probation significantly decreases the predicted probability of successful discharges. From this analysis then, we find that no matter what type of conditions probationers are imposed, the tendency will be to fail if they have to comply with more conditions. While, on the contrary, the predicted probabilities also tell us that probationers with fewer counts of conditions are more likely to successfully complete their terms. For example, probationers ordered with twelve punitive conditions are twenty-nine percent more likely to fail probation as compared to those who had only one punitive condition ordered. These patterns of failing probation if offenders have been ordered more conditions of probation are contrary to previous findings on the deterrent effects of more severe measures. 53 Table 5.3. Predictors of Dischgge Type Using Logistic Regression Predictors Sex (Female, 1) Age Highest Grade Completed EthnIcIty (African American and other minorities, 1) Employment Status (Employed, 1) Marital Status (Married, 1) Dug Abuse (Yes, 1) Previous Juvenlle Record (Yes, 1) Offense Category Non-Assaultive Drugs Assaultive Prlor Misdemeanor Convlctlons Prior Felony Convictions Count of all. Probation Condltlons Count of Punitive Conditions Count of Treatment Condltlons ModeIA ModeIB Includes Count of all Includes Count of Punitive J Probation Conditions and Treatment Conditions ‘ I: . T ‘ f»: _e_ S.E. odds Ratio p s15. Odds Ratio .30 .17 1.34 .29 .17 1.34 .01 .01 1.01 .01 .01 1.01 .09 .04 1.09 ** .09 .04 1.10 ** -.77 .14 .46 *** -.77 .14 .46 *** .66 .14 1.93 *** .66 .14 1.93 *** .60 .23 1.82 ’* .60 .23 1.81 ** -.56 .15 .57 *** -.56 .15 .57 *** -.38 .19 .68 *" -.38 .19 .68 ** .13 .20 1.14 .13 .20 1.14 .33 .17 1.39 ** .33 .17 1.38 ** Reference Category Reference Category -.08 .03 .92 ** -.08 .03 .92 ** -.18 .07 .83 *** -.18 .07 .83 *** 13 .03 .88 ”* -.15 .05 .86 *** -.11 .06 .90 -2Log marina—68"" . Model Chi-Square Notes: * = p<.05 1319.78 1319.59 205.07 205.26 ** = p<.o1 “7": p<.001 54 CHAPTER VI DISCUSSION Probation is the most common form of community supervision in the Criminal Justice System. As most forms of intermediate sanctions, probation has also been experiencing constant changes. Nowadays, probation has turned into a more intensive and punitive oriented type of supervision. In such a manner, more conditions are often added to probation terms. However, previous research has failed to look at the effects of those conditions on probationers’ recidivism. The aim of the present study was to assess the effectsof multiple conditions of probation on probationers’ recidivism. Interestingly, the £959th did not provide any support for the hypothesis that an increase in probation conditions decreases the risks of failing probation/Specifically, this study found that probationers ordered higher counts of conditions to probation are significantly more likely to recidivate. Although weak, this relationship was significant in both models using logistic regression analysis. This pattern seems to be highly prevalent and consistent. In addition, the distinction among punitive and treatment conditions on discharge type illustrated once again that the more of such conditions the more likely the probationer will be discharged as a failure meaning more prison and jail sentences, revocations, etc. Therefore, this study also provides support for what has been found in previous literature regarding the trend of more severe conditions leading to an increase in failures. This 0 I U" tendency may help explain the overall prison overcrowding in the state of Michigan. Likewise, this study intended to examine the effects of socio-demographic and criminal history variables on probationers’ recidivism. As previously mentioned in the results section, age and sex were not significant in the regression models. In” contrast, resultsbfrornprevious studieshayg indicated that younger males are. more likely to recidivate. An important difference that may help explain the inconsistencies in these findings was controlling for multiple conditions of probation. The tendency to fail probation while having more conditions was higher for females and males and also for younger and older probationers. 4, On the other hand, other relevant demographic characteristics 31191ng support forprevious studies on probation recidivism and recidivism in general. Such attributes as being White, employed, and married are more related to probationers’ success rates while African American, Asian and Hispanics, unemployed, and single probationers are more likely to fail their probation terms. Criminal history measures also provided support for the hypothesis that probationers’ with lengthier criminal histories are more likely to recidivate. The negative effects of prior juvenile records, drug abuse, and prior misdemeanor and felony convictions on discharge type buttressed this argument. The measure of type of offense also yielded support for this hypothesis. However, it seems that probationers who are convicted for drug offenses are more likely to 56 successfully complete their terms as compared to those charged with non- assaultive and assaultive offenses. In overall, this study provides powerful implications for policy as it relates to the future of probation in the United States. It emphasizes that a more pragmatic identification of factors that influence recidivism is needed in order to develop policies that will help lower recidivism rates and at the same time reduce prison overcrowding. This study also provides support for what has been found in previous literature regarding the trend of more severe conditions leading to an increase in failures. Implications for Future Research This study constitutes only a small part of the many issues that need to be examined with regard to recidivism. It was pointed out earlier that most studies that examine probation recidivism typically overlook its sanctioning effects. This study was most interested in the severity of sanctions as measured by multiple conditions to probation. There was no clear agreement on past research regarding this issue. so it seemed that new approaches in measuring severity were needed. Since this research did not find support for the deterrent effect of severity as measured by multiple conditions of probation, it is suggested that other important aspects of the deterrence theory -celerity and certainty- be examined. In addition, the author considers that the defiance effects of sanctions (Sherman, 1993) may need to be examined in order to understand why multiple conditions 57 of probation are more likely to lead to failures instead of reducing recidivism rates. A clear pattern in this study was for minorities, unemployed, and non- married probationers to fail probation. In fact, these were the variables with larger effects in the logistic models. The defiance effects of sanctions may help explain this pattern. The defiance effect of sanctions entails reoffending by those who think sanctions are unfair (Sherman, 1993). Future studies may have to consider that employment and marital status -both measures of social stability- and race, are key variables in understanding recidivism. More importantly, further research should consider examining whether certain probationers, specifically, minorities, unemployed, and non-married are more likely to receive more conditions of probation. If so, it might be probable that individuals with these characteristics perceive that sanctions have been unfairly imposed on them and they react by deviance, meaning further law breaking. Limitations of the Study It is important to mention that the present study did not intend to examine specific violations of probation nor did it look at responses to any violations from the part of probation agents. Since the study used the final discharge type as a measure of recidivism, it may so happen that probationers who were discharged as successful had in effect committed some technical violations to their probation terms. 58 Additional limitations of the present study include the facts that: (a) The data used were originally collected (in files) for administrative rather than research purposes and are subject to the short-comings of such information, (b) Data were pre-collected from files and probationers with no reported recidivism could have been (1) truly non-criminal, (2) criminally active but not apprehended, or (3) criminally active but apprehended in a jurisdiction other than Michigan. Another limitation concerns missing files that were unavoidably excluded from the study. Although we have no reason to believe that they would differ from the all the other cases sampled, it is important to make clear that they were not available, therefore there is no way to make certain this is in fact the situation. Although these limitations must not be ignored, they should not be considered overwhelming. The results in this study provide a number of important and substantive directions for further research, particularly in the policy arena regarding probation and other intermediate sanctions and their effectiveness in reducing recidivism rates. 59 APPENDIX A 60 Predicted Probability of Success Figure 1. Effects of Probation Conditions on Succesful Discharges 0.5 0.4 0.3 0.2 0.1 0 1 i I I I I I r l 1 I I 0 l 2 3 4 5 6 7 8 9 101112 Count of Conditions + All Conditions * Punitive & Treatment Conditions + Punitive Conditions + Treatment Conditions 61 REFERENCES Abadinsky, Howard. (1991). Probation and Parole. (4th edition) Englewood Cliffs, NJ: Prentice-Hall Anderson, Elijah. (1978).. A Place on the Corner. Chicago: University of chicago Press. Andrews, DA. and Bonta, James. (1994). The Psychology of Criminal Conduct. Cincinnati, Ohio: Anderson Andrews, D.A.; Bonta, James; and Hoge, Robert. (1990). Classification for Effective Rehabilitation: Rediscovering Psychology. Criminal Justice and Behavior. 17:19-52. Bachman, Ronet, and Paternoster, Raymond. (1997). Statistical Methods for Criminology and Criminal Justice. McGraw- Hill. Ball, John C. (1955). The Deterrence Concept in Criminology and Law. Journal of Criminal Law, Criminology, and Police Science. 46:347-354. Bureau of Justice Statistics (1991). Probation and Parole Violators in State Prison, 1991. Washington, DC: US Department of Justice. Bureau of Justice Statistics (1992). Census of Probation and Parole, 1991. Washington, DC: US Department of Justice. Bureau of Justice Statistics (1995). Correctional Populations in the United States. Washington, DC: US Department of Justice Boostrom, Ron. (1995). Enduring Issues in Criminology. Greenhaven Press, Inc. San Diego. Braithwaite, John. (1985). To Punish or Persuade: Enforcement of Coal Mine Safety. Albany:Suny. Bridges, George S. and Stone, James A. (1986). Effects of criminal punishment on perceived threat of punishment: Toward an understanding of specific deterrence. Journal of Research in Crime and Delinquency . 23:207- 239 Brooker, Frank. (1971). The Deterrent Effect of Punishment. Criminology. 9:469-490. 62 Byme, James M., Lurigio, Arthur J., Petersilia, Joan. (1992). Smart Sentencing: The Emergence of Intemiediate Sanctions. Sage Publications: Newbury Park, CA Chambliss, William J. (1966). The deterrent influence of punishment. Crime and Delinquency. 12: 70 — 5 Chambliss, William J. (1967). Types of deviance and the effectiveness of legal sanctions. University of Wisconsin Law Review. 67:703-719 Chiricos, Theodore G. and Waldo, Gordon P. (1970). Punishment and crime: An examination of some empirical evidence. Social Problems. 18:200- 217 Claster, D. (1967). Comparison of Risk Perceptions Between Delinquents and Nondelinquents. Journal of Criminal Law, Criminology, and Police Science. 58: 80-86. Clear, Todd. (1988). Statistical Prediction in Corrections. Research in Corrections Monograph. Washington, DC: National Institute of Corrections Craig, M. M., and Budd, L. A. (1967). The Juvenile Offender: Recidivism and Companions. Crime and Delinquency. 13:344-351. Decker, Scott H. and Kohfeld, Carol W. (1990). Certainty, severity, and the probability of crime: A logistic analysis. Policy Studies Journal. 19:1 2-21 Dembo. R., Washbum, M., Wish, E. D., Schmeidler, J., Getreu, A., Berry, E., Williams, L., and Blount, W.R. (1987). Further Examination of the Association Between Heavy Marihuana Use and Crime Among Youths Entering a Juvenile Detention Center. Journal of Psychoactive Drugs. 19:361-373. Erwin, Billie S. (1987). Evaluation of Intensive Probation Supervision in Georgia. Atlanta, Georgia: Department of Corrections. Fagan, J, (1994). Do criminal sanctions deter crime? In drugs and the Criminal Justice System: Evaluating Public Policy Alternatives, edited by MacKensie, D.L.; Uchida, C. Newbury Park, CA: Sage Publications Falkowsky, Carol. (1984). The Impact of Two-Day Jail Sentences for Drunk Drivers in Hennepin County, Minnesota. Final Report. Washington, DC: National Highway Safety Administration. Farrington, David P. (1977). The Effects of Public Labeling. British Journal of Criminology. 17:1 12-125. Farrington, David P.; Ohlin, Lloyd E.; and Wilson, James O. (1986). Understanding and controlling crime. New York: Springer-Vedag Ganzer, V. J., and Sarason, l. G. (1973). Variables Associated with Recidivism Among Juvenile Delinquents. Journal of Consulting and Clinical Psychology. 40:1 -5. Geerken, Michael R. and Hayes, Hennessey D. (1993). Probation and parole: Public risk and the future of incarceration alternatives. Criminology. 31:4 549-564 Gendrau, Paul; Paparozzi, Mario; Little Tracy; and Goddard, Murray. ( ). Does “punishing smarter” work? An assessment of the new generation on alternative sanctions in probation. Gendrau, Paul; Little, Tracy; and Goggin, Claire. (1996). A meta-Analysis of the predictors of adult offender recidivism: What works! Criminology. 34:4 575-607. Gibbs, Jack P. (1968). Crime, punishment and Deterrence. Southwestern Social Science Quarterly. 48:515-530. Gibbs, Jack P. (1975). Crime, punishment and deterrence. New York: Elsevier Gold, Martin, and Williams, Jay. (1970). National Study of the Aftermath of Apprehension. 323-12. Hamilton, Lawrence C. (1992). Regression with Graphics: A Second Course in applied Statistics. Belmont, CA: Duxbury Press. Hanson, C. L., Henggeler, S. W., Haefele, W. E., and Rodick, J. D. (1984). Demographic, Individual, and Family Relationship Correlates of Serious and Repeated Crime Among Adolescents and Their Siblings. Journal of Consulting and Clinical Psychology. 52:528-538. Jensen, Gary. F. (1969). Crime Doesn’t Pay: Correlates of a Shared Misunderstanding. Social Problems. 17:189-201. Jensen. Gary. F. (1976). Sex Differences in Delinquency. Criminology. 13:427-448. Kinsey, Karyl. (1992). Deterrence and Alienation Effects of IRS Enforcement: An Analysis of Survey Data. In Why People Pay Taxes, edited by Joel Slemrod, Ann Arbor:Unversity of Michigan Press. 64 Klein, Malcolm. (1986). Labeling Theory and Delinquency Policy: An Empirical Test. Criminal Justice and Behavior. 13:47-79. Landis, Judson R.; Mercer, James D.; Wolff, Carole. (1969). Success and failure of adult probationers in California. Journal of Research in Crime and Delinquency. 6:34-40. Langan, Patrick. (1994). Between Prison and Probation: lnterrnediate Sanctions. Science, 264:791-793. Lauen, Roger. (1990). Community-Managed Corrections: And Other Solutions to Americas Prison Crisis (2"‘1 edition). American Correctional Association. Liska, A. E. (1987). Perspectives on Deviance. Englewood Cliffs, NJ: Prentice Hall. Loeber, R., Dishion, T. (1983). Early Predictors of Male Delinquency: A Review. British Journal of Criminology. 15:43-50. Logan, Charles H. (1972). General Deterrence Effects of Imprisonment. Social Forces. 57:64-73. Lurigio, Arthur J. and Joan Petersilia. (1992). “The emergence of intensive probation supervision programs in the United States. In Smart Sentencing: The Emergence of Intermediate Sanctions, edited by James E. Byme, A. J. Lurigio, and J. Petersilia. Newbury Park, CA: Sage Martinson, Robert. (1974). What Works? Questions and Answers about Prison Reform. Public Interest. 35:22-54. Maxwell, Sheila R. (1994). Formal and Informal Social Controls on Drug Treatment Retention of Offenders.(A dissertation). Rutgers University: New Jersey McCarthy, Belinda R. and McCarthy, Bernard J., (1997). Community - Based Corrections. Third Edition. Wadsworth Publishing Company. Michigan Department of Corrections (1997). Supervision of Felons in the Community. Moore, D. E., and Arthur, J. L. (1989). Juvenile Delinquency. In T.H. Ollendick and M. Hersen (Eds.). Handbook of Child Psychopathology. (2nd edition). New York: Plenum. 65 Morgan, Kathryn D. (1994). Factors Associated with Probation Outcomes. Journal of Criminal Justice. 22:341-353. Myner, Julye. (1998). Variables Related to Recidivism Among Juvenile Offenders. International Journal of Offender Therapy and Comparative Criminology. 42:65-80. Niarhos, E. J., and Routh, D. K. (1992). The Role of clinical Assessment in the Juvenile Court: Predictors of Juvenile Dispositions and Recidivism. The Journal of Clinical Child Psychology. 21 :151-159. Orland, L. (1973). Justice, Punishment, Treatment. The Correctional Process. New York: The Free Press Palmer, J. (1978). Economic analysis of the deterrent effects of punishment: A review. Journal of Research in Crime and Delinquency, 14, 4- 21. Paternoster, Raymond and Piquero, Alex. (1995). Reconceptualizing Deterrence: An Emprical Test of Personal and Vicarious Experiences. Journal of Research in Crime and Deliquency. 32:251-286. Pearson, Frank S. (1987). Research on New Jersey’s Intensive Supervision Program. National Institute of Justice, Washington, DC. Petersilia, Joan. (1985). Community Supervision: Trends and Critical Issues. Crime and Delinquency, Vol. 31, 3:339-347. Petersilia, Joan. (1997). Probation in the United States. Practices and Challenges-“National Institute of Justice Journal. 28. Petersilia, Joan and Turner, Susan. (1986). Prison versus probation in California. Santa Monica, Calif. Rand. Petersilia. Joan, Susan Turner, James Kahan, and Joyce Peterson. (1985). Executive Summary of Rand’s Study. Granting Felons Probation: Public Risks and Alternatives. Crime and Delinquency, 31 :379-392. Pontell, Henry N. (1984). A Capacity to Punish. Indiana University Press: Bloomington. Roshier, Bob. (1989). Controlling Crime: The Classical Perspective in Criminology. Open University Press: Milton Keyner-Philadelphia. Schneider, Anne. (1990). Deterrence and Juvenile Crime. New York: Springer-Verlag. 66 Schwartz, Barry. (1968). The effect in Philadelphia of Pennsylvania’s Increased penalties for rape and attempted rape. Journal of Criminal Law, Criminology and Police Science, 55: 509-515. Sedgwick, Jeffrey L. (1980). Deterrence versus retribution: A debate on the meaning of punishment in policy implementation. Edited by John Brigham and Don W. Brown. Beverly Hills: London Severy, L.J. and Whitaker, J.M. (1982). Juvenile Diversion: An Experimental Analysis of Effectiveness. Evaluation Review. 6:753-774. Sherman, Lawrence W. (1993). Defiance, Deterrence, and lrrelevance: A Theory of the Criminal Sanction. Journal of Research in Crime and Delinquency. 30:445-473. Sherman, Lawrence W.; Gartin, Patrick; Doig, David; and Miller, Susan. (1986). The effects of jail time on drunk drivers. Presentation at the Annual Meeting of the American Society of Criminology, Atlanta, November. Silberman, M. (1976). Toward a Theory of Criminal Deterrence. American Sociological Review. 41 :442-461. Smith, Douglas A. and Gartin, Patrick R. (1989). Specifying Specific Deterrence: The Influence of Arrest on Future Criminal Activity. American Sociological Review. 54:94-105. Stafford, Mark and Warr, Mark. (1993). A Reconceptualization of General and Specific Deterrence. Journal of Research in Crime and Delinquency. 30:123-135. Taxman, Faye S. and Cherkos, Robert. (1995). lnterrnediate Sanctions: Dealing with Technical Violations. Corrections Today. 57:46-54. Tittle. Charles R. (1969). Crime rates and legal sanctions. Social Problems. 16: 409-423 Tittle, Charles R. (1974). Certainty of arrest and crime rates: A further test of the deterrent hypothesis. Social Forces. 52: 455- 462 Tittle, Charles R. and Logan, Charles H. (1973). Sanctions and deviance: Evidence and remaining questions. Law and Society Review. 7:371-392. Tittle, Charles R. (1980). Sanctions and Social Deviance: The Question of Deterrence. Praeger Publishers: New York. 67 Tonry, Michael (1990). Stated and Latent Features of IPS. Crime and Delinquency. 36:174-191 . Tonry, Michael and Hamilton, Kate. (1995). Intermediate Sanctions in Overcrowded Times. Northeastern University Press: Boston Tyler, Tom. R. (1990). Why People Obey the Law. New Heaven: Yale University Press. Van Andel, Henk. (1989). Crime Prevention that Works: The Case of Public Transport in the Netherlands. British Journal of Criminology. 29:47- 56. Van den Haag. (1975). Punishing Criminals: Concerning a Very Old and Painful Question. New York: Basic Books. Vito, Genaro F. (1991). Felony probation and recidivism: Replication and response. Federal Probation. 50:17-25 Waldo, Gordon P. and Chiricos, Theodore G. (1972). Perceived penal sanction and self-reported criminality: A neglected approach to deterrence research. Social Problems. 19: 522-540 Wierson, M., and Forehand, R. (1995). Predicting Recidivism in Juvenile Delinquents: The Role of Mental Health Diagnoses and the Qualifications of Conclusions by Race. Behavior Research Therapy. 33:63-67. Wilson, James O. (1983). Thinking About Crime. Rev. ed. New York: Basic Books. Wolfgang, Marvin E., Figlio, Robert M., and Sellin, Thorsten. (1972). Delinquency in a Birth Cohort. Chicago:University of Chicago Press. Yu, Jiang. (1994). Punishment Celerity and Severity: Testing a Specific Deterrence Model on Drunk Driving Recidivism. Journal of Criminal Justice. 22:355-366. 68 HICHIGRN STRTE U NIV. LIBRRRIES llllllllllllllllllllllllllllll 0604017 1|)le Ill 9302