Skip to content | Skip to institutional links

Common menu bar links

Institutional links

Research Reports

Warning This Web page has been archived on the Web.

An Investigation into the Characteristics of Substance-Abusing
Women Offenders:
Risk, Need and Post-Release Outcome

This report is also available in French. Ce rapport est également disponible en français. Veuillez vous adresser à la direction de la recherche, Service Correctionnel du Canada, 340 avenue Laurier ouest, Ottawa (Ontario) K1A 0P9. Should additional copies be required they can be obtained from the Research Branch, Correctional Service of Canada, 340 Laurier Ave., West, Ottawa, Ontario, K1A 0P9.

1999 N^o R-81

Craig Dowden
and
Kelley Blanchette

PREPARED FOR

Research Branch
Correctional Service of Canada

April 1999

EXECUTIVE SUMMARY

Programming research for women offenders is a fairly modernphenomenon. Moreover, until recently, there was only one federal correctional facility for female federal offenders. The opening of five new federal facilities, however, emphasized the need to provide an appropriate environment to manage and rehabilitate women offenders. Community reintegration is a valued goal of Correctional Services Canada (CSC) and development of treatment programs for women is viewed as an important component to facilitate offenders' release. Accordingly, the present report focuses on issues relating to substance abuse in the federal female offender population, as this is a demonstrated treatment need for many women offenders.

The present investigation compared women offenders who were substance abusers to those who were not. The groups were compared on a number of different criteria: risk and need variables, demographic characteristics, and recidivism data. To obtain the required information for the present study, CSC's automated data base, the Offender Management System (OMS) was accessed in conjunction with the Canadian Police Information Centre (CPIC) records. The final sample comprised 251 women offenders for whom institutional program participation information was available. Almost 60% of those had successfully completed a substance abuse treatment program at some point during their incarceration.

The first set of comparisons focused on demographic information (age, race). Statistical analyses revealed that the substance abusers were younger than their non-abusing counterparts. The mean age for the former group was 32.7 years compared to 34.9 years for the non-abusers. Results also indicated that Aboriginal women were over-represented amongst substance abusers. An extremely high proportion (93%) of Aboriginal women was classified as substance abusers, compared to 49% of non-Aboriginal women. This difference was statistically significant, and highlights the need for substance abuse programming tailored to Aboriginal women.

Comparisons on overall risk ratings also demonstrated significant between-group differences. The majority of substance abusers were classified as medium (46%) and high-risk (26%) offenders whereas only 28% were low-risk. Non-abusers, on the other hand, were predominantly low-risk (60%), with only 17% classified as medium-risk, and 23% high-risk.

Substance abusers differed from non-abusers on several criminal history variables as well. Substance abusers tend to start their criminal careers at an earlier age compared to non-abusers, with 39% of them having youth court experience compared to only 8% of non-abusers. Data revealed that substance abusers were also more likely to have adult court experience. Not surprisingly, they were also more likely to have previous escape attempts or 'unlawfully at large' convictions on their criminal records. Finally, they were five times as likely to have been placed in (disciplinary) segregation as compared to non-abusers.

Comparisons were also performed on the global need ratings assigned (at intake) to both groups of women offenders. Results indicated that the majority of non-abusers were classified as low-need (56%), whereas 25% were medium-need and only 19% were high-need. Many substance abusers, on the other hand, were classified as high-need (47%). An almost equal number (41%) were categorized as medium-need and only 13% were low-need. The substance abusers exhibited more difficulties in five of seven need domains examined, including associates, attitudes, employment, marital/family, and (as expected) substance abuse.

When groups were compared on individual need indicators, analyses produced significant between-group differences. Substance abusers consistently had more difficulties than their non-abuser counterparts. Given the selection criteria for grouping the study participants, many of the obtained results were expected. For example, the findings confirmed that substance abusers had more substance-abusing friends (85%) and possessed attitudes more favourable towards substance abuse (24%) than non-abusers. However, substance abusers also had more difficulty in a number of domains unrelated to substance abuse. This suggests that substance abuse problems are not uni-dimensional and interact with a number of criminogenic need areas. Notably, previous studies (Fabiano, 1993; Lightfoot & Lambert, 1992; Shaw et al., 1991) have suggested that substance abuse serves to exacerbate problems in other areas of one's life.

A final set of comparisons examined the recidivism rates of the study sample. There was a trend for substance abusers to recidivate at a higher rate than non-abusers; however, this was not statistically significant. Next, the released substance abusers that completed relevant institutional programming were compared to their untreated counterparts on post-release outcome. Although the recidivism rates for both groups were relatively low, those who had participated in substance abuse programming were significantly less likely to return to custody than their untreated counterparts. Importantly, a comparison on global risk ratings revealed that differences in recidivism were not due to a between-group difference in overall risk. This bodes well for the Service's institutional substance abuse programs, although prospective research is needed to support a definitive causal link between substance abuse programming and lowered recidivism amongst women offenders.

In conclusion, results of the present study reveal clear and reliable differences between substance abusers and non-abusers in a variety of areas assessed at intake. Moreover, of women with substance abuse problems, those who complete relevant programming fare better after release than their untreated counterparts. It is therefore posited that appropriate assessment, classification, and intervention with women substance abusers can significantly increase their potential for successful community reintegration. This study provides preliminary support for the efficacy of substance abuse treatment for incarcerated women.

EXECUTIVE SUMMARY
TABLE OF CONTENTS
LIST OF TABLES
INTRODUCTION
METHOD
RESULTS
CONCLUSIONS
REFERENCES
APPENDIX A
APPENDIX B
APPENDIX C

LIST OF TABLES

Table 1: Percentage Distribution of Security Level by Substance Abuse
Table 2: Percentage Distribution of Overall Risk Ratings by Substance Abuse
Table 3: Percentage Distribution of Selected Criminal History Background Indicators: Substance Abusers and Non-Abusers
Table 4: Identified Needs of Substance Abusing and Non-Abusing Federal Women Offenders at Admission
Table 5: A Breakdown of Discriminating 'Associates' Need Indicators as Assessed by the Offender Intake Assessment Process
Table 6: A Breakdown of Discriminating 'Employment' Need Indicators as Assessed by the Offender Intake Assessment Process
Table 7: A Breakdown of Discriminating 'Marital/Family' Need Indicators as Assessed by the Offender Intake Assessment Process
Table 8: A Breakdown of Discriminating 'Attitude' Need Indicators as Assessed by the Offender Intake Assessment Process
Table 9: A Breakdown of Discriminating 'Community Functioning' Need Indicators as Assessed by the Offender Intake Assessment Process
Table 10: A Breakdown of Discriminating 'Personal/Emotional' Need Indicators as Assessed by the Offender Intake Assessment Process
Table 11: A Breakdown of 'Substance Abuse' Need Indicators as Assessed by the Offender Intake Assessment Process
Table 12: Percentage Distribution of Risk/Need Levels at Time of Admission

INTRODUCTION

An Investigation into the Characteristics of Substance-Abusing Women Offenders: Risk, Need, and Post-Release Outcome

There is an extremely high prevalence of substance abuse among offenders. More specifically, it has been reported that approximately 70% of federal inmates have difficulties with drug or alcohol abuse (Fabiano, 1993). In addition, it has been noted that many, if not most, of these offenders have problems in other areas of community and personal functioning. For example, research suggests that substance abusers have a number of problems in cognitive, emotional, environmental/social, and behavioral functioning. Moreover, these other needs may mediate the negative effects of substance abuse problems experienced by particular offenders. Therefore, effective treatment programs for substance abusers must address a variety of need areas as well as intensively focusing on the substance abuse problem itself (Fabiano, 1993).

Although the majority of offenders are substance abusers, several studies have reported that female offenders in particular have a high rate of substance abuse problems (Blanchette, 1996; Tien, Lamb, Bond, Gillstrom, & Paris, 1993). A comparison between the characteristics of substance abusing and non-abusing female offenders should provide important insight into an appropriate program structure for substance abusing women. The present investigation examines this issue in detail.

There have been few studies focusing on the characteristics of substance abusing women offenders. Lightfoot and Lambert (1991; 1992) assessed the prevalence of drug problems of offenders incarcerated at the Prison for Women in Kingston, Ontario using the Drug Abuse Screening Test (DAST). The DAST was administered to 80 incarcerated women, and it was reported that 14% of the women surveyed had severe drug-related problems. On the other hand, only 35% of the women offender population at the Prison for Women reported that they did not have any problems related to drugs. These data confirm the high prevalence of substance abuse within the federal female offender population in Canada.

Lightfoot and Lambert (1991; 1992) also requested the women offenders to detail their drinking patterns. Of those surveyed, 25% described themselves as alcoholics. Notably, 45% of the women surveyed reported that they had experienced substantial problems related to their substance use. These women highlighted difficulties in the areas of family (81%), legal (69%), and work and health issues (53%). Again, these results support the view that substance abuse is a significant problem for the majority of federal women offenders, and adversely affects their community functioning in various of ways.

The problems caused by substance abuse do not singularly affect community functioning. These needs are also intricately related to their criminal offending. Shaw et al. (1991) conducted a survey of provincial prisoners and federal women offenders residing at the Prison for Women. This survey examined the relationship between the offenders' substance abuse problems and their prior criminal records. The authors reported that 120 of the 170 women (71%) surveyed indicated that drugs or alcohol played a role in their offences. Furthermore, 52% of the sample reported that they were under the influence of a substance when they committed their current offence.

Further evidence for the relationship between substance abuse and criminality in women offenders was derived from research by Lightfoot and Lambert (1991; 1992). They indicated that most of the women who used a substance on the day of their offence report that it had negatively affected their judgment. Sixty percent believed that their use of a substance seriously impaired their decision-making abilities. Only 17% reported that the use of a substance did not negatively impact on their behaviour. Therefore, the vast majority (83%) of those offenders who used a substance on the day of their offence agreed that it impaired their judgment to some degree.

The above findings certainly suggest that substance abuse has a significant negative impact on women offenders' ability to make rational, prosocial choices and likely contributes to certain at-risk individuals' criminal behaviour.

The purpose of the present investigation is to compare the characteristics of female substance abusing offenders to non-abusing offenders on intake assessment data, including demographic characteristics, criminal history risk, and various criminogenic needs. These comparisons will provide important information regarding the differences between substance abusers and non-abusers. More importantly, analyses may identify specific cognitive, emotional, behavioural, and environmental/social needs of women offenders with substance abuse problems. This information is paramount for relevant and empirically derived programming for women substance abusers.

Finally, the women who completed institutional substance abuse programs will be compared to untreated substance abusers on post-release outcome (i.e., recidivism). This analysis will provide an examination of the association between completion of institutional substance abuse programming and recidivism.

METHOD

The present investigation compared 'substance-abuser' women offenders to women offenders who were not substance abusers on demographic criteria, risk/need data, and post-release outcome. A wide definition of 'substance abuser' was used, and included all women who were assessed and recommended for substance abuse treatment at intake. The assessment data used in this study was collected from the Offender Management System (OMS; an automated database). Outcome data were derived from Canadian Police Information Center (CPIC) records.

OMS is an automated database that stores a wealth of information pertaining to federal offenders currently or formerly incarcerated in Canada. Most of the required information for the present study was obtained through this source. For example, OMS stores all Offender Intake Assessment (OIA) data for every offender. The correctional plans included within OMS were used to test whether the programs that were initially recommended for the offenders at intake were actually completed during incarceration. Finally, offender progress summaries and other information relating to successful and unsuccessful program completion were also obtained through OMS.

The initial sample included all (354) federal female offenders who had received an Offender Intake Assessment (OIA) at admission to a federal correctional facility. A number of subjects had to be dropped from the study due to either a lack of available finger print serial numbers, the high profile nature of the case, or missing OIA data. In addition, 84 offenders were dropped from the initial analyses because they did not participate in any form of correctional treatment program or all program information was missing. The final sample included 251 federal women offenders who had participated in at least one treatment program during their incarceration. Since the OIA data played an extremely important role in the development of this project, a brief description of the information obtained through the OIA process is provided below.

The Correctional Service of Canada has been using the OIA since 1994. It is the standard intake assessment used by all federal institutions in Canada. The OIA procedure examines a broad range of factors of the offenders and it serves as the fundamental basis for determining their individualized correctional plans. The OIA has two main components: Criminal Risk Assessment (CRA) and Case Needs Identification and Analysis (CNIA). Both of these sections provide important information for offender assessment and classification. These two components will be discussed in more detail in later sections.

CPIC records were used to determine the official offence history of each of offender included in the present sample. A comprehensive coding manual (Appendix A) was developed to measure all offences that had resulted in the current federal conviction for each particular subject. The information provided by the CPIC records was divided into three different conviction categories. Past history included all offences previously committed by the offender. The current offence category recorded all offences included in the present conviction (i.e., those resulting in the first incarceration post-1994). A broad definition of 'official recidivism' was used. Specifically, it included return to prison for any reason after release from the 'current' incarceration.

The final stage in the research process involved examining the correctional plans for all offenders in detail. A coding guide (Appendix B) was developed to document all information concerning the primary need areas that were recommended to be addressed during treatment for each offender. In addition, the principle components for each need were recorded. The program modules and progress summaries provided information on actual enrollment and this information was also recorded in the coding guide. This type of overlapping information provided an opportunity to examine how closely the correctional plans were followed in the assignment of substance abuse treatment programs for women offenders.

The final sample of 251 women offenders was evaluated on a number of criteria. One set of analyses focused on the details of their past and present criminal convictions. The degree of agreement between the proposed treatment programs found in the correctional plan and the programs that the offenders actually completed while incarcerated was also tested. This analysis permitted a determination of whether the offenders recommended for substance abuse treatment actually received such intervention while incarcerated. Finally, recidivism data for released offenders was gathered to determine whether participation in substance abuse programming was associated with more successful community reintegration.

It should be noted that stringent criteria for statistical significance were used due to the nature and number of pairwise comparisons performed in the current study. This reflects an effort to reduce potential Type 1 error associated with conducting multiple, non-orthogonal, between-group comparisons. Specifically, unless otherwise stated, criterion for statistical significance was set at .001.

RESULTS

A series of statistical analyses focused on comparing substance abusing female offenders to their counterparts who were not substance abusers. As mentioned, the 'substance abuser' group included women assessed (at intake) as requiring substance abuse treatment (n = 143). The 'non-abuser' group included all remaining women who had received an OIA, and were not recommended for substance abuse treatment (n = 108). Groups were compared on demographic variables, security level, global risk/need scores, need domains and indicators, various aspects of criminal history, and post-release outcome.

Concordance between Program Recommendation and Enrollment

In a recent study of women offenders in cognitive skills training, investigators reported markedly low concordance rates between program recommendations (as per correctional plan) and actual program completion (Dowden & Blanchette, 1997). Specifically, according to OMS data, less than one-third of women offenders recommended for cognitive skills training actually received it during incarceration.

Fortunately, concordance between recommendations and enrollment for women offender substance abuse programming surpassed that found in the cognitive skills study. Of the 143 women recommended for substance abuse intervention, 86 (60%) actually completed it. Also, of those women who completed institutional substance abuse programming (n = 129), the majority (67%) were recommended as per their correctional plans at intake.

High concordance between the program recommendations provided by the parole officer and enrollment within the targeted programs is desirable. The parole officers who make these recommendations review comprehensive information regarding an offender and then provide the most informed judgment possible concerning his or her criminogenic needs and required programs.

There may be a number of reasons for the lack of concordance between program recommendations and actual program completion. For one, it is possible that the facility in which the offender resides does not provide the specific recommended intervention. Second, the offender may resist placement within the recommended treatment program. Some offenders are placed in segregation, transferred, or paroled mid-treatment. Of course, this would then prevent them completing the program. Finally, it is possible the staff member responsible for implementation of the correctional plan feels that recent developments in the life of the offender necessitate a deviation from the correctional treatment plan. Thus, there are numerous potential reasons for why an offender does not complete the recommended treatment program. Such 'slippage', however, remains an issue to be monitored in the evaluation of correctional programs.

Security Level

Security level at admission was available for 102 of the 'non-abuser' women and 142 of the substance abusers. Examination of the data revealed that 17% of the women substance abusers were classified as maximum-security as compared to only 4% of non-abusers. Accordingly, 31% of the substance-abusing women were classified as minimum-security whereas the majority (60%) of non-abusers were so classified. Statistical analyses revealed that these results were highly significant, as presented in Table 1.

Table 1: Percentage Distribution of Security Level by Substance Abuse
Security Level	Substance Abusers		Non-Abusers
Minimum	44	31%	61	60%
Medium	72	51%	37	36%
Maximum	24	17%	4	4%

Total	140		102

Note: Chi-square test of significance; p < .001

Demographic Information

On average, the substance abusers were younger than the comparison group. Substance abusers ranged in age from 19 to 58 years, with a mean of 32.7. The age range for the non-abusers was 20 to 64 years, with a mean of 34.9. Although there was a noted age differential between groups, analyses failed to render statistical significance. However, it should be noted that the results were extremely close to obtaining the appropriate significance level (t(249) = 2.0; p= .052).

Offenders were also compared on their Aboriginal status. Results indicated that a significantly higher proportion of substance abusers were Aboriginal offenders. Almost all (93%) of the Aboriginal women in this study were recommended for substance abuse programming. Only 49% of non-Aboriginal women, on the other hand, were categorized as substance abusers. These results were highly statistically significant (p<.001).

Criminal Risk Assessment

The Criminal Risk Assessment (CRA) is one of the two central components of the OIA process. The purpose of the CRA is to provide an overall risk rating for each offender. The CRA examines details of the past and current conviction record of the offender, and other relevant information regarding potential risk factors from the criminal profile report. These sources of information are combined to yield an overall risk rating of 'low', 'medium', or 'high'.

Global risk ratings were available for almost all (n = 248) of the women in the sample. Statistical analyses revealed significant differences in overall risk levels between substance abusers and non-abusers (c ² = 29.8; p<.001). More specifically, 27% of the substance abusers were high risk, 45% were medium risk and 28% were classified as low-risk. The non-abusers, on the other hand, were mostly classified as low-risk (60%). Table 2 presents the risk ratings received for each category of offender.

Table 2: Percentage Distribution of Overall Risk Ratings by Substance Abuse
Overall Risk Rating	Substance Abusers		Non-Abusers
Low Risk	40	28.0%	63	60.0%
Medium Risk	65	45.4%	18	17.1%
High Risk	38	26.6%	24	22.9%

Total	143		105

Note: Chi-square test of significance; p < .001

Specific criminal history risk variables were also examined. Interestingly, statistical analyses revealed significant between-group differences on almost all selected criminal history background indicators. In each case, substance abusers had significantly more problems in these areas.

Table 3 presents percentage distributions of selected criminal history background indicators for substance abusers and non-abusers. Clearly, the substance abusers started their criminal careers at a younger age than non-abusers. Over one-third (39%) of the substance abusers had previous youth court experience, compared to only 8% of the non-abusers. The same pattern of results was found in adult court experience, with substance abusers having significantly more experience in adult court.

Table 3: Percentage Distribution of Selected Criminal History Background Indicators: Substance Abusers and Non-Abusers
	Substance Abusers (N = 107)	Non-Abusers (N = 84)

Previous Youth Court*	38.7%	8.3%
Community Supervision*	25.7%	6.0%
Open Custody*	22.2%	3.6%
Secure Custody*	18.7%	6.0%

Previous Adult Court*	77.7%	40.5%
Community Supervision*	65.2%	28.6%
Provincial Terms*	65.2%	20.2%
Federal Terms*	18.8%	6.0%

Previous	21.9%	4.9%
Segregation (disciplinary)*	18.9%	3.6%
Escape/UAL*	18.9%	3.6%
Failure on Conditional Release* < 6 Months Since Last Incarceration	17.9%	3.6%

Note: Chi-square tests of significance; *p<.001

Examination of Table 3 reveals that the substance abusers had more disciplinary segregation and more extensive escape histories as compared to non-abusers. Breaches of trust were also more common among substance abusers, with almost one-third having failed on conditional release at some point. Finally, substance abusers were more likely to have been incarcerated within the six months prior to their current offence as compared to non-abusers.

Between-group differences were also noted in the types of offences committed. More specifically, 41% of substance abusers were incarcerated for a robbery offence, compared to only 16% of the non-abusers. In fact, the groups differed in terms of violent offence histories. While 58% of substance abusers had a violent offence in their past or current criminal history, this was true for only 32% of the non-abusers. These results suggest differential patterns of offending between substance abusers and their non-abusing counterparts.

Case Needs Identification and Analysis

The Case Needs Identification and Analysis (CNIA) is the second major component of the OIA process. The purpose of the CNIA is to identify the criminogenic needs of the offender and to make recommendations on how to best address these needs. In order to make these determinations, the CNIA considers a large amount of information related to the personality of the offender as well as his/her current life situation. Multiple, specifically defined need indicators are examined, covering seven major need domains: associates/social interaction (11 indicators), attitudes (24 indicators), community functioning (21 indicators), employment (35 indicators), marital/family (31 indicators), personal/emotional orientation (46 indicators), and substance abuse (29 indicators). A complete listing of all need indicators is located in Appendix C.

Once every indicator has been scored, the offender receives an overall rating on each target domain based on a four-point continuum. The classifications received reflect the offender's degree of need for each domain of interest. The scale ranges from "asset to community adjustment" to "significant need for improvement". The two intermediate ratings are "no need for improvement" and "some need for improvement". It should be noted that "asset to community adjustment" is not applicable to the substance abuse and personal/emotional domains.

For the present investigation, the rating scale was collapsed and the scores were dichotomized to indicate the presence or absence of a particular need. Specifically, 'asset to community adjustment' and 'no need for improvement' were collapsed to indicate absence of need. Domains ranking as 'some' or 'considerable' need for improvement was collapsed to indicate presence of need. These data were available for 112 of the substance abusers and 81 of the women offenders in the 'non-abuser' comparison group. Percentage distributions, by substance abuse classification, are reported in Table 4.

Table 4: Identified Needs of Substance Abusing and Non-Abusing Federal Women Offenders at Admission
Type of Need	Substance Abusers (N = 112)	Non-Abusers (N = 83)
Associates*	93.7%	67.7%
Attitudes*	34.3%	13.0%
Community Functioning	80.4%	70.4%
Employment*	88.1%	67.6%
Marital/Family*	83.2%	63.9%
Personal/Emotional	94.4%	86.1%
Substance Abuse*	87.4%	20.4%

Note: Chi-square test of significance; *p<.001

Table 4 clearly indicates that both substance abusers and non-abusers have difficulties in a large number of need areas. Notably, statistical analyses revealed highly significant between-groups differences in five of the seven need domains. As expected, those recommended for substance abuse treatment had significantly greater difficulties in the substance abuse domain with almost 90% exhibiting difficulties compared to only 20% of the 'non-abuser' group. Importantly, however, they also had significantly more problems with associates, attitudes, employment, and their marital/family situation. Although data showed a trend for substance abusers to have more problems in the 'community functioning' and 'personal/emotional' domains as well, differences failed to reach statistical significance at the adjusted (.001) level. Taken together, these results suggest that substance abuse in women offenders is generally symptomatic of multiple criminogenic needs.

To explore differences between groups in more detail, between-group comparisons on the individual need indicators were also conducted (see Appendix C for a complete listing). This provided a more specific examination of the individual need domains, allowing a precise determination of where the differences emerge between groups of women. However, it should be noted that not all of the individual need indicators were available for the entire sample. Rather, OIA indicator data were available for 195 of the 251 study participants.

Many of the 197 pairwise comparisons resulted in significant between-groups differences. To adjust for the probable inflation in Type I error rate, conservative alpha level of .001 was used. Notably, in every case, the substance abusers were found to have significantly more difficulties than the comparison group of non-abusers. As expected, the substance abuse domain had the largest number of indicators that discriminated between groups and these were also the most statistically robust differences found.

Table 5: A Breakdown of Discriminating 'Associates' Need Indicators as Assessed by the Offender Intake Assessment Process
Associates/Social Interaction Indicators	Substance Abusers (N = 112)	Non-Abusers (N = 83)
Has less than Grade 10	84.7%	28.0%
Criminogenic Living	39.2%	12.0%

Note: Chi-square test of significance; p <.001 for all comparisons shown.

Not surprisingly, compared to non-abusers, the substance abusers were significantly more likely to have substance-abusing and criminal friends. Accordingly, over one-third of substance abusers were assessed as having a criminogenic lifestyle.

Five employment indicators yielded reliable between-group differences. Once again, the women offenders recommended for substance abuse treatment showed more needs than their 'non-abuser' counterparts. A breakdown of reliably discriminating employment indicators is presented in Table 6.

Table 6: A Breakdown of Discriminating 'Employment' Need Indicators as Assessed by the Offender Intake Assessment Process
Employment Indicators	Substance Abusers (N = 112)	Non-Abusers (N = 83)
Has less than Grade 10	54.8%	28.9%
Unemployed 90% or more	58.4%	26.5%
Unemployed 50% or more	79.9%	54.2%
Unstable Job History	78.8%	33.7%