FORUM on Corrections Research

Comparing violent and non-violent female offenders on risk and need

Violent crime has traditionally been viewed as a uniquely male phenomenon. In 1989, however, it was estimated that women were committing about 10% of serious (violent) crime in Canada,² and more recent research shows a steady increase in the rate of female offending in general.³ These findings indicate the need to understand better violent offending by women .

The advent of the Offender Intake Assessment (OIA) process in Canadian corrections allows for a comprehensive and systematic evaluation of each offender at the time of admission to federal custody.⁴ Briefly, the OIA process collects case-specific information to determine criminal risk and need areas that may be unique to each offender. Since its implementation in 1994, over 200 full OIAs have been completed for women serving sentences of two years or more. Partial intake assessment data are available for approximately 100 additional federal female offenders. This article compares violent and non-violent federally sentenced women on risk and need dimensions derived from the OIA.

Procedure

Since its implementation, all data derived from the OIA process have been entered into the Correctional Service of Canada's Offender Management System (OMS), an automated database. For this study, all available OIA data for federal female offenders were extracted from the OMS. Then, based on their most serious admitting offence, offenders were divided into two groups, violent and non-violent. The violent group consisted of offenders convicted of homicide, attempted homicide, assault, sex offences, robbery, kidnapping/forcible confinement and arson. The non-violent group consisted mostly of women serving sentences for property and drug-related offences.

Statistical analyses were used to test for differences between the two groups in the following categories: basic demographic information, criminal history (risk) indicators, overall ratings on criminogenic need areas (employment, marital/family, associates, substance abuse, community functioning, personal/emotional orientation and attitude) and various indicators of suicide potential. Subsequently, more detailed analyses compared violent and non-violent offenders on all indicators within each specific need area. Results of this study are presented in sections that follow.

Demographic data

Complete information about offender age, race and region was available on the OMS for 182 federal female offenders. Classified by their most serious admitting offence, 106 (58%) females were designated as violent, and 76 (42%) were designated as non-violent.

Violent and non-violent offenders were scattered across the regions of the country, with no between-group differences in location by offence category. The largest proportion (40%) of offenders in this study was incarcerated in Ontario, followed by the Prairies (34%), Quebec (16%) and the Atlantic (10%). Although there are federal female offenders incarcerated in the Pacific region, offender intake data were not available for this region at the time of this study.

The average age for the entire sample of female offenders under study was 33.8 years. For violent offenders, the average age was 32.4 years, while for non-violent offenders, the average age was slightly older at 34.5 years. Statistical analyses revealed no significant differences between the two groups based on age.

To simplify a comparison on race, the two groups were compared on the basis of Native status. In total, 40 (22%) of the 182 female offenders were Aboriginal, with the majority of the remaining non-Native offenders being Caucasian. Although there were no significant differences between violent and non-violent offenders based on race, Aboriginal offenders were more likely to be serving sentences for violent offences. More specifically, while 68% of Aboriginal offenders were classified as violent, only 56% of their non-Native counterparts were similarly classified.

Criminal history

The Criminal Risk Assessment is a main component of the OIA process. It is based primarily on the offender's criminal record and provides specific information about past and current offences. Moreover, any other details relevant to specific risk factors are included in the criminal profile report.

By design, there are significant and expected differences between violent and non-violent female offenders on variables related to their current offences. Very few differences are noted, however, when looking at previous experiences with youth court, adult court and other selected criminal history indicators. While there is a small tendency for violent offenders to have had more experience with the criminal justice system, statistical analyses showed that the majority of these differences were negligible. Only two specific background indicators yielded significant differences between the two groups, and both of those were associated with youth court sentencing. Percentage distributions for selected criminal history indicators are outlined in Table 1.

Table 1

Comparing Criminal History Indicators of Violent and non-violent Female Offenders
Criminal History Indicators	Violent (n=106)	Non-violent (n=71)
Previous Youth Court	31%	21%
Community supervision*	14%	4%
Open custody*	21%	7%
Secure custody	19%	9%
Previous Adult Court	66%	56%
Community supervision	52%	44%
Provincial terms	47%	39%
Federal terms	20%	13%
Previous:
Segregation (disciplinary)	22%	15%
Escape/UAL	15%	11%
Failure on conditional release	26%	18%
< 6 Mo. since last incarceration	26%	18%
Notes: *p<0.05: UAL = unlawfully at large

Case need levels

The other core component of the OIA process involves the identification of the offender's criminogenic needs. The Case Needs Identification and Analysis (CNIA) considers specific aspects of the offender's personality and life situation. Data for CNIA are clustered into seven target domains, with multiple indicators for each: employment (35 indicators), marital/family (31 indicators), associates/ social interaction (11 indicators), substance abuse (29 indicators), community functioning (21 indicators), personal/emotional orientation (46 indicators) and attitude (24 indicators). Based on the number of indicators endorsed, the completed CNIA classifies offenders on each target domain along a four-point continuum. Classifications reflect the degree of offender need, ranging from "asset to community adjustment" ­ meaning that the offender would not likely have trouble in that need area once released from prison ­ to "significant need for improvement." (Asset to community adjustment is not applicable to substance abuse and personal/emotional orientation.)

Percentage distributions for each domain need level, by group, are presented in Table 2. Violent offenders exhibited more difficulty than their non-violent counterparts in five of seven areas. The two exceptions were associates/ social interaction and community functioning. This difference was especially marked in the area of substance abuse, where about 74% of violent female offenders exhibited at least some difficulty, compared to about 46% of non-violent women.

Perhaps the most noteworthy factor revealed by the CNIAs of federally sentenced women is the fact that a large percentage of these offenders showed difficulty across all need areas. This reinforces the necessity to address these issues in correctional management and in programming for female offenders.

Table 2

Identified Needs of Violent and Non-violent Female Offenders at Admission
Type of Need	Violent (n=106)	Non-violent (n=127)
Employment*	80.2%	70.1%
Marital/Family***	87.8%	68,5%
Associates	73.5%	76.4%
Substance abuse***	73.6%	45.7%
Community functioning	75.5%	68.5%
Personal/Emotional***	97.2%	82.7%
Attitude**	35.8%	20.5%
Notes: *p<0.05, **p<0.01, ***p<0.001

Case need indicators

To explore the nature of the differences between the two groups of female offenders in target need domains further, multiple analyses were performed to compare violent and non-violent offenders for each domain. Results showed that, across all target domains, almost all differences between the two groups of offenders indicated a higher degree of need for violent female offenders.

Of the 35 indicators targetting employment needs, five significant between-group differences were found. The most critical difference (p<.001) was in employment history, where 34% of violent women had no employment history, compared to only 11% of non-violent women. Other significant indicators were unemployed at the time of arrest, unstable job history, lack of initiative and difficulty with supervisors.

In the marital/family target domain, six of the 31 indicators were found to distinguish significantly between groups. In this area, the most striking differences were in childhood family functioning, where over 30% of violent female offenders lacked family ties during childhood. This was true for only 12% of non-violent women offenders. Significant differences were also noted in childhood relations, where violent women were more likely to have experienced negative relationships with their mothers, siblings and other relatives. Between-group differences were also noted in parenting responsibilities: about half the violent women had no parenting responsibilities (before incarceration) compared to about one third of non-violent offenders. Finally, violent offenders were more likely to have had prior family/marital assessments (15%) than non-violent offenders (4%), although no differences were noted in participation in marital/family therapy or intervention.

As mentioned, there were no discernible differences between violent and non-violent offenders on overall level of need in the area of associates/social interaction. However, between-group comparisons for each of the 11 indicators within the specified need area (associated/social interaction) revealed significant differences on two indicators. First (and not surprising) violent offenders were more likely to associate with substance abusers (68%) than non-violent offenders (51%). A second and unexpected finding, however, showed that almost 70% of non-violent offenders had many criminal acquaintances, compared to 54% of violent offenders: this difference was also statistically significant.

Table 3

Selected Substance Abuse Indicators Assessed by OIA: Violent and Non-violent Female Offenders
Significant Domain Indicators	Violent	Non-violent
Abuses Alcohol***	57.1%	29.7%
Began drinking at an early age***	50.5%	20.3%
Drinks on regular basis**	37.3%	18.3%
Has a history of drinking binges***	47.1%	21.1%
Has combined the use of alcohol and drugs**	44.2%	23.3%
Drinks to excess during leisure time ***	45.1%	18.2%
Drinks to excess insocial situations**	46.7%	21.1%
Drinking interferes with employment*	46.1%	21.1%
Drinking interferes with marital/family relations**	28.9%	13.9%
Drinking interferes with social relations***	42.3%	19.4%
Drinking has resulted in law violations***	35.9%	11.1%
Abuses Drugs**	50.5%	20.6%
Began using drugs at an early age**	67.6%	44.6%
Has gone on drug taking sprees*	47.6%	25.7%
Has combined the use of different drugs**	48.1%	30.1%
Uses drugs during leisure time**	46.2%	23.6%
Uses drugs in social situations**	51.4%	29.7%
Drugs to releve stress*	51.0%	31.5%
Drug use interferes with marital/family relations*	30.1%	20.8%
Drug use interferes with marital/family relations*	48.1%	32.9%
Drug use has resulted in aw violations*	52.9%	36.5%
Drug use interferes with health*	39.8%	24.7%
Has participated in substance abuse treatment**	50.0%	28.4%
Has completed substance abuse treatment*	37.6%	21.6%
Notes: *p<0.05, **p<0.01, ***p<0.001; indicator numbers may very slightly

In the target domain of substance abuse, almost all the indicators differentiated violent women from their non-violent counterparts, with the former showing more deficits. Results of these analyses are found in Table 3. Interestingly, differences are more marked for indicators pertaining to alcohol abuse than those pertaining to drug abuse: violent women were almost twice as likely to be alcohol abusers than non-violent women.

Few indicators in the area of community functioning were found to differentiate between groups. The most notable finding in this domain (p<.001) was that over half the violent women had unstable living arrangements (before incarceration) compared to only 26% of non-violent women. There were also small but statistically significant differences in monetary issues, where non-violent female offenders were less likely to have a bank account and more likely to have outstanding debts.

Twenty-eight of the personal/emotional orientation indicators were significantly associated with violence. Not surprisingly, all but one of these indicators suggest that violent women have more personal and emotional problems than their non-violent counterparts. Table 4 reveals that cognition is the principal component underlying many of the observed differences. More specifically, violent female offenders were more likely to be unable to recognize problem areas and make informed choices, have unrealistic goals, be narrow and rigid thinkers, show disregard for others and be unable to understand the feelings of others. Violent offenders also showed more difficulty in mental health and intervention issues. For instance, they were more than twice as likely to have been diagnosed as disordered, and the majority had received medication and/or a personal/emotional assessment at some point.

Table 4

Selected Personal/Emotional Indicators Assessed by OIA: Violent and Non-violent Female Offenders
Significant Domain Indicators: Personal/emotional Orientation	Violent	Non-violent
Physical prowess problematic*	5.7%	0.0%
Family ties are problematic*	57.7%	38.4%
Has difficulties solving interpersonal problems***	64.8%	34.3%
Unable to generate choices*	67.6%	50.7%
Goal setting is unrealistic***	21.6%	2.8%
Has disregard for others*	26.0%	11.0%
Incapable of understanding the feelings of others**	18.3%	4.2%
Narrow and rigid thinking**	25.2%	6.9%
Aggressive***	40.4%	13.9%
Copes with stress poorly*	71.4%	52.1%
Poor conflicy resolution***	78.9%	38.4%
Manages time poorly*	20.0%	6.9%
Has low frustration tolerance***	46.6%	20.6%
Hostile**	34.3%	13.9%
Takes risks inappropriately**	51.0%	72.2%
Thrill seeking*	28.2%	14.1%
Non-reflective**	39.4%	20.6%
Manipulative*	34.0%	17.1%
Sexual attitudes are problematic*	11.0%	2.7%
Diagnosed as disordered in the past**	31.0%	12.7%
Diagnosed as disordered currently*	20.8%	8.2%
Prior personal/emotional assessments***	55.0%	18.6%
Prescribed medication in the past***	64.1%	37.5%
Prescribed medication currently*	41.2%	26.4%
Past hospitalization***	41.6%	17.4%
Received outpatient services in the past**	40.8%	19.7%
Received outpatient services prior to admissior**	22.2%	5.6%
Past program participation*	36.9%	20.8%
Current program particiaption*	30.4%	16.4%
Notes: *p<0.05. **p<0.01. ***p<0.001; indicator numbers may vary slightly

Comparisons across indicators of attitude yielded significant results for half of the 24 items. Again, each significant difference presented the violent group as higher need than the non-violent group. Violent women were much more likely to have negative attitudes toward the law, the police, the courts, corrections, community supervision and rehabilitation. They were also significantly more likely to value substance abuse, and to view employment and basic life skills as useless. Given the nature of their offences, it was not surprising to find that over 20% were supportive of instrumental violence, compared to none of their non-violent counterparts. Suicide potential One component of the initial OIA evaluates each offender for suicide potential at admission. Nine specific indicators are used to determine the potential for suicide. Violent women were compared to non-violent women across all indicators of suicide potential, and the results are listed in Table 5.

Violent women were assessed as being at higher risk for suicide by most indicators. What is most striking about the data contained in Table 5 is the fact that 53% of female offenders convicted of a violent offence have attempted suicide in the past. This figure is more than double that of female offenders serving sentences for non-violent crimes. These results support an earlier investigation⁵ involving a sample of federally sentenced women which showed that a history of attempted suicide was a potent predictor of violent recidivism.

Table 5

Suicide Risk Potential Assessed by OIA: Violent and Non-violent Female Offenders
Suicide Risk Indicators	Violent	Non-violent
May be suicidal	11.8%	11.4%
Previous suicide attempts(s)***	53.0%	24.3%
Expressed suicide intent	10.8%	4.2%
Has plan for suicide	3.9%	0.0%
Has had recent psychological/psychiatric intervention*	37.4%	21.1%
Recent lossof relationship/death of close relative	18.6%	12.7%
Excessively worried about problems re: legal situation **	21.0%	5.8%
Under influence of alcohol or drugs/signs of withdrawal	5.0%	8.8%
Showing signs depression	17.0%	17.1%
Notes: *p<0.05, **p<0.01, ***p<0.001; indicator numbers may very slightly

Conclusions

This study has revealed results that are both expected and surprising. Although there were some differences between violent and non-violent female offenders on variables related to criminal history, sentencing experience and institutional adjustment, very few were statistically significant. Moreover, aside from current offence data, the observed differences were relegated to youth court experiences.

Results clearly indicate that violent female offenders present higher need levels than their non-violent counterparts. This difference was noted across five of the seven target domains. The exceptions were the associates/social interaction and community functioning domains. Micro-level analyses (indicators), however, revealed a trend for non-violent offenders to have more criminal associates. This was somewhat surprising, considering that other research with federally sentenced women⁶ suggests that criminal associates are predictive of violent recidivism. It is hoped that future research will explore this contradiction.

In a close examination of each individual indicator for each target domain, it was found that those associated with substance abuse exhibited the most (and strongest) between-group differences. More specifically, violent female offenders showed a higher tendency to problematic substance use. This was evidenced in the fact that 24 of the 29 substance abuse indicators yielded statistically significant between-group differences. This finding is not surprising, as most researchers posit an unequivocal association between substance abuse and violent behaviour. Moreover, the need for ongoing, intensive substance abuse programming is highlighted ­ especially for those women convicted of violent offences.

Finally, the results demonstrate significant associations between suicide potential and violent offending, with the strongest indicator being previous attempts at suicide. It was suggested elsewhere⁷ that more consideration be given to suicide risk and self-injury in the assessment of women offenders. It merits reiteration here, as results from the present investigation imply again that suicide potential could be paramount in risk prediction, particularly for female federal offenders.

---

1. 340 Laurier Avenue West, Ottawa, Ontario, K1A 0P9.

2. A. Hatch and K. Faith, "The Female Offender in Canada:
A Statistical Profile," Canadian Journal of Women and the Law, 3, 2 (1989): 432-456.

3. R. Boe, CSC Offender Population Forecast for 1993 to 2002 (Ottawa: Correctional Service of Canada, 1992).

4. For a more detailed description of OIA, see L.L. Motiuk, "Classification for Correctional Programming: The Offender Intake Assessment Process," Forum on Corrections Research, 9, 1 (1997): 18-22.

5. K. Blanchette and L.L. Motiuk, Female Offender Risk Assessment: The Case Management Strategies Approach, Paper presented at the Annual Convention of the Canadian Psychological Association (Charlottetown, 1995).

6. Blanchette and Motiuk, Female Offender Risk Assessment: The Case Managment Strategies Approach.

7. K. Blanchette, "Classifying Female Offenders for Correctional Interventions," Forum on Corrections Research, 9, 1 (1997): 36-40.