Needs Assessment and Treatment Plan

chaearvie

classificationofwomenoffenders.pdf

Home >Social Science homework help >Sociology homework help >Needs Assessment and Treatment Plan

D E

RT MEN

T OF JUSTIC E

N ATIO

NAL INSTITUTE OF C OR

RE CT

IO N

U.S. Department of Justice

National Institute of Corrections

A National Assessment

of Current Prac tices

U.S. Department of Justice National Institute of Corrections

320 First Street N.W. Washington, DC 20534

Morris L. Thigpen, Director

Susan M. Hunter, Chief, Prisons Division

Sammie D. Brown, Project Manager

National Institute of Corrections World Wide Web Site

http://www.nicic.org

Classification of Women Offenders: A National Assessment of Current Practices

Patricia Van Voorhis, Ph.D.

Lois Presser, MBA

University of Cincinnati

Division of Criminal Justice

August 2001

This report was funded by the National Institute of Corrections under cooperative agreement #99P10IGL2 with the University of Cincinnati Division of Criminal Justice. Points of view or opinions stated in this document are those of the authors and do not necessarily represent the official position or policies of the U.S. Department of Justice.

Executive Summary.............................................................................................v

Approach ....................................................................................................v

Findings ................................................................................................... vi

Classification Practices Currently Under Consideration ........................viii

Discussion .................................................................................................ix

Chapter 1. Introduction .....................................................................................1

Chapter 2. Issues in Classifying Women Offenders: The Literature ............3

Classifying Women for Security and Custody ...........................................3

Internal Classification Systems ..................................................................6

Classification for Case Management and Treatment Purposes..................6

Legal Impediments .....................................................................................8

Chapter 3. Approach ........................................................................................11

Development of Assessment Questions ...................................................11

Identification of and Contact With Respondents......................................11

Telephone Discussion/Assessments .........................................................12

Focus Groups ...........................................................................................12

Chapter 4. Findings...........................................................................................13

Context of Classification ..........................................................................13

Are Women and Men Classified Differently? ..........................................15

Innovative Practices Currently in Use......................................................17

Innovative Practices Currently Under Consideration ..............................19

Chapter 5. Discussion........................................................................................23

Notes ..................................................................................................................25

References......................................................................................................... 29

iii

Contents

This report presents the findings of a national assessment of state and federal prac- tices for classifying women offenders conducted under the National Institute of Correction’s (NIC) Classification of Women Offenders Initiative. Information was collected from discussions on classification strategies with correctional administra- tors and representatives from classification and research offices throughout the 50 states and the Federal Bureau of Prisons, focus groups with staff and inmates, and ongoing work with agencies engaged in reforming the system. A detailed descrip- tion was sought of current practices in use for women offenders. Respondents’ per- ceptions of whether their systems “worked” for women and served as a tool for making custody, programming, and housing decisions were noted, and the extent to which agencies found the classification needs of women offenders to be different from men was explored. Finally, inquiries were made about the psychometric qual- ity of current systems: What were the origins of the system? Was it developed with women offenders in mind, or was it designed for men and applied to women? Had the systems been validated for women offenders?

These issues were framed by an emerging body of literature on women offenders. An overview of this literature revealed uncertainty about the purpose of objective classification systems for women in prisons. Authorities questioned whether women are as dangerous as men and whether they can be identified as custody or security risks by using the same variables as those used for men. Additional con- cerns faulted the lack of attention to women’s unique needs, criticizing classifica- tion models for their inattention to relationship, abuse, mental health, and parenting issues. Earlier surveys of state and federal classification practices note that most classification systems for women were designed for men and applied to women, with no attention to whether the systems were valid for women offenders. This assessment sought to discuss these concerns with those who work with women offenders on a daily basis.

Approach

A number of the issues posed above were addressed through a telephone assess- ment of representatives from 50 state correctional agencies and the Federal Bureau of Prisons. The telephone discussions took place between February 29 and May 23, 2000. Thirty-three respondents were state directors of classification; five were administrators of women’s facilities; four were research analysts; one was a clini- cal director; and eight were serving in some other administrative capacity. Consultant or researchers who had worked with the agencies were also interviewed.

Executive Summary

Findings

Are Women and Men Classified Differently?

Although many respondents discussed differences between men and women offenders in terms of needs and risk to institutional and public safety, few states have incorporated these differences into their objective prison classification instruments.

All states consider the risk and custody assessment to be the foundation of their classification approach. For 21 states, the custody instrument is a variation on a sys- tem developed by NIC and contains mostly static (unchangeable) factors. Most states reclassify offenders using an objective reclassification tool, administered at intervals ranging from 3 to 6 months, although 14 states only reclassify women on an annual basis. For the many women serving short sentences, annual reclassifica- tion means no reclassification.

Most of the existing classification models were developed for men (4 states) or for men and women both (39 states). To say that a system was developed for men and women, however, is not the same as saying that the system contains variables that are relevant to women, or that it has been validated for women.

Thirty-four states reported using an objective tool to summarize offender needs. More specific tests and inventories are used to measure single needs (e.g., educa- tion, substance abuse, and mental health). Only eight states reported use of a sys- tem that identified needs in a gender-responsive manner.

Although 17 agencies reported using internal classification systems to guide hous- ing assignments, none reported using any gender-responsive systems. Twelve states have made some change to the way women are classified for purposes of custody and public safety:

¢ Four states have a separate custody classification system for women (Idaho, New York, Massachusetts, and Ohio).

¢ Four states have different cutoff scores for men and women, representing an attempt to tie each custody level to a group of offenders with similar behavioral outcomes.

¢ Two states have changed variables to better reflect the nature of women’s dis- ciplinary behavior in prison. For example, New York State reduces points on a common variable, seriousness of the current offense, for women who murdered an abuser.

¢ Four states have expanded options on existing variables to accommodate the nature of women’s offending and infractions. Employment variables have been expanded in some instances to avoid classifying full-time homemakers or

Executive Summary

parents as unemployed. In addition, escape variables have been modified to provide lower scores to offenders who walk away from community and other nonsecure settings.

What Should Be the Purpose of Classifying Women Offenders?

The overwhelming majority of respondents noted security and public safety as the central purpose to be served by incarcerating and classifying women offenders. However, respondents in 15 states indicated that some goals are more central to women offenders than to men. These respondents voiced a desire for classification models that would better support gender-responsive programming and move less serious women offenders more quickly through the system.

What Should Govern the Classification of Women Offenders?

Most women offenders are less dangerous then male offenders. A concern for mov- ing women to lower custody and community levels as soon as possible accords with an underlying sense that women offenders do not pose the same security risks as men. Fifty-one percent of the respondents reported either that women pose less risk than men or that a much smaller portion of women offenders than men pose seri- ous risks to institutional and public safety.

Women have different needs than men. Responses to questions about women’s pro- gramming and treatment needs strongly echoed the emerging writings on gender- responsive programming. Forty-nine respondents (92 percent) asserted that women have unique needs that should be addressed in correctional settings. These issues included trauma and abuse, self-esteem and assertiveness, vocational skills, medical care, mental health, parenting and childcare, and relationships.

What Problems Are Encountered in Classifying Women Offenders?

Many states find that existing systems overclassify women offenders. Too many women are unnecessarily assigned to high custody levels, which then requires offi- cials to override the classification decisions. Representatives of 10 states indicated that they overrode more than 15 percent of their classification scores (rates ranged from 18 to 70 percent of their cases). In effect, these agencies (20 percent of the total) indicated that their systems were not working for women offenders.

Many states do not use the custody classification systems to separate women offenders. Given the observation that public safety and security is their primary concern, it was surprising to learn that in 35 states, women with different custody scores are housed together in at least one, if not all, of their state’s facilities for women. Regardless of the assigned custody level, women, unlike men, are often assigned to an institution where custody does not impact housing, privileges, pro- gramming, or movement throughout the facility. Furthermore, this practice occurs without an increased risk of prison misconduct or breach of security. Custody

vii

Classification of Women Offenders

scores affect whether a woman may work outside of the facility, the types of restraints required when transporting her outside of the institution, and whether she may move to a community placement.

Only 14 states reported validating their custody classification systems for a sample of women offenders. Another 11 states combined women and men in the same val- idation sample. Combined samples, however, cannot be considered adequate vali- dation studies of women offenders because they often contain far fewer women than men. The resulting statistics are driven by the majority of the sample: men. Six of these 14 validation studies occurred during the past year as part of an NIC coop- erative agreement to improve several state external classification systems.

In spite of the fact that 49 respondents identified needs and problems that were unique to women offenders, only 8 states have made provisions to assess men and women differently. Unfortunately, needs assessments greatly impact programming because when needs are not identified they are likely to be overlooked.

Classification Practices Currently Under Consideration

States participating in the NIC cooperative agreements are changing to their classi- fication models. These changes are under study and will not be recommended for implementation until they have been validated and deemed effective. According to focus groups with prison staff and women offenders, overclassification problems are the result of a few variables, common to most custody classification systems, which can produce misleading scores among some women offenders. These prob- lematic variables include—

¢ The seriousness of the current offense for women who killed an abuser: The majority of women who commit a one-time offense in response to a sustained period of abuse are not seen as violent by women offenders or prison staff. Yet, in most current classification models, “seriousness of the current offense” is the mainstay of the system. It is a heavily weighted variable designed to keep mur- derers in high-custody settings for extended periods of time. A number of states are considering lowering the weights of this measure for a first-time violent offense against an abuser.

¢ Variables without options pertinent to women: Overclassification sometimes occurs because such key variables as employment and escape history do not proffer attributes that speak to women’s lives. Women who have been engaged in full-time parenting are scored as unemployed, thereby inflating their custody score when parenting often suggests stability. Another example involves women’s escapes. Fewer women than men escape a secure perimeter. However, women are perceived to be more likely than men to walk away from a commu- nity residential setting to visit children, family, or friends. In some states, many of the higher custody female offenders are women who walked away, returned, and then received additional points on the heavily weighted escape variables. In

viii

Executive Summary

many cases, walkaways can be counted again as disciplinary infractions. Differentiating according to type of escape should solve this problem.

¢ Timeframes on reclassification variables: Prison misconduct drives up reclassi- fication scores in most states. In many cases, the misconduct drops from the reclassification score after 2 or 3 years. For the many women serving short sen- tences, misconduct inflates reclassification scores and there is no time to reduce them. In effect, misconduct acts as a static variable for most women and a dynamic variable for most men in prison. Altering these timeframes can help correct this discrepancy.

¢ Staff training: Staff often observe women to be more difficult to supervise than men. Women in prison ask more questions, question authority, want to talk things over, and challenge decisions. Staff inexperienced with these differences seem more likely to “write up” women inmates. Both staff and inmates report- ed situations where custody scores are driven up by rather minor misconduct reports. Staff training and changes to disciplinary variables and policies could help alleviate this problem.

¢ Gender-responsive needs: The literature and focus groups concur that women offenders represent a more troubled group than male offenders. Rates of men- tal illness, relationship issues, physical abuse, and sexual abuse far surpass those for men (e.g., see Belknap, Dunn, and Holsinger, 1997; Dembo et al., 1992; Holsinger, 1999; Miller et al., 1995; Morash et al., 1998). It is essential that women’s needs assessments adequately capture needs that are unique to women. Procedures to identify problems that cannot always be identified through central files or intake interviews are key to this effort. In some cases, agencies could also consider incorporating these items into the custody classi- fication model if doing so improves the validity of the custody classification and if high custody on the classification system translates into assignment to a facility that provides intensive mental health, substance abuse, and other services.

¢ Use of community risk assessment instruments rather than custody classifica- tion instruments: What is to be made of states that employ custody classifica- tion systems but do not use them to make institutional assignments? If a custody classification is only used to inform movement, such as work assign- ments outside of the prison, rather than institutional assignments, perhaps a community risk assessment would be preferable to the custody score, especial- ly for offenders serving short terms.

Discussion

This assessment, work on the cooperative agreements, and focus groups paint a pic- ture of little change since Burke and Adams (1991) conducted a similar survey 9 years ago. Most states continue to use identical systems for men and women, and

Classification of Women Offenders

most states have no empirical basis for knowing whether the system is offering accurate classifications. Almost all the respondents claim that women’s needs are different than men’s, but only eight agencies have incorporated these differences into needs assessments. The number of states making improvements has not changed significantly since the earlier survey. In fact, two states that had gender- unique classification systems for women have reverted to a system that is uniform for men and women.

It is noteworthy that few states have designed systems that started with women in mind. Most map existing male-based assumptions regarding the goals and pur- poses of corrections onto women and the systems that classify them. The gaps are apparent. Most states do not use the notion of custody in the same way for women and men. Many states do not base housing decisions on custody. Many respon- dents do not consider women as dangerous as men, and a common observation of focus group participants is that women with high custody scores are not necessar- ily more dangerous than women with low custody scores. It is hard to avoid the observation that in practices involving women offenders, custody and risk is a very confused concept.

Executive Summary

Existing classification

models for women—

many of them

designed for male

prisoners—are not

relevant to the needs

of women offenders.

Nor are these models

informed by emerg-

ing research on

gender-specific pro-

gramming or mana-

gerial considerations

unique to women

offenders.

Accelerating growth in the number of incarcerated women offenders (Beck and Mumola, 1999)1 underscores the need for sound strategies of classification and assessment. Two recent nationwide surveys of correctional administrators and prac- titioners independently observed that existing classification models for women— many of them designed for male prisoners—are not relevant to the needs of women offenders (Burke and Adams, 1991; Morash, Bynum, and Koons, 1998). Nor are these models informed by emerging research on gender-specific programming or managerial considerations unique to women offenders (Burke and Adams, 1991; LIS, Inc., 1998; Morash, Bynum, and Koons, 1998; Ryan, 1994).

The call for improving classification and programming for women predates the recent surge in numbers of women in prison. Women offenders, especially those who are incarcerated, have long been overlooked in practice and research due to economies of scale (there are fewer of them) and gender-based assumptions about the causes of their criminal behavior (Belknap, 1996; Dobash, Dobash, and Gutteridge, 1986; Pollock-Byrne, 1990; Rafter, 1990; Weisheit and Mahan, 1988). There is a need to be concerned about both the adequacy of programs for women offenders (Belknap, Dunn, and Holsinger, 1997; Covington and Bloom, 1999; Dembo, et al., 1992; Morash, Bynum, and Koons, 1998; Owen, 1998) and about whether women are being classified in valid and relevant ways.

The Prisons Division of the National Institute of Corrections (NIC) has assumed the task of addressing issues pertinent to women offenders through two initiatives:

¢ Classification of Women Offenders—two cooperative agreements2 designed to research and develop improved strategies for classifying women offenders.

¢ Gender Responsive Principles—a series of cooperative agreements with select- ed state and local jurisdictions endeavoring to improve programming and serv- ices to women offenders.

This report presents findings from a national assessment of state and federal prac- tices for classifying women offenders conducted under the auspices of NIC’s Classification of Women Offenders initiative. Classification models and strategies were discussed with correctional administrators and representatives from classifi- cation and research offices throughout the 50 states and the Federal Bureau of Prisons. A detailed description was elicited of current practices in use for classify-

Introduction

oneChapter

ing women offenders. Respondents provided a nationwide picture of the systemic and operational contexts of classification and how these differ for men and women. Respondents’ perceptions of whether their systems “worked” for women in direct- ing appropriate custody, programming, and housing decisions were noted, and the extent to which agencies found the classification needs of women offenders to be different from men was explored. Finally, inquiries were made about the psycho- metric quality of current systems: What were the origins of the system? Was it developed with women offenders in mind, or was it designed for men and subse- quently applied to women? Were the systems validated for women offenders?

As the following discussion illustrates, these issues were initially framed by a grow- ing body of literature on women offenders. Exposed to the viewpoints of those who work with women offenders on a daily basis, an additional perspective is offered on that literature.

Chapter 1

This assessment took place in the context of uncertainty regarding the role and sta- tus of objective classification3 systems for women in correctional settings, espe- cially in prisons. Two issues underlie this observation. First, the literature is not entirely clear about the proper purpose of a classification system for women offend- ers. Should women offenders be classified to ensure the security of women’s pris- ons or to inform treatment decisions; or should classification serve both purposes?4

Should classification systems inform housing assignments? Agencies can be expected to have different answers to these questions, or they might identify addi- tional purposes, but the answer is crucial to their approach to classification. Second, on the basis of available literature, the current generation of classification systems for women may not adequately serve any of these purposes.

Classifying Women for Security and Custody

Renewed concern about the security of women’s prisons seems reasonable in light of contemporary increases in the number of incarcerated females (Chesney-Lind, 1997; Gilliard and Beck, 1998; Kline, 1992). Adding to this concern is the fear that changes in the composition of prison populations (e.g., gangs, substance abusers, and younger offenders) might promote increases in prison violence. At the outset of this project, it was assumed that most existing prison classification systems for women would address women’s risk of committing aggressive and other discipli- nary infractions while incarcerated. In doing so, most systems would be identical to current systems for male offenders. In support of this assertion, a somewhat dated survey of state correctional agencies found that 40 of the 48 states surveyed used the same objective classification system for women as for men. In most instances this involved variations on an NIC prison classification model used for predicting disciplinary problems (Burke and Adams, 1991).5 A subsequent survey found few changes in this situation (Morash, Bynum, and Koons, 1998).6

Yet, in focusing on prison security, it must be recognized that correctional officials disagree about the level of violence found among female prison inmates, many sug- gesting that women’s aggression is not prevalent or serious enough to be a primary concern (Burke and Adams, 1991; Morash, Bynum, and Koons, 1998). Empirical support for these staff perceptions is suggested by some validation studies of security- based classification systems for women (e.g., Alexander and Humphrey, 19887;

Issues in Classifying Women Offenders: The Literature

twoChapter

Applying any type of

prediction instrument

to a population other

than the one used for

its construction and

validation is scientifi-

cally improper and

professionally unethi-

cal and can result in

inappropriate deci-

sions for the group

being (mis)diagnosed.

Hardyman,8 1999) and several studies conducted prior to the 1980s and reviewed by Bowker (1981). On the other hand, some researchers argue that women offenders are becoming more aggressive and problematic (Kruttschnitt and Krmpotich, 1990; Tischler and Marquart, 1989; Williams, 1981).9

Apart from the question of whether there are dysfunctional behaviors to predict is the question of which inmate attributes might predict such behaviors. Commonly used instruments (such as the NIC models) tap mostly static factors: history of vio- lence, severity of current offense, history of escapes, type of prior commitments, type of outstanding detainers, and such stability factors as age, education, and employment. A number of sources question the relevance of these factors to women’s behavior while in prison or to goals that administrators and policymakers set for women in prison (Burke and Adams, 1991).

Additional or alternative classification factors for women offenders might consider marital status and suicide attempts (Forcier, 1995); family structure of childhood home (Balthazar and Cook, 1984; Kruttschnitt and Krmpotich, 1990); childhood abuse, depression, and substance abuse (McClellan, Farabee, and Crouch, 1997);10

single parenting and reliance upon public assistance (Bonta, Pang, Wallace- Capretta, 1995); and relationships (Covington, 1998). More controversial is the inclusion of homosexual relationships in prison, which is based on the view that such relationships generate many acts of aggression.11 At the outset of the present assessment, it appeared that none of these variables had been incorporated into cor- rectional classification systems for women offenders.

More fundamental than the use of gender-responsive variables is the issue of whether existing classification models, even those constructed for men, had been validated and normed for women. Because few such validation studies were on record at the NIC Information Center, the question was raised in this assessment. Many correctional researchers and consultants have observed that agencies fail to validate correctional assessments to specific populations (Van Voorhis and Brown, 1996). Without such studies, questions regarding predictive validity and relevance of assessments to women offenders are difficult to answer. Moreover, applying any type of prediction instrument to a population other than the one used for its con- struction and validation is scientifically improper and professionally unethical (AACP, 2000; APA, 1992) and can result in inappropriate decisions for the group being (mis)diagnosed.

Far from being unique to corrections, failure to adequately study predictors that dif- ferentially affect the performance of men and women is found in education, mental health, and medicine. In many cases, women have been adversely affected. Effects have included less accurate college admission criteria for women than for men (Sternberg and Williams, 1997); inaccurate mental health assessments (Gilligan, 1993); less accurate information regarding women’s heart attacks and strokes (Arnstein, Buselli, and Rankin, 1996); and inaccurate understandings of drug

Chapter 2

Objective classifi-

cation of women

offenders is far from

an exact science.

dosages and side effects for women (Martin et al., 1998). These effects stem from the error of conducting most studies on men and prematurely generalizing the find- ings to women.

Another issue concerns whether security and custody classification systems are overclassifying female offenders, recommending unwarranted assignment to high- er security levels. Overclassification can occur in three ways. First, prediction instruments developed for populations that are known to evince low base rates12 on the criterion variable tend to produce more false-positive classification actions than those where variation on the criterion variable is more evenly distributed (Brennan, 1998; Clear, 1988). This is perhaps the most important cause of overclassification, because nothing can be done about it except to avoid predicting behaviors that char- acterize a very small proportion of a group. Predicting behaviors committed by an extreme minority creates a situation where excessive numbers of individuals must be held for the few who will commit the problem behaviors.

Second, even in cases where high risk is indicative of a greater likelihood of prob- lematic behavior than low risk, the meaning of seriousness is relative to the popu- lation (Van Voorhis and Brown, 1996). For example, high risk in a group with a low base rate could translate into 30 percent of the high-risk group committing a new offense, whereas high risk in a group with higher base rates (e.g., male prison inmates) could translate into 60 percent of the high-risk group committing a new offense. Simply put, high-risk females may be quite different from high-risk males, and the difference has clear implications for policy and practice.

Third, if the dependent or criterion variable captures different behaviors for men and women within the same disciplinary or offense criteria, one group or the other could have inflated reclassification levels. For example, if women tend to receive escape charges because of walking away from a nonsecure facility or returning late, while men receive them for escaping the perimeter of a secure facility, women’s scores will be inflated by less serious offenses. Similar results could be incurred in any situation where women get “the book thrown at them” and men do not.

In sum, objective classification of women offenders is far from an exact science. In contrast to parole, probation, and other community-based risk assessments, prison-based risk assessments only emerged and received wide use in the 1980s (Alexander and Austin, 1992). This generation of security and custody instru- ments, though designed to predict prison disciplinary infractions, is largely com- posed of static factors.13 Those factors that might render an instrument more useful for classifying women offenders have been discouraged for lack of research on women-specific predictors of institutional infractions; legal reasons, such as equal protection concerns (see Brennan, 1998; Burke and Adams, 1991);14 ongoing dis- agreements concerning whether there are enough dangerous behaviors to predict; and arguments concerning whether women should be classified according to risk (Stanko, 1997).

Issues in Classifying Women Offenders: The Literature

In a growing number

of community cor-

rectional agencies,

classification systems

simultaneously assess

risks and needs by

tapping the dynamic

individual characteris-

tics that predispose

one to criminal

behavior.

Internal Classification Systems

Even after inmates have been classified into institutional settings on the basis of risk, a second level of internal classification often occurs (Levinson, 1988). Such assessments are viewed as secondary to risk and custody systems and recognize that even when groups are separated according to risk, they remain highly heteroge- neous (Levinson, 1988). Choice of an appropriate internal classification instrument depends upon the type of institution, its goals, and the types of programs into which clients are screened. Some internal classification systems are used to inform hous- ing decisions within the institution and others to assign inmates to appropriate treat- ment interventions. Some systems serve both functions, as when a substance abuse assessment determines who will be placed in a therapeutic community. The most common types of classification systems for making housing decisions use crimi- nal personality typologies, primarily to separate predatory inmates from the popu- lation (e.g. Jesness, 1996; Megargee and Bohn, 1979; Quay, 1983; Van Voorhis, 1994; Warren et al., 1966). Personality-based systems for internal classification or for assigning inmates to housing units have been found to reduce prison infractions committed by males (e.g., Bohn, 1980). Few studies of this nature have been con- ducted on women offenders. However, one in South Carolina found a modified Quay Adult Internal Management System (AIMS) to be applicable to women offenders (Quay and Love, 1989). Extensive testing of the Megargee Minnesota Multiphasic Personality Inventor (MMPI)-based system has found the system, with modifications, to be applicable to women offenders (Megargee, 1997). It is noteworthy that attempts to validate the personality systems with women, particu- larly the Megargee MMPI-based typology and the Quay AIMS, resulted in modi- fications to the original systems (see Megargee, Mercer, and Carbonell, 1999; Megargee, 1997; Quay and Love, 1989).

Classification for Case Management and Treatment Purposes

Needs assessment systems are the most common internal classification systems. Needs assessments consist of either an objective system for classifying offenders according to major problem areas (e.g., education, mental health, employment, sub- stance abuse) or separate tests for key need areas, or both. Static custody classifi- cation systems have had some success predicting offending behaviors in institutions (Buchanan, Whitlow, and Austin, 1986; Alexander and Austin, 1992) and in com- munity settings (Hoffman, 1994; Baird, Prestine and Klockziem, 1989), but they cannot help case managers and other institutional officials make treatment-related decisions (Andrews and Bonta, 1998; Van Voorhis, 2000). Agencies that place a high priority on meaningful programming for correctional clients require addition- al needs-based classification tools.

Chapter 2

Though problematic

for both males and

females, abuse and

neglect seem to be

more predictive of

the future offending

of females than males.

In a growing number of community correctional agencies, classification systems simultaneously assess risks and needs by tapping the dynamic individual charac- teristics that predispose one to criminal behavior15 (Andrews, Bonta, and Hoge, 1990; Andrews and Bonta, 1998). Prison needs assessments separate the needs assessment function from the risk assessment function. Some argue that it is more important to keep the custody classification and the needs classification process separate in prison classification models than in community correctional models (Adams and Henning, 1982), lest agencies find themselves advancing custody according to one’s problems rather than one’s behavior. This debate is unresolved. Indeed, if those with mental health and substance abuse issues actually pose a greater threat to public safety, it is not clear that risk and needs assessment should be distinct.

Recognition of the importance of needs assessment has increased in recent years. Publications and conference programs of professional organizations—such as the American Correctional Association, the International Community Correctional Association, and the American Probation and Parole Association—as well as NIC activities, suggest an identifiable trend toward increased treatment programming in correctional agencies. In addition, recent surveys have found correctional officials and practitioners stressing the importance of programs responsive to the specific needs of women offenders, such as victimization counseling, nursery programs to enable mothers to keep babies near them while incarcerated, mentoring and self- esteem programs, gender-specific health education programs, gender-responsive substance abuse and mental health programs, and gender-specific cognitive skills programs (LIS, Inc., 1998; Morash, Bynum, and Koons, 1998).

We may be a long way from incorporating these gender-responsive needs into insti- tutional needs assessment models. Some would argue that the most important sources of prospective criminal behavior are the same for men and women. Research on this question is equivocal. Results of a meta-analysis of delinquency causation studies by Simourd and Andrews (1994) report that the most important risk factors are antisocial peers, a history of criminal behavior, antisocial attitudes, antisocial personality, family relationships, and school or employment problems. These were seen as gender-neutral. However, the authors concede that victimization and self-esteem require further research before they can be ruled out as predictors of female offending.

Though problematic for both males and females, abuse and neglect seem to be more predictive of the future offending of females than males (McClellan, Farabee, and Crouch, 1997; Rivera and Widom, 1990). This may be partly attributable to the fact that girls are more likely to be physically abused than boys.16 Furthermore, abuse itself introduces many girls to systems (e.g., school failure, unemployment, and juvenile court) that have their own effects on future offending (Belknap and Holsinger, 1998; Chesney-Lind and Shelden, 1992). In correctional settings for both girls and women, the proportion of inmates reporting past physical and sexual

Issues in Classifying Women Offenders: The Literature

Systems that are

valid for men but not

for women are not

equal, even though

the classification

factors are identical.

abuse is much higher than it is for males (Dembo et al., 1992; Holsinger, 1999; McClellan, Farabee, and Crouch, 1997; Morash, Bynum, and Koons, 1998; Snell, 1994). Finally, the causal paths between abuse and neglect and offending for females may be intertwined with anxiety, depression, emotional problems, self-esteem, and substance abuse (Miller et al., 1995; McClellan, Farabee, and Crouch, 1997).

From focus groups with both staff and inmates (e.g., see Belknap, Dunn, and Holsinger, 1997; Dembo et al., 1992; Holsinger, 1999; Miller et al., 1995; Morash, Bynum, and Koons, 1998) comes a clear call for interventions targeting physical and sexual abuse, self-esteem, relationships, and mental health. Women and girls appear to have different paths to substance abuse (Wanberg and Milkman, 1998). Their offending behavior is likely to involve relationship issues (Taylor, Gilligan, and Sullivan, 1995; Covington, 1998), and they are less likely than males to adhere to antisocial rationalizations for their behavior (Erez, 1988). Women offenders are far more likely to receive mental health diagnoses than men offenders.

Whether systems conduct separate needs assessments or incorporate criminogenic needs into risk and custody assessment, it makes sense to add variables pertinent to women’s needs. Assessments for specific or special needs provide more detailed pictures of offender needs and could assess substance abuse, mental health, intelli- gence, and educational background. The assessment tools should also be normed and validated on women.

If corrections is headed toward a gender-responsive approach to programming for women offenders, it may not have the optimal classification technology for doing so. An overview of items contained in the current generation of institutional classi- fication systems shows an absence of gender-specific, criminogenic needs probably because these systems were designed to facilitate custody rather than programming assignments. Given that risk and custody issues may be of less concern for women than men, it is not surprising to find widespread dissatisfaction with current systems for classifying women offenders.

Legal Impediments

Following years of litigation, case law, and legislation, male and female prison inmates now are mandated to be afforded equality in housing, access to legal services, programming, employee wages, and medical facilities. Classification has been interpreted as falling under this rubric. Systems may be viewed as gender- discriminatory if they contain different variables, assign offenders to different types of programs, and send people with similar characteristics to qualitatively different types of housing (Brennan, 1998). These concerns have prompted many adminis- trators to staunchly defend parity as exemplified by identical systems.

Chapter 2

Use of identical systems may offer a false sense of assurance. Indeed, systems that are valid for men but not for women are not equal, even though the classification factors are identical. Moreover, a system that recognizes men’s needs better than it does women’s is inherently unequal. Finally, if proportionally more maximum- custody men commit predatory acts than maximum-custody women, the classifica- tion system already lacks parity. In that instance, although the classification instruments look identical and procedures of administration are similar, they are disparate in terms of outcome. With this in mind, some agencies voice concern that not having separate systems may lead to litigation (Austin, Chah, and Elms, 1993). This was seen in the recent case Cain v. Michigan Department of Correctionswhich resulted in Michigan’s adoption of different classification cutoff scores for women.

This discussion has endeavored to frame a number of issues that may need to be addressed through development of systems for classifying women offenders. The literature identifies three crucial issues: validity, overclassification, and lack of gender-responsiveness. This assessment sought to understand the extent to which these issues typified current correctional practices and whether agencies had for- mulated approaches for resolving them. The remainder of this report details the national assessment of correctional officials and its results.

Issues in Classifying Women Offenders: The Literature

The issues and questions posed above were addressed through a telephone assess- ment of representatives from 50 state correctional agencies and the Federal Bureau of Prisons. The assessment proceeded with the following steps:

¢ Construction of a telephone inventory. ¢ Identification of respondents most knowledgeable about their state’s classifica-

tion procedures.

¢ Review of the results of the telephone discussions. ¢ Review of classification documents, where warranted. ¢ Collateral contacts, where warranted, with other state officials or consultants

providing classification-related services to the agency.

Development of Assessment Questions

The assessment consisted of both open-ended and closed-ended questions. Although the final version contained 73 questions, only sections applicable to the state were administered.17 Before any questions were written, University of Cincinnati project staff reviewed the most recent academic articles, books, and other literature on female offending and correctional classification. The survey was grounded on this review. NIC personnel also contributed to the instrument.

Identification of and Contact With Respondents

Identification of respondents with expert knowledge of agency classification proce- dures was of utmost concern. The following steps were followed:

¢ Letters were mailed to agency heads to announce the purpose and nature of the assessment, provide a copy of the survey questions, request designation of a staff member to serve as a respondent, and invite them to discuss any concerns about the assessment.

¢ Followup telephone calls were placed within 2 weeks to the office of the agency head requesting the name of a respondent.

Approach

threeChapter

¢ Telephone calls were made to the designated respondent to schedule an interview. ¢ The telephone interview was conducted. ¢ Followup calls were placed to commissioners or directors if no response to the

initial letter was received.

¢ Additional calls were placed to other state officials to address questions that could not be answered by the initial respondents.

Telephone Discussion/Assessments

The telephone discussions were conducted between February 29 and May 23, 2000. Generally, the telephone contacts afforded an economical way to secure the data while assuring a greater response rate and accuracy than a mail-in questionnaire (Maxfield and Babbie, 2000).

Respondents showed considerable expertise in the classification systems used by their states. Thirty-three of the respondents were state directors of classification, five were either wardens or administrators of women’s facilities, four were research analysts familiar with the system’s development or modification, one was a clinical director, and eight were serving in some other administrative capacity. Many of these other administrators had once served as a warden or had extensive experience working with classification systems for women offenders. When respondents lacked the expertise to answer questions, they referred us to other officials, or we contacted another respondent with greater expertise. Even so, questions were occa- sionally left unanswered. Some respondents, for example, had no knowledge of whether their classification methods had been validated for women offenders, espe- cially in instances where the validation study may have been conducted as many as 15 years ago. It was not unusual for respondents to identify as classification prob- lems issues that were not necessarily attributable to classification, such as needing more bed space, getting women closer to families, and providing more services.

Focus Groups

As noted earlier, the NIC cooperative agreement included technical assistance to three states. The technical assistance portion of the initiative was underway while this assessment was being conducted. Technical assistance visits involved separate focus groups with women offenders, staff, and administrators. Twelve focus groups were conducted, and findings from these meetings are included in this report where they lend additional understanding to the results of the assessment.

Chapter 3

Context of Classification

Correctional classification exists in an organizational and operational context. Classification practice is tied inextricably to the purpose of classification (Brennan, 1998; Van Voorhis, 1994, 2000), agency goals and philosophy, the behaviors and needs of the women being classified, and any organizational constraints that affect agency operations. As already indicated, a classification system designed to attend to custody concerns may look considerably different from one endeavoring to inform treatment and programming tasks (Van Voorhis, 2000) or to facilitate housing assignments (Levinson, 1988). For these reasons, a good portion of this assessment sought to understand the correctional goals and philosophies that classification models are intended to support, the nature of correctional environments, and the women being classified.

Goals and Philosophies

The correctional goals of deterrence, incapacitation, rehabilitation, retribution, restoration, and public safety are familiar to practitioners of correctional practice. Such terms are repeated in most textbooks of correctional practice, correctional mission statements, and agency reports. This assessment, however, sought to deter- mine the extent to which these goals might differ for male and female offenders.

The overwhelming majority of respondents noted security and public safety as the central purpose to be served by incarcerating women offenders. However, respon- dents in 15 states indicated that some goals are more central to women offenders, including—

¢ Habilitation and rehabilitation, particularly programs targeted to needs that are unique to women.

¢ Transitional programming pertinent to parenting and family issues. ¢ Moving women who have committed minor offenses to lower custody levels

and out of the system as soon as possible, to serve more women in community facilities rather than in institutions.

Not surprisingly, these respondents tended to want classification models that would better support gender-responsive programming and that would move women more quickly through the system.

Findings

fourChapter

Several respondents

faulted current

custody-centered

classification models

for overclassifying

women by assigning

too many women to

unnecessarily high

custody levels. One

response to the

problem of over-

classification is to

override selected

classification scores.

Several respondents faulted current custody-centered classification models for overclassifying women by assigning too many women to unnecessarily high cus- tody levels. One response to the problem of overclassification is to override select- ed classification scores. Representatives of 10 states indicated that it was necessary to routinely override more than 15 percent of their classification scores. Override rates ranged from 18 to 70 percent of these agencies’ classification scores.18 These agencies (20 percent of the total), in effect, indicated that their systems were not working for women offenders.

Are Women as Dangerous as Men?

A concern for moving women to lower custody and community levels as soon as possible accords with an underlying sense that women offenders do not pose the same security risks as men. When asked this question, 51 percent of the respondents reported either that women pose less risk than men or that a much smaller portion of women offenders than of men offenders pose a serious risk to institutional and public safety.19

Do Women Have Different Needs Than Men?

Responses to questions about women’s programming and treatment needs strongly echoed the emerging writings on gender-responsive programming. Almost all respondents (92 percent) asserted that women have unique needs that should be addressed in correctional settings. These needs included help in dealing with issues of trauma and abuse, self-esteem and assertiveness, vocational and job skills, med- ical care, mental health, parenting and child care, and relationships.

Organizational Context

Notwithstanding recent increases in the number of women housed in correctional facilities, their numbers relative to men remain small.20 Eighteen states house fewer than 500 women offenders in correctional institutions, while 13 states house more than 1,500 women. In Texas and California, this figure is close to 10,000 women offenders. Throughout the United States, the median number of women incarcerat- ed in each state is 851.

Given the observation that public safety and security is the primary concern of cor- rectional officials, it was somewhat surprising to learn that in 35 states women with different custody scores are housed together in at least one, if not all, of the state’s facilities for women. When women offenders and staff were asked if women were treated differently on the bases of their custody scores, the response was often, “No.”

Classification unfolds roughly as follows. On sentencing, women are assigned to a diagnostic and reception center (N = 40 states), not unlike male offenders. Remarkably, however, on receiving a custody score that may range from commu- nity or minimum custody, to close or maximum custody, women are often assigned to an institution where custody scores do not affect housing, privileges,

Chapter 4

To say that a system

was developed for

men and women,

however, is not the

same as saying that

the system contains

variables that are rele-

vant to women or

that it has been vali-

dated for women.

programming, or movement throughout the facility. Custody scores in these cases do affect whether a woman may work outside of the facility, the types of restraints required when transporting her outside of the institution, and whether she may move to a community placement.

Just the same, the era of all states operating just one facility for women appears to be coming to an end. Only 19 states house women in a single facility. Most states now maintain at least 2 institutions for women, and 21 states and the Federal Bureau of Prisons operate 3 or more facilities, not including community correc- tional facilities, prerelease centers, or hospitals.

Are Women and Men Classified Differently?

Although many respondents discussed clear differences between men and women offenders in terms of needs and risk to institutional and public safety, few states have incorporated these differences into their objective prison classification instruments.

Risk/Custody Assessment

All states consider the risk/custody assessment to be the foundation of their classi- fication approach. For 21 states, the custody instrument is a variation on a system developed by NIC that contains mostly static factors. Although 30 states report hav- ing some dynamic variables that could be viewed as relevant to offender needs, the custody tool typically is not viewed as relevant to programming and treatment issues.21 Most states reclassify offenders using an objective reclassification tool administered at intervals ranging from 3 to 6 months, although 14 states reclassify women only on an annual basis. Of course, for the many women who serve short sentences, annual reclassification means no reclassification.

Thirty-four states report using an objective tool to summarize offender needs. More specific tests and inventories are used to measure such single needs as education (34 states), substance abuse (36 states), and mental health (44 states). Internal classifi- cation specific to housing needs that assign offenders to housing units on the basis of personality, dangerousness, or other considerations were in used in 12 states.

Not surprisingly, most of the existing classification models were developed for men (4 states) or for men and women both (39 states). To say that a system was devel- oped for men and women, however, is not the same as saying that the system con- tains variables that are relevant to women or that it has been validated for women.

Twelve states have made some change to the way women are classified for purpos- es of custody and public safety. Four states (Idaho, New York, Massachusetts, and Ohio) have a separate custody classification system for women. Ohio’s classifica- tion model is different because a new system was developed for male offenders and had not been found to be valid for women. Consequently, Ohio continues to use its old system for women but is currently designing a new system for women.

Findings

Regardless of whether

a state’s custody clas-

sification system was

changed or adjusted,

only 14 states report-

ed validating their

custody classification

systems on a sample

of women offenders.

Four states have different cutoff scores for men and women. Changing the criteria for cutoff scores is an attempt to tie each custody level to a similar type of offend- er, whether male or female. If, for example, infraction rates for men in maximum custody are greater than that for women, raising the cutoff score for women may result in greater similarity between the groups.22 Two of the four states have changed variables to better reflect the nature of women’s disciplinary behavior in prison. For example, age is noted to be a predictor of prison disciplinary problems. However, in Delaware, women are seen as more likely than men to continue to act out with advancing years. As such, the state has increased the age cutoff for women. New York State reduces points on a common variable, seriousness of the current offense, for women who murdered an abuser.

The remaining four states (probably more)23 have expanded options on existing variables to accommodate the nature of women’s offending and infractions. Employment variables, for example, have been expanded in some instances to avoid classifying full-time homemakers or stay-at-home parents as unemployed. In addi- tion, escape variables have been modified to provide lower scores to offenders who walk away from community and other nonsecure settings, thus obviating the need to give scores similar to those who escape a secure perimeter.

Regardless of whether a state’s custody classification system was changed or adjusted, only 14 states reported validating their custody classification systems on a sample of women offenders. Another 11 states combined women and men in the same validation sample.24 Combined samples cannot be considered adequate vali- dation studies of women offenders because they often contain far fewer women than men. The resulting statistics are driven by men, the majority of the sample. Furthermore, if results for men and women are different, a combined sample is unlikely to identify those differences. This situation is changing rapidly. Indeed, 6 of these 14 validation studies occurred during the past year as part of NIC’s coop- erative agreement to validate and improve several state external classification sys- tems. Some of the recent validation studies found existing custody systems to be valid for women offenders. The valid systems contained a notable pattern. Key vari- ables had been expanded to include options that were more reflective of women’s lives, thereby reducing some problems with overclassification. This issue is dis- cussed in later sections of this report.

Needs Assessments

Another way to accommodate women offenders in correctional classification processes is to construct gender-responsive needs assessments. In spite of the fact that 49 respondents identified needs/problems that were unique to women offend- ers, only 8 states have made some provisions to assess the needs of men and women differently. These include four states with separate objective needs assessment instruments for women; one state that has altered the interview format for women to secure more accurate information; one state that conducts a psychosocial assess- ment for women and not for men; one state with a separate needs assessment for some men; one state that conducts a separate vocational battery at its largest

Chapter 4

Only eight states

have made some pro-

visions to assess the

needs of men and

women differently.

facility for women; and one state that has expanded the needs assessment to include gender-responsive needs for both men and women.

In none of the states were gender-responsive needs tapped through a test or inventory of a specific need, such as trauma. Although 17 agencies report using classification systems to guide housing assignments, none reported using gender- responsive systems.

Innovative Practices Currently in Use

The assessment revealed that 14 correctional agencies assess men and women dif- ferently. The nature of these differences is unique to each agency and there is no standardized classification model that is perceived to fit the unique needs of women offenders. Some states, in fact, have made changes that do not address the issues and problems encountered in classifying and managing women offenders. Only 11 states and the Federal Bureau of Prisons (BOP) have actively sought to address problems of overclassification, poor validity, and failure to consider gender- responsive needs.

Arizona. The Arizona Department of Corrections uses a Correctional Classification Profile (CCP) to determine inmate risk and custody level and summarize needs. Risk and need factors are distinct parts of the instrument and the two sets of factors are integrated only to the extent that risk factors constrain programming indicated by needs. The CCP was purchased from Correctional Services Group, Inc., a pri- vate firm based in Kansas City, Missouri. Factors are scored somewhat differently for men and women. For example, escape history is positively related to determined custody level for both men and women. But escape from a secure facility during the past 3 years is coded the same for men (with 5 points) as escape during the past 2 years is for women. That is, the agency examines a greater expanse of time for men offenders. While this is a relatively minor change, it acknowledges women’s short- er sentences and perhaps the lesser severity of most escapes committed by women.

Idaho. The Idaho Department of Correction is one of only four state correctional agencies that assesses women’s custody levels using an instrument uniquely designed for women. This instrument was developed in 1993 with NIC consultants. Before 1993, men and women had been classified using the same assessment form. The men’s system, however, was believed to be overclassifying women. Idaho has also developed a gender-responsive needs assessment tool. The systems have recently been revalidated.

Illinois. For the Illinois Department of Corrections, scored security point values are slightly different for men and women because more minimum security beds are available to women. Minimum security is achieved by a wider score interval for women (zero to three), than for men (zero to two). Medium security is achieved by four to five points for women and three to five points for men. Maximum security scoring is the same for both sexes.

Findings

Massachusetts.Before 1994, the Massachusetts Department of Correction used a narrative-type classification summary to determine appropriate inmate security levels. The same summary was used for men and women. The Objective Point- Based Classification Form for men was developed in 1994 and implemented in 1997. A form developed and validated for women offenders was implemented in 1998. A recent revalidation study of the classification form for women suggested that its validity may be compromised by staff reaction to the variables contained on the form. Interviews with staff revealed a lack of faith in the variables, which can result in the form being inaccurately or unreliably completed, and the criticism that too few dynamic variables relative to static variables are on the form, result- ing in little opportunity for one’s custody level to change.

Michigan. Faced with a lawsuit (Cain v. Michigan Department of Corrections) alleging overclassification of women, the Michigan Department of Corrections revised its security classification screen as of March 31, 2000. It made changes relating to custody cutoff scores and weights on specific variables. Women, for example, receive more points for positive conduct than men. Women’s security levels are downgraded due to the revision.

New York. The security classification guidelines for women are similar to those for men, except for the inclusion of an “isolated personal violence factor.” That factor is a single violent act—either the instant offense or a past offense—that arose from an abusive personal relationship. If a woman’s criminal history includes this sort of isolated violence, a point is subtracted from the risk score. It is believed that, on that basis, the individual poses less risk to the public. No such factor is included on the men’s instrument, although classification administrators are not opposed to adding one in the future. In addition to the isolated personal violence factor, the men’s and women’s security classification guideline forms (both initial and reclassification) are scored differently. The cutoff scores were developed in such a way that women’s custody levels are reduced relative to those of men.

Pennsylvania.As in Arizona, the Pennsylvania Department of Corrections uses an aggregated tool, the Pennsylvania Additive Classification Tool (PACT), which encompasses risk/initial classification and needs assessment. These assessments are not integrated. Reclassification (of risk) is included in PACT as well. All three PACT assessments are identical for men and women. Interestingly, criteria used to score risk/classification and reclassification were gender-differentiated until 1996. A vali- dation study conducted in that year suggested that women were being overclassified. In response, the scoring criteria were made consistent for men and women.

Virginia. Different security levels are available to men and women under the juris- diction of the Virginia Department of Corrections. Men are classified across six security levels; women are classified across three. As a result, the same risk assess- ment forms are scored differently for men and women. Women who score 14 points or higher—which would distinguish men as levels 3, 4, 5, or 6—are classified as level 3 only.

Chapter 4

West Virginia. The West Virginia Division of Corrections assesses men and women identically. Two consecutive assessments take place for all inmates, using the same forms. These classifications, which influence custody level and risk assessment, determine whether an inmate may transfer to a less secure facility within the same custody level. What distinguishes women from men is that the custody level (hence classification) does not generally influence placement at one of two facilities for women. The exception is level 5 (maximum security) women, who are placed at the more secure Ohio County Correctional Complex. Women classified at levels 1 through 4 may be placed at either facility.

Wyoming. The Wyoming Department of Corrections is unique in that needs assess- ment, but not custody assessment, is gender specific. A recent study revealed that Wyoming’s custody classification instrument is valid for both men and women. By contrast, an Assessment of Inmate Needs was developed solely for women offend- ers. It scores levels of needs in health, intellectual ability, behavioral/emotional problems, substance abuse, education, work experience, motivation to change, sex- ual abuse history, domestic life, and life skills. The data for this assessment are culled from a psychosocial interview.

Federal Bureau of Prisons.BOP uses the Inmate Load and Security Designation Form to assess inmates’ initial custody level and the Custody Classification Form to conduct its periodic (at 6 or 12 months, depending on release date) reclassifica- tion. The form is identical for men and women but the scoring criteria differ, with the overall effect being to downgrade women’s custody levels relative to men’s. In this way, women in the highest custody category evidence infraction rates that are more similar to the rate for men.

Innovative Practices Currently Under Consideration

Some states participating in the NIC cooperative agreements are evaluating changes to their classification models. These changes are being studied and will not be recommended for implementation until they have been validated and found to be effective.

Change to Specific Variables

Focus group discussions with staff and women offenders have named specific vari- ables included in current custody classification systems as having contributed to the overclassification of women offenders. There is a pattern and a consensus to these discussions. The same variables (discussed below) are held to be problematic across focus groups.

Seriousness of the current offense and women who killed an abuser.The majority of women who commit a one-time offense in response to a sustained peri- od of abuse are not seen as violent by women offenders or staff. Yet, in most cur- rent classification models, “seriousness of the current offense” is the mainstay of

Findings

On many systems,

women who have

been engaged in

full-time parenting

are scored as unem-

ployed, thereby inflat-

ing their custody

score when parenting

may suggest stability.

the system. It is a heavily weighted variable that can keep offenders convicted of homicide in high-custody settings for extended periods. Support is growing for adoption of a policy similar to New York’s, where points are subtracted from the variable if the offender committed a violent offense against an abuser.

Variables expanded to include options pertinent to women.Overclassification sometimes occurs because such key variables as employment and escape history do not proffer attributes that speak to women’s lives. On many systems, women who have been engaged in full-time parenting are scored as unemployed, thereby inflat- ing their custody score when parenting may suggest stability. Another example involves women’s escapes. Fewer women than men escape a secure perimeter. However, women are perceived to be more likely than men to walk away from a community residential setting to visit children, family, or friends. In some states, many higher custody offenders are women who walked away, returned, and received additional points on the heavily weighted escape variables. Walking away was counted again on another variable pertaining to disciplinary infractions. Differentiating according to type of escape should correct this problem.

Timeframes on reclassification variables.Prison misconduct drives up reclassifi- cation scores in most states. In many cases, the misconduct drops from the reclas- sification score after 2 or 3 years. For the many women who serve short sentences, misconduct inflates reclassification scores and there is no time to reduce them. In effect, misconduct is a static variable for most women and a dynamic variable for most men in prison. Altering timeframes for assessment of misconduct can help correct this discrepancy.

Staff training. Staff often view women as more difficult to supervise than men, reacting to women’s different ways of problem solving, relating to staff, and doing time. Women in prison ask more questions, question authority, want to discuss things, and challenge decisions. Staff who are inexperienced with these differences become irritated and more likely to write up the inmates in an effort to better con- trol their behavior. At the same time, smaller facilities can create a situation in which behavior is more easily noticed and more likely to result in a write up. Both staff and inmates reported situations in which custody scores were driven up by rather minor misconduct reports. Staff training and changes to disciplinary vari- ables and policies could help to reduce this problem.

Gender-responsive needs.The literature and focus groups concur that women offenders represent a more troubled group than male offenders. Rates of mental ill- ness, relationship issues, physical abuse, and sexual abuse far surpass those for men (e.g., see Belknap, Dunn, and Holsinger, 1997; Dembo et. al, 1992; Holsinger, 1999; Miller et al., 1995; Morash, Bynum, and Koons, 1998). It is essential that women’s needs assessments adequately capture needs that are unique to them. Key to this effort is the inclusion of items that identify problems that cannot always be identified through central files or intake interviews. Nonetheless, agencies could also consider incorporating these items into the custody classification model if

Chapter 4

doing so improves the validity of the custody classification and if high custody on the classification system translates into a medium security facility or one in which intensive mental health, substance abuse, and other services are provided.

Use of Community Risk Assessment Instruments Rather Than Custody Classification Instruments

What sense is to be made of those states that administer and score custody classifi- cation systems and then do not use the systems? One such state reported that, although it did not use custody scores to facilitate housing assignments, it differen- tiated between inmates who would not be allowed to leave the perimeter and those who could safely be placed in community settings. Such decisions are related to public safety and community risk, rather than to institutional risk. Additionally, actual time served by most women in this state was less than one year, in which case an agency might be better served by a community correctional risk assessment instrument. A community risk instrument would be more relevant to concerns sur- rounding an offender’s risk while in the community. Some have also been shown to predict institutional infractions (Bonta, 1989; Bonta and Motiuk, 1992).

Findings

Most states continue

to use identical sys-

tems for men and

women, and most

states have no empiri-

cal basis for knowing

whether the system

is offering accurate

classifications of

women offenders.

At the outset of this report, three areas of concern regarding the classification of women offenders were outlined. First, many systems appear to overclassify women by housing them in unjustifiably high security-level facilities. Second, few current classification systems were developed with women in mind and many may not fit the primary goals set for women offenders. It is not at all unusual to hear, for example, that the current generation of classification systems for women does very little to facilitate meaningful programming. Third, systems may be invalid or inaccurate for women. Unfortunately, most have not been validated on samples of women offenders.

In sum, our survey and our focus groups paint a picture of little change since Burke and Adams (1991) conducted a similar survey 9 years ago. Most states continue to use identical systems for men and women, and most states have no empirical basis for knowing whether the system is offering accurate classifications of women offenders. Almost all of the respondents claim that women’s needs are different from men’s needs, but only eight agencies have incorporated these differences into needs assessments. The number of states making improvements has not changed significantly since the earlier survey. In fact, two states that had unique classifica- tion systems for women have changed back to a system that is uniform for men and women. Validation studies did direct a few states to make changes that improved classification of women offenders. Some established different cutoff scores, which help to ensure that men and women at the same custody levels are more similar than not.25 Other states have broadened attributes of key classification variables, thus increasing the validity of the system for women offenders.

Just the same, few states have designed systems that started with women in mind. Most map existing male-based assumptions regarding the goals and purposes of corrections onto women and the systems that classify them. The gaps are apparent. For example, what sense is to be made of the fact that most states do not use the notion of custody in the same way with women as they do with men? Many states do not base housing decisions on custody. Many respondents do not consider women to be as dangerous as men, and a common observation of focus groups par- ticipants is that women with high custody scores are not more dangerous than women with low custody scores. The notion of custody often does not carry over to prison architecture. Many women’s maximum security facilities do not look like maximum security facilities for men. It is hard to avoid the observation that in prac- tices involving women offenders, custody and risk seems a very confused concept.

Discussion

fiveChapter

If we started with women, we might expect to see classification systems that focused more attention on factors that seem key to women’s reintegration: their children, relationships, abuse, earlier trauma, mental illness, and job skills. We found only a few isolated attempts to move in this direction.

Chapter 5

Notes

1. The number of women incarcerated in America has increased at a faster rate than for men for the past several years. Since 1990, the annual rate of growth for incar- cerated women has averaged 8.5 percent, as opposed to 6.6 percent for men. Even so, the proportion of women incarcerated (57 per 100,000) remains much lower than the proportion of men incarcerated (885 per 100,000) (Mumola and Beck, 1997).

2. These cooperative agreements are administered by Patricia Van Voorhis at the University of Cincinnati and Patricia Hardyman at the Institute on Crime, Justice, and Corrections at The George Washington University.

3. We differentiate objective, actuarial classification systems from subjective approaches involving clinical judgment of correctional practitioners. As with other areas of human assessment, objective approaches have been found to be more valid and reliable measures of human behavior (e.g., see Gottfredson, 1987; Meehl, 1954).

4. The most recent generation of correctional classification systems perform both functions (Bonta, 1996).

5. The earlier NIC classification models contained mostly static variables (e.g., his- tory of institutional violence, severity of current and prior convictions, escape his- tory, current or pending detainers, prior felonies, substance abuse, assaultive behavior, and age). Later NIC versions added more dynamic variables (e.g., educa- tion, employment, and performance in treatment programs).

6. This survey found 39 states using the same classification system for both men and women.

7. This was a memo submitted to the authors on December 5, 1995. The study found that 60 percent of Ohio’s female inmates had no reports for any kind of dis- ciplinary infractions.

8. This study found that 60 percent of women incarcerated in New York State had no disciplinary reports; only 33 percent of the incarcerated men had no reports. Moreover, the offenses committed by women were less serious than those commit- ted by men (Alexander and Humphrey, 1988).

9. This discrepancy across studies may also be attributable to the use of official prison data (e.g., disciplinary infractions, tickets, etc.) as a measure of the types of behaviors we wish to prevent. Such measures are known to be affected by offi- cial policies regarding disciplinary infractions, the visibility of such infractions, staff initiative, and other organizational considerations (Light, 1990; Tischler and Marquart, 1989; Hewitt, Poole, and Regoli, 1984; Van Voorhis, 1994). In a sample of female prison inmates in England, for example, Dobash, and Gutteridge (1986) found that women were more readily cited for infractions than men. In addition, Van Voorhis’s study of 390 federal prison inmates reports that the choice of criteri- on measure (e.g., self-report, official disciplinary reports, and staff observational

measures) greatly altered estimates of the level of disciplinary involvement (Van Voorhis, 1994). In addition, the importance of the predictors of disciplinary infrac- tions (the classification variables themselves) vary somewhat according to the type of behavioral outcome measure used.

10. A study by Bonta, Pang, and Wallace-Capretta (1995) questions the role of abuse as a predictor of future offending.

11. As noted by Tischler and Marquart (1989), women’s aggression in prison set- tings often involves acts committed against other inmates, whereas male inmates are more likely to attack prison staff. Historically, research on violence among incarcerated women emphasized homosexuality and pseudofamilies (e.g., see Ward and Kasselbaum, 1965; Giallombardo, 1966; Bowker, 1981). More recent research suggests that violence associated with relationship-based jealousy varies tremen- dously across institutional environments (Propper, 1981).

12. Low base rate refers to a rate of occurrence that is typically below 10 percent.

13. The preference for such instruments may reflect the desire for data more easily obtained from prison records and management information systems, rather than less accessible data obtained from interviews or paper-and-pencil tests.

14. Burke and Adams (1991) attach a legal analysis conducted by Nicholas and Loeb (1991) that effectively dispels these concerns.

15. In other words, criminogenic needs are offender characteristics that should be targeted for interventions, are predictors of criminal behavior, and are likely to reduce future offending behavior if they are successfully addressed (Andrews and Bonta, 1998). Of course, there may be noncriminogenic needs that should be addressed for a variety of additional reasons (e.g., housing and health), but address- ing the latter does not reduce future offending (Van Voorhis, Cullen, and Applegate, 1995).

16. See Miller, et al. (1995). Such abuse and neglect also appears to start at an ear- lier age and to last longer for girls.

17. If the state had not instituted procedures to accommodate women offenders, for example, many assessment questions would be inapplicable. Also, sections on clas- sification for responsiveness or housing assignments were unnecessary unless the state actually reported using such systems.

18. A classification system loses predictive validity once the proportion of overrides surpasses 15 percent. The higher the proportion of overrides, the closer classifica- tion comes to being subjective and discretionary, the override is essentially a dis- cretionary decision.

Notes

19. This question was added after beginning the assessments; we had already con- ducted 14 interviews. The question was added because respondents in the earlier interviews and focus groups had raised the issue spontaneously. Although it seemed like a relevant question to add to subsequent interviews, we did not wish to further inconvenience earlier respondents by recontacting them.

20. This too is growing. In 1990, women composed 5.7 percent of all incarcerated adults. By 1998, they constituted 6.5 percent of all incarcerates (Beck and Mumola, 1999).

21. In contrast, many community correctional classification agencies are moving to a dynamic risk instrument that combines dynamic criminogenic needs (as predic- tors) with static risk factors (Van Voorhis, 2000).

22. This option does not solve all problems of this nature. If, for example, a vali- dation study results in reducing the maximum security category to an extremely small number of women offenders who have the same misconduct rates as a larger group of maximum security males, the resulting women’s system would be far less stable than the system for men. Predictive accuracy is likely to decrease in subse- quent efforts to revalidate the women’s system. In practice, the women’s system is likely to result in high numbers of false-positive classification.

23. Respondents may not have identified modifications to their classification sys- tems as beneficial to women.

24. Respondents for two additional states did not know whether their systems had been validated for either men or women.

25. With low base rates, changing cutoff scores often does not solve the problems. Reflecting the inadequacy of statistical methods to predict small numbers, or out- liers, a revalidation of the new cutoffs is likely to show the original problem. Very few individuals in the high-risk group commit a predatory offense relative to those who are held in anticipation of such an offense.

Classification of Women Offenders

Adams, L. and J. Henning. 1982. “Illinois Adult Classification System Design.” In ACA (ed.) Classification As A Management Tool: Theories and Models for Decision-Makers.College Park, MD: American Correctional Association.

Alexander, J. and J. Austin. 1992. Handbook for Evaluating Objective Prison Classification Systems.Washington, DC: National Institute of Corrections.

Alexander, J. and E. Humphrey. 1988. Initial Security Classification Guidelines for Females: Working Paper XVI.Albany, NY: New York State Department of Correctional Services.

American Association of Correctional Psychologists (Standards Committee). 2000. “Standards for Psychology Services in Jails, Prisons, Correctional Facilities and Agencies.”Criminal Justice and Behavior,27(4): 433–493.

American Psychological Association. 1992. “Ethical Principles of Psychologists and Code of Conduct.”American Psychologist,December.

Andrews, D. and J. Bonta. 1995. The Level of Supervision Inventory-Revised (LSI–R). North Tonawanda, NY: Multi-Health Systems.

Andrews, D. and J. Bonta. 1998. The Psychology of Criminal Conduct(2d ed.). Cincinnati, OH: Anderson Publishing Company.

Andrews, D., J. Bonta, and R. Hoge. 1990. “Classification for Effective Rehabilita- tion: Rediscovering Psychology.”Criminal Justice and Behavior,17: 19–52.

Arnstein, P., E. Buselli, and S. Rankin. 1996. “Women and Heart Attacks: Prevention, Diagnosis, and Care.”Nurse Practitioner,21(5): 57–69.

Austin, J., L. Chan, and W. Elms. 1993. Indiana Department of Corrections: Women Classification Study.San Francisco, CA: National Council on Crime and Delinquency.

Baird, C., R. Prestine, and B. Klockziem. 1989. Revalidation of the Wisconsin Probation/Parole Classification System.Madison, WI: National Council on Crime and Delinquency.

Balthazar, M. and S. Cook. 1984. “An Analysis of the Factors Related to the Rate of Violent Crimes Committed by Incarcerated Female Delinquents.” In S. Chaneles (ed.), Gender Issues, Sex Offenses, and Criminal Justice: Current Trends. New York, NY: Hayworth Press.

Beck, A. and C. Mumola. 1999. Prisoners in 1998.Bureau of Justice Statistics Bulletin. Washington, DC: Bureau of Justice Statistics.

References

Belknap, J. 1996. The Invisible Woman: Gender, Crime and Justice. Belmont, CA: Wadsworth.

Belknap, J., and K. Holsinger. 1998. “An Overview of Delinquent Girls: How Theory and Practice Have Failed and the Need for Innovative Changes.” In R. Zaplin (ed.),Female Crime and Delinquency: Critical Perspectives and Effective Interventions. Gaithersburg, MD: Aspen Publishing, Inc.

Belknap, J., M. Dunn, and K. Holsinger. 1997. Moving Forward Juvenile Justice and Youth-Serving Systems That Address the Distinct Experience of the Adolescent Female: Gender-Spesific Services Work Group to the Governor.Columbus, OH: Office of Criminal Justice Services.

Bohn, M. 1980. “Inmate Classification and the Reduction of Institution Violence.” Corrections Today,44(4): 8ff.

Bonta, J. 1989. “Native Inmates: Institutional Response, Risk, and Needs.” Canadian Journal of Criminology,31: 49–62.

Bonta, J. 1996. “Risk-needs Assessment and Treatment.” In A.T. Harland (ed.), Choosing Correctional Options that Work: Defining the Demand and Evaluating the Supply.Thousand Oaks, CA: Sage.

Bonta, J., and L. Motiuk. 1992. “Inmate Classification.”Journal of Criminal Justice,20: 343–353.

Bonta, J., B. Pang, and S. Wallace-Capretta. 1995. “Predictors of Recidivism Among Incarcerated Female Offenders.”The Prison Journal,75: 277–294.

Bowker, L. 1981. “Gender Differences in Prison Subcultures.” In L. Bowker (ed.) Women and Crime in America.New York, NY: Macmillan.

Brennan, T. 1998. “Institutional Classification of Females: Problems and Some Proposals for Reform.” In R. Zaplin (ed.),Female Crime and Delinquency: Critical Perspectives and Effective Interventions.Gaithersburg, MD: Aspen Publishing, Inc.

Buchanan, R., K. Whitlow, and J. Austin. 1986. “National Evaluation of Objective Prison Classification Systems: The Current State of the Art.”Crime and Delinquency,32: 272–290.

Burke, P. and L. Adams. 1991. Classification of Women Offenders In State Correctional Facilities: A Handbook for Practitioners. Washington, DC: National Institute of Corrections.

Chesney-Lind, M. 1997. The Female Offender: Girls, Women, and Crime. Thousand Oaks, CA: Sage Publications.

References

Chesney-Lind, M. and R. Shelden. 1992. Girls, Delinquency, and Juvenile Justice. Pacific Grove, CA: Brooks/Cole.

Clear, S. 1988. “Statistical Prediction in Corrections.”Research in Corrections,1: 1–39.

Covington, S. 1998. “The Relational Theory of Women’s Psychological Development: Implications for the Criminal Justice System.” In R. Zaplin (ed.), Female Crime and Delinquency: Critical Perspectives and Effective Interventions. Gaithersburg, MD: Aspen Publishing, Inc.

Covington, S., and B. Bloom. 1999. “Gender-Responsive Programming and Evaluation for Females in the Criminal Justice System: A Shift from What Works? to What is the Work?” Paper presented at the 51st Annual Meeting of the American Society of Criminology. November 17–20. Toronto, Canada.

Dembo, R., L. Williams, W. Wothke, J. Schmeidler, and R. Brown. 1992. “The Generality of Deviance Replication of a Structural Model Among High-Risk Youth.” Journal of Research in Crime and Delinquency,29(2): 200–216.

Dobash, R., R. Dobash, and S. Gutteridge. 1986. The Imprisonment of Women. Totowa, NJ: Blackwell.

Erez, E. 1988. “Myth of the New Female Offender: Some Evidence from Attitudes Toward Law and Justice.”Journal of Criminal Justice,16(6): 499–509.

Forcier, M. 1995. Massachusetts Department of Correction Female Offender Objective Classification Technical Assistance Project: Final Report. Washington, DC: National Institute of Corrections.

Giallombardo, R. 1966. Society of Women: A Study of a Women’s Prison.New York, NY: Wiley.

Gilliard, D. and A. Beck. 1998. Prison and Jail Inmates at Midyear 1997. Washington, DC: Bureau of Justice Statistics.

Gilligan, C. 1993. In a Different Voice: Psychological Theory and Women’s Development.Cambridge, MA: Harvard University Press.

Gottfredson, D. 1987. “Prediction: An Overview of Selected Methodological Issues.” In Gottfredson, D. and M. Tonry (eds.) Prediction and Classification: Criminal Justice Decision Making. Chicago, IL: University of Chicago Press.

Hardyman, P. 1999. Wyoming Department of Corrections Prison Objective Classification System: Final Report on 1999 Revalidation Effort. Washington, DC: Institute on Crime, Justice, and Corrections.

Classification of Women Offenders

Hewitt, J., E. Poole, and R. Regoli. 1984. “Self-reported and Observed Rule Breaking in Prison: A Look at Disciplinary Response.”Justice Quarterly,1: 437–447.

Hoffman, P. 1994. “Twenty Years of Operational Use of a Risk Prediction Instrument: The United States Parole Commission’s Salient Factor Score.”Journal of Criminal Justice,22: 477–494.

Holsinger, K. 1999. “Addressing the Distinct Experience of the Adolescent Female: Explaining Delinquency and Examining the Juvenile Justice System.” Unpublished Doctoral Dissertation. Cincinnati, OH: University of Cincinnati.

Jesness, C. 1996. The Jesness Inventory Manual. North Tonawanda, NY: Multi- Health Systems.

Kline, S. 1992. “Profile of Female Offenders in the Federal Bureau of Prisons.” Federal Prisons Journal,3(1): 33–36.

Kruttschnitt, C. and S. Krmpotich. 1990. “Aggressive Behavior Among Female Inmates: An Exploratory Study.”Justice Quarterly,7(2): 371–389.

Levinson, R. 1988. “Development in the Classification Process.”Criminal Justice and Behavior,15: 24–38.

Light, S. 1990. “Measurement Error in Official Statistics: Prison Rule Infraction Data.”Federal Probation,54(4): 63–68.

LIS, Inc. 1998. Current Issues in the Operation of Women’s Prisons.Boulder, CO: National Institute of Corrections.

Martin, R., P. Biswas, S. Freemantle, G. Pearce, and R. Mann. 1998. “Age and Sex Distribution of Suspected Adverse Drug Reaction to Newly Marketed Drugs in General Practice in England: Analysis of 48 Cohort Studies.”British Journal of Clinical Pharmacology,46(5): 505–511.

Maxfield, M. and E. Babbie. 2000. Research Methods for Criminal Justice and Criminology.Belmont, CA: Wadsworth.

McClellan, D., D. Farabee, and B. Crouch. 1997. “Early Victimization, Drug Use, and Criminality.”Criminal Justice and Behavior,24: 455–476.

Meehl, P. 1954.Clinical v. Statistical Prediction.Northvale, NJ: Jason Aronson, Inc.

Megargee, E. 1997. “Using the Megargee MMPI-Based Classification System with the MMPI-2s of Female Prison Inmates.”Psychological Assessments,9(2): 75–82.

References

Megargee, E. and M. Bohn. 1979. Classifying Criminal Offenders: A New System Based on the MMPI.Beverly Hills, CA: Sage.

Megargee, E., S. Mercer, and J. Carbonell. 1999. “MMPI-2 with Male and Female State and Federal Prison Inmates.”Psychological Assessments,11(2): 177–185.

Miller, D., D. Trapani, K. Fejes-Mendoza, C. Eggleston, and R. Dwiggins. 1995. “Adolescent Female Offenders: Unique Considerations.”Adolescence,30: 429–435.

Morash, M., T. Bynum, and B. Koons. 1998. Women Offenders: Programming Needs and Promising Approaches.Washington, DC: National Institute of Justice.

Mumola, C. and A. Beck. 1997. Prisoners in 1996:Bulletin. Washington, DC: Bureau of Justice Statistics.

Nicholas, S. and A. Loeb. 1991. Legal Analysis. In Burke, P. and I. Adams (eds.). Classification of Women Offenders in State Correctional Facilities: A Handbook for Practitioners.Washington, DC: National Institute of Corrections.

Owen, B. 1998. In the Mix: Struggle and Survival in a Women’s Prison.Albany, NY: SUNY Press.

Pollock-Byrne, J. 1990. Women, Prisons, and Crime.Pacific Grove, CA: Brooks/Cole.

Propper, A. 1981. Prison Homosexuality—Myth and Reality.Lexington, MA: DC Heath.

Quay, H. and C. Love. 1989. Behavioral Classification for Female Offenders. Washington, DC: National Institute of Corrections.

Quay, H. 1983. Technical Manual for the Behavioral Classification System for Adult Offenders.Washington, DC: U.S. Department of Justice.

Rafter, N. 1990. Partial Justice: Women, Prison, and Social Control.New Brunswick, NJ: Transaction Books.

Rivera, B. and C. Widom. 1990. “Childhood Victimization and Violent Offending.” Violence and Victims 5(1): 19–35.

Ryan, T. 1994. Adult Female Offenders and the Institutional Programs: A State of the Art Analysis.Washington, DC: Department of Corrections.

Simourd, D. and D. Andrews. 1994. “Correlates of Delinquency: A Look at Gender Differences.”Forum on Corrections Research,6: 26–31.

Classification of Women Offenders

Snell, T. 1994. Women in Prison.Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics.

Stanko, E. 1997. “Safety Talk: Conceptualizing Women’s Risk Assessment as a ‘Technology of the Soul.’”Theoretical Criminology,1(4): 479–499.

Sternberg, R. and W. Williams. 1997. “Does the GRE Predict Meaningful Success in the Graduate Training of Psychologists?”American Psychologist,52(6): 630–641.

Taylor, J., C. Gilligan, and A. Sullivan. 1995. Between Voice and Silence: Women and Girls, Race and Relationship.Cambridge, MA: Harvard University Press.

Tischler, C. and J. Marquart. 1989. “Analysis of Disciplinary Infraction Rates Among Female and Male Inmates.”Journal of Criminal Justice,17: 507–513.

Van Voorhis, P. 1994. Psychological Classification of the Adult Male Prison Inmate. Albany, NY: SUNY Press.

Van Voorhis, P. 2000. “An Overview of Offender Classification Systems.” In Van Voorhis, P., M. Braswell, and D. Lester (eds.),Correctional Counseling and Rehabili- tation,2d edition. Cincinnati, OH: Anderson Press.

Van Voorhis, P. and K. Brown. 1996. Risk Classification in the 1990s.Washington, DC: National Institute of Corrections.

Van Voorhis, P., F. Cullen, and B. Applegate. 1995. “Evaluating Interventions With Violent Offenders: A Guide for Practitioners and Policymakers.”Federal Probation,59: 17–28.

Wanberg, K. and H. Milkman. 1998. Criminal Conduct and Substance Abuse Treat- ment: Strategies for Self-Improvement and Change.Thousand Oaks, CA: Sage.

Ward, D. and G. Kasselbaum. 1965. Women’s Prisons: Sex and Social Structure, Chicago, IL: Aldine.

Warren, M. and the Staff of the Community Treatment Project. 1966. Interpersonal Maturity Level Classification: Diagnosis and Treatment of Low, Middle, and High Maturity Delinquents. Sacramento, CA: California Youth Authority.

Weisheit, R. and S. Mahan. 1988. Women, Crime, and Criminal Justice.Cincinnati, OH: Anderson.

Williams, T. 1981. “Sentencing Time in Corrections and Parole.” In Figueira- McDonough, J., A. Inglehart, R. Sarri, and T. Williams (eds.),Women in Prison: Michigan 1968-1978.Ann Arbor, MI: University of Michigan School of Social Work and the Institute for Social Research.

References

User Feedback Form

Please complete and return this form to assist the National Institute of Corrections in assessing the value and utility of its publications. Detach from the document and mail to:

Publications Feedback National Institute of Corrections 320 First Street N.W. Washington, DC 20534

1. What is your general reaction to this document?

______Excellent ______Good ______Average ______Poor ______Useless

2. To what extend do you see the document as being useful in term of:

3. Do you believe that more should be done in this subject area? If so, please specify the types of assistance needed.____________________________________________________________________

4. In what ways could this document be improved?________________________________________________

5. How did this document come to your attention? ____________________________________________

6. How are you planning to use the information contained in this document?__________________________

7. Please check one item that best describes your affiliation with corrections or criminal justice. If a governmental program, please also indicate the level of government.

_____ Citizen group _____ Legislative body _____ College/University _____ Parole _____ Community corrections _____ Police _____ Court _____ Probation _____ Department of corrections or prison _____ Professional organization _____ Jail _____ Other government agency _____ Juvenile justice _____ Other (please specify)

8. Optional:

Name: ____________________________________________________________________________

Agency:____________________________________________________________________________

Address: __________________________________________________________________________

Telephone:__________________________________________________________________________

Useful Of some use Not useful

Providing new or important information

Developing or implementing new programs

Modifying existing programs

Administering ongoing programs

Providing appropriate liaisons

Classification of Women Offenders: A National Assessment of Current Practices

U.S. Department of Justice

National Institute of Corrections

Washington, DC 20534