Correctional Service Canada
Symbol of the Government of Canada

Common menu bar links

Research Reports

Warning This Web page has been archived on the Web.

An Investigation into the Factors leading to Increased Security Classification of Women Offenders


Joy Irving, Cherami Wichmann

Research Branch
Correctional Service of Canada

February, 2001


ACKNOWLEDGEMENTS

We would like to thank Kelley Blanchette for her input into this project, and both Kelley and Hilda Vanneste for their editorial expertise. We also acknowledge the hard work of Kendra Delveaux, Fariya Syed and Craig Dowden, who were responsible for coding the data for this project. Finally, we thank Mark Nafekh for his role in the data management process.

EXECUTIVE SUMMARY

Security reviews have a significant impact, with the potential of increasing or decreasing the women offender's security level. Change in the offender's security level has the potential to alter her physical environment, the privileges awarded to her, the treatment services provided, and the likelihood of favourable consideration for release by the National Parole Board. These facts emphasize the importance of understanding why women are, upon review, reclassified to higher levels of security. Moreover, there has been some, unsubstantiated but critical, suggestion that the factors that lead some women to be reclassified to higher levels of security are subjectively based and have little direct relationship with the women's behaviours. These concerns set the impetus for the present investigation.

The sample included 167 women offenders, for whom 275 security level reviews were available. A regression analysis was performed to identify variables important to the prediction of an increase in their security level as a result of a security review. The analysis revealed that five factors were significantly predictive of decisions to reclassify women to higher levels of security. The five significant predictors included, in order of significance,

1) displaying an uncooperative attitude (refusal to participate in institutional activities such as programs or work, disruptive to staff or other inmates),
2) being convicted of serious institutional charges (such as assault),
3) expressing little or no motivation to comply with the correctional plan or programs,
4) possessing and/or distributing contraband, and
5) having a history of being unlawfully at large.

Collectively, these factors accounted for 36% of the variance in reclassifications to a higher security in the present sample. Further, there were several other factors, not included in the regression analysis due to lower frequency occurrences, which had an impact on decisions to reclassify women to higher levels of security. These factors include assaulting and causing serious harm to others, instigating disruptions, attempting to escape, incurring positive urinalysis tests, and being segregated due to posing a danger to others are factors which appear to result in increased security reclassification of women inmates.

To illustrate, it was found that all of the security reviews of the women sampled, which resulted in a reclassification from minimum to maximum security, were prompted by serious institutional events, such as assaults or incurring charges for new offences. Over two-thirds (69%) of security reviews on women sampled which resulted in a reclassification from medium to maximum security were prompted by disruptive behaviour or institutional incidents, while only 24% of reviews resulting in a reclassification from minimum to medium were eventdriven.

In conclusion, results of the present study indicate that both subjective and objective factors were predictive of decisions to reclassify women to higher levels of security. Further, all factors identified were easily categorized as "adjustment" or "security-risk" related. Statistical analyses revealed that psychological concerns and race were not found to influence decisions to reclassify women to higher levels of security. The data from this investigation suggest that security review practices continue to manage risk for women offenders while assigning the least restrictive security level as equitably as possible.

TABLE OF CONTENTS

LIST OF TABLES

INTRODUCTION

The importance of the reclassification process cannot be underestimated, as security assessments hold major implications for inmates. For example, reclassification decisions directly effect the housing conditions for offenders, their access to programming and treatment, their level of privilege and entitlement, even the likelihood of their release. The most appropriate security level is one at which women offenders' level of rise is managed within the least restrictive and most humane environment possible (Blanchette, 1997a).

Women represent approximately two to three percent of all federally sentenced offenders in Canada. With the creation of five new regional facilities to replace the Prison for Women, as recommended by the 1990 Task Force on Federally Sentenced Women, the issue of appropriate security reclassification is of pressing importance (Correctional Service of Canada, 1990). Instead of placing all women offenders into one multi-level facility, women must be accurately assessed as to the most appropriate security level within which their needs and risk level can be managed accordingly1 For this reason, it is a necessity that reclassification assessments for women be based primarily on objective and standardized criteria in order to help ensure the highest degree of accuracy and equity.

The reclassification process implemented by Correctional Service Canada (CSC) has often been criticized by those who suggest that women offenders in general are not at high risk to re-offend or cause problems within the institution and should be reclassified at lower levels of security. (Canadian Association of Elizabeth Fry Societies, 1998; Hannah-Moffat, 1999). However, the existing data support the claim that women offenders are reclassified to higher levels of security when they present higher risk (e.g., to public safety) and require more intense correctional intervention (Blanchette, 1997a; Blanchette, 1997b; Dowden & Blanchette, 1999). Further, support for the accuracy of placement into differing levels of security has been provided by the women themselves.


1 Currently, maximum-security women are not housed in the five regional facilities.

In fact, McDonagh (1999) reported that the women she interviewed while they resided in maximum security facilities viewed reducing their security levels as primarily their own responsibility. In other words, they acknowledged that while CSC offers appropriate programming and provides the necessary assistance, they are ultimately responsible for changing their attitudes and behaviour, and following their correctional plans as closely as possible according to their individual potential. Unfortunately, however, they also reported poor understanding of how to reduce their security level.

There has been no investigation into which factors lead to reclassification to higher levels of security, nor which factors lead to decisions of reductions in security level. As such, the purpose of the present study was to clearly delineate the factors that predict security level increases following a security review. Findings from this study may help provide women offenders an understanding and knowledge-base they now lack regarding how to reduce their security level.

METHODOLOGY

Sample

Data were coded and analyzed for all (n = 167) women for whom offender security level (OSL) decisions and complete information on 'candidate' predictor variables (Appendix A) were available on the Offender Management System (OMS). Data extraction spanned from January 19, 1993 to September 24, 1998. For inclusion, the OSL review must have covered at least a six month period, unless a change in security classification occurred.

The women ranged from 18 to 52 years of age (M = 31.11, SD = 7.10) when admitted for their current sentence to a federal institution. Thirteen percent of the women were serving life sentences, the remaining 87% of the sample had an average sentence length of 67 months with a range of 20 to 304 months (SD = 47.42). In terms of ethnicity, 58% of the women were Caucasian, 25% were of Aboriginal descent2, and 17% comprised other ethnic backgrounds.

Data Sources

For the purpose of this study, the primary source of information used was the Correctional Service of Canada's automated Offender Management System (OMS).

Information from OMS was compiled from progress summary reports, institutional behaviour (presence of institutional incidents, serious charges or segregation), program reports (level of participation, motivation and assessed progress), and assessments for decisions under the Offender Security Level Reviews.

From this sample of 167 women offenders, a total of 275 OSL security review decisions were coded. The number of OSL decisions coded per offender ranged from 1 to 5 (M = 1.67, SD = 0.92). For the present sample, the period of time between security reviews (i.e., the review period) averaged 10.5 months (SD = 9.11). At the review date, the women serving fixed sentences were on average, 21 months from their statutory 4 release date (SRD), and 42 months from their warrant expiry date (WED). The women serving life sentences had served an average of 34 months at the time of their security review (SD = 42.59).

Results of Security Reviews

In the present sample of security reviews, the number of women classified as minimum security more than doubled from their previous rating to their present one. Of the 275 security reviews, 54.5% resulted in a lowered security level, 24% resulted in an increase, and 21.5% did not change (see Table 1).3

Table 1: Outcome of Security Level Reviews
Security
Status
N = 275(%)4 Med. To
Min.(%)
Max. to
Med.
Max. to
Min.
Min. to
Med.
Med. to
Max. (%)
Min. to
Max.
Lowered 150(54.5) 92(33.5) 58(21) 0
Raised 66(24) 25(9) 37(13.5) 4(1.5)
No Change 59(21.5)

2 Aboriginal women were intentionally over-sampled for this study. They comprise about 18% of the federal women inmate population.
3 The results of security reclassifications differ for male offenders. In Ontario, from July to December 1997, 26% of security reviews resulted in a lower security classification, 18% resulted in a highersecurity classification, and the majority (56%) remained at the same security level. While men tended to remain at the same security level more than women, women tended to be decreased in security level more often (Luciani, 1997).
4 Percent of total sample.

Selection of Variables

Dependent Variable

The focus of this study was to examine factors predictive of increases in security level.

As such, for the present analyses, women were separated into two groups based on the outcome of their security reviews: those for whom security levels was increased (24%; n = 66) and those for whom it was not (i.e., either decreased or remained unchanged; 76%; n = 209).

Predictor Variables

Numerous 'candidate' variables were selected for inclusion in the analyses.5

These included objective and subjective factors thought to influence the reclassification process (e.g., institutional adjustment, and assessed level of motivation). Variables suggested by critics of the classification process for women to be influential in the classification of women were also examined (e.g., ethnicity, the presence of psychological concerns). For a full list of 'candidate' variables, see Appendix A.


5 In the selection process, variables with 10% or more of the data missing were excluded from the analyses. Data with highly skewed distributions or an extremely low base rate were also excluded.

RESULTS

Frequency of Increases in Security Level

Reclassification to higher levels of security does not occur often. In fact, this sample of security reclassification data from 1993 to 1998 suggest that only 9% of women were reclassified from minimum to medium security upon review, and 13.5% of women were reclassified from medium to maximum security upon review (see Table 1). In other words, results of security reviews render the vast majority of women offenders at the same security level or reclassify them to lower levels of security.

Reasons for the Increase in Security Level

Security reviews can either be event-driven or routine in nature. It was expected that a high percentage of the reclassifications up to maximum security would be due to serious events that occurred just prior to the review (e.g., assault, escape, threats/disruptive behaviour). Alternatively, it was expected that relatively few women classified from minimum to medium security would have been involved in serious institutional incidents prior to their review.

The expectations were supported by the data. Specifically, all of the women who were reclassified from minimum to maximum security (n = 4) had been involved in serious events that prompted their security review and reclassification. Two of the women had been involved in assaults, one had engaged in threats and disruptive behaviour, and one woman had been charged with new offences. Further, 69% of the women reclassified from medium to maximum security had been involved in a serious institutional event that prompted the security level review. In contrast, only 24% of the reclassifications from minimum- to medium-security were event-driven.

Factors Predictive of Reclassification to Higher Levels of Security

A stepwise regression analysis was performed to identify those variables that significantly contributed to the prediction of increased security level. Table 2 displays the variables predictive of reclassification to higher security levels for the present sample of offenders, in order of their importance to the model.

Collectively, the five listed variables (non-cooperative attitude, incurring a number of serious institutional convictions, limited or no motivation to comply with their respective correctional plan or programming, possession or distribution of contraband, and history of unlawfully at large) accounted for 36% of the variance in decisions to reclassify women offenders to higher levels of security. Given that explained variance of 40% or higher in most social science research is considered impressive, the findings of the present analyses suggest good predictive accuracy. Below, each variable is discussed in greater detail.

Table 2: All Selected Predictor Variables and Incremental Contribution of Predictors
Predictor Simple r Multiple R²
Level of cooperation -.5381 .2900
Serious institutional offences with convictions .1849 .3258
Level of correctional plan and program motivation -.3162 .3425
Possession or distribution of contraband .2456 .3532
History of unlawfully at large .1456 .3619

(All correlation coefficients are significant at p<0.001 or higher)

Level of Cooperation

Through the use of offender progress summary reports, researchers coded the assessed level of cooperation for each offender in the sample. The level of co-operation was divided into three separate categories ("uncooperative", "partially cooperative", and "cooperative"). An "uncooperative" offender refused to participate in various institutional activities, such as work or programs, and was disruptive to other inmates and staff, and/or a disciplinary problem. An offender who was "partially cooperative" may have been somewhat of a disciplinary problem or somewhat disruptive, but also may have been occasionally cooperative. An offender who was 'cooperative' was not a disruption or disciplinary problem within the institution, and usually displayed positive interaction with institutional staff and participated in institutional activities. Of the 275 security review decisions sampled, 18.5% were rated as "uncooperative", 22.2% were rated as 'partially' cooperative, and 59.3% were rated as "cooperative".

Level of cooperation, which accounted for the majority of the variance in classification of women to higher levels of security, was primarily a subjective variable. However, further investigation indicated that numerous objective factors were associated with offenders' assessed level of cooperation. For example, poor institutional adjustment as measured through the presence of serious and minor institutional charges, the number of institutional incidents, assaults, positive urinalyses, and involuntary segregation terms were all variables that influenced (to varying degrees) women's perceived level of cooperation (see Table 3).

Table 3: Correlation Analysis of Objective Variables with Level of Cooperation
Objective Variable6 Simple r
Serious disciplinary offences -0.5809
Segregation/danger to others -0.5381
Number of institutional incidents -0.4692
Minor disciplinary offences -0.4641
Failure to comply with taking medications -0.4355
Number of positive urinalyses -0.3066
Source of distribution of contraband -0.2749
Number of assaults causing serious harm -0.2252

(All correlation coefficients are significant at p<0.0005 or higher)


6 All variables were measured during the security review period.

Results indicate that more than half of the women reclassified to higher levels of security were rated as 'uncooperative' with institutional staff and other offenders (54.5%; see Table 4). However, only seven percent of women who were classified to lower security levels or whose level remained unchanged were assessed as "uncooperative". In contrast, slightly less than 20% of women who were increased were rated as "cooperative". Interestingly, the women viewed as generally "cooperative", which increased in security level, were increased as a result of a threshold event, such as incurring an institutional charge for assault or attempting to escape.

Table 4: Level of Cooperation by Security Reclassification Outcome
Level of Cooperation No Change or
Decrease (%)
Security Level
Increased (%)
Uncooperative (n = 51) 15 (7.2) 36 (54.5)
Partial cooperation (n = 61) 44 (21.0) 17 (25.8)
Cooperative (n = 163) 150 (71.8) 13 (19.7)
Total (N = 275) 209 (100) 66 (100)

Serious Institutional Convictions

Institutional offences categorized as "serious" constituted a range of behaviours, including assault on staff or other offenders, threatening behaviour, or possession of a weapon, among others. These are offences for which the security of the institution, staff or other inmates may have been threatened. The number of serious offences resulting in convictions is clearly an objective factor. An offence is categorized into either "minor" or "serious" depending on the extent and nature of the offence. An institutional conviction for an offence indicates that unquestionable evidence exists to charge and convict the inmate for the offence committed. This means that a conviction is not based on the subjective opinion of institutional staff, but is based on the presence of a witness to the offence or a plea of guilt from the offender.

The number of times a woman was convicted of a serious institutional offence during the review period was the second of five variables that contributed to the prediction of an increase in security level (see Table 5). Examination of the distribution for this variable indicated that almost half (45.5%) of the women who were reclassified to higher levels of security had been convicted of one or more serious institutional offences during the review period. In contrast, only a quarter of women who were not reclassified to a higher level had been convicted of a serious institutional offence.

Table 5: Number of Serious Institutional Convictions by Security Reclassification Outcome
Number of Convictions No Change or
Decrease (%)
Security Level
Increased (%)
0 (n = 192) 156 (74.6) 36 (54.5)
1 to 2 (n = 49) 31 (14.8) 18 (27.3)
3 or more (n = 34) 22 (10.5) 12 (18.2)
Total (N = 275) 209 (100) 66 (100)

Level of Correctional Plan and Program Motivation

The level of correctional plan and program motivation was the third of five variables found to predict reclassification to higher levels of security. This variable, coded by researchers from examination of offender progress summary reports, had three levels.

An offender classified as possessing "no or limited motivation" refused to participate in the programming recommended to address those needs outlined in her correctional plan, or participated very sporadically. A "partially motivated" offender participated in programming with adequate attendance, and completed the necessary work sometimes or partially. An offender rated as having "full motivation" actively participated in programming recommended in her correctional plan, completed work most of the time, and consistently applied what she had learned to daily life situations.

Similar to the first variable, level of cooperation, program motivation comprised a subjective component. Women who did not feel the need to address particular needs identified in her correctional plan would be termed "unmotivated". However, many objective factors also contributed to assessment of this variable. For example, the amount of work completed for the programs, the attendance record, and the frequency with which the offender either applies or does not apply what she has learned through programming are concrete behavioral indices of the offender's level of motivation.

Therefore the impact of subjective assessment of motivation was moderated by the numerous objective factors that must be considered in making an accurate assessment of the offender's motivation.

Results of the analyses illustrate that an offender deemed as expressing "no or limited motivation" in programs or in fulfillment of her correctional plan is more likely to be classified to a higher security level than to remain at the same security level. In fact, more than one third (37.9%) of the women reclassified to higher levels of security were rated as having "no or limited motivation", whereas only six percent of women who were not increased had the same rating. Conversely, almost half (48.3%) of the security reviews which resulted in no change or a decrease in security rating were comprised of women rated as "fully motivated". A woman's motivation is, to a large extent, indicative of her level of institutional adjustment. A woman motivated to address her needs and participate in meeting the goals of her correctional plan is displaying more positive institutional adjustment and is less likely to present a disciplinary problem. Given this, women with "full and active motivation" were less likely to be reclassified to higher levels of security. As expected, women with "partial motivation" fell appropriately in between.

Table 6: Level of Correctional Plan and Program Motivation by Security Reclassification Outcome
Level of Motivation No Change or
Decrease (%)
Security Level
Increased (%)
No/Limited (n = 38) 13 (6.2) 25 (37.9)
Partial/Active (n = 127) 95 (45.5) 32 (48.5)
Fully/Actively (n = 110) 101 (48.3) 9 (13.6)
Total (N = 275) 209 (100) 66 (100)

Possession and Distribution of Contraband

The fourth factor found to influence the decision to reclassify women to higher security levels was the possession and distribution of contraband within the institution during the review period. This objective variable was assessed through the observation of incident reports logged in the institution (on OMS), and is an indicator of institutional adjustment and compliance with rules and regulations of the facilities.

The majority of women, who possessed and distributed contraband, specifically illegal substances or other material forbidden in the institution, were classified to a higher level of security. Specifically, 12% of the women increased in security were found to be sources of distribution of contraband, whereas only 1% of the women not placed in higher levels of security were connected with the possession or distribution of contraband.

Table 7: Possession and/or Distribution of Contraband by Security Reclassification Outcome
Possessing or Distributing
Contraband
No Change or
Decrease (%)
Security Level
Increased (%)
No (n = 265) 207 (99.0) 58 (87.9)
Yes (n = 10) 2 (1.0) 8 (12.1)
Total (N = 275) 209 (100) 66 (100)

History of Unlawfully at Large

Having a history of being unlawfully at large (UAL) from unescorted temporary absence (UTA) or community supervision was the fifth variable found to predict increases in security level. Women who have a history of breaches of trust pose a security risk; therefore case management officers appropriately consider this factor when making decisions regarding security level. In the present sample, 24.4% of the women who were not reclassified to higher levels of security had ever been UAL, and almost half (43.9%) of the women placed in increased security as a result of the review had been UAL at some point during their previous or current incarceration. It should be noted however, that there are strict criteria governing UTA eligibility. Only offenders classified as minimum or medium security are eligible to apply for, and participate in UTAs. Thus the findings regarding this variable are affected, to some extent, by this fact.

Table 8: History of UAL by Security Reclassification Outcome
Ever UAL No Change or
Decrease (%)
Security Level
Increased (%)
No (n = 195) 158 (75.6) 37 (56.1)
Yes (n = 80) 51(24.4) 9 (43.9)
Total (N = 275) 209 (100) 66 (100)

Summary

The aforementioned factors in combination with the five variables (uncooperative attitude, incurring a number of serious institutional charges with convictions, no or limited correctional plan and program motivation, possessing or distributing contraband, and a history of UAL) clearly predicted the outcome of reclassification decisions.

However, it is very important to stress that assessment of one of these variables to the exclusion of all others cannot explain why women are reclassified to higher levels of security. In other words, it is the presence of a combination of these five predictors that contributed to the ability to account for 36% of the variance in increased security classification.

Although the five factors described above accounted for 36% of the variance in security reclassification decisions, the majority of the variance remains unexplained. However, there are numerous variables not included in the regression analyses that might have been important to security level decisions. Often, single events prompt a reclassification to a higher security level, regardless of behaviour during the review Criteria include (but not limited to) a restriction against women classified as maximum security being granted UTAs period7. All of these factors, while not statistically predictive, are nonetheless a critical consideration when rendering security reclassification decisions.


7Criteria include (but not limited to) a restriction against women classified as maximum security being granted UTAs.

Other Factors Influencing Reclassification to Higher Levels of Security

In addition to the five significant predictors previously discussed, there are numerous variables that may affect the reclassification of women offenders to higher levels of security. These factors were not highlighted in the analyses for number of reasons (for example, low base rate), but nonetheless, are important to reclassification decisions.

Thus, variables such as assaultive behaviour, instigating disruptions, escape attempts, positive urinalysis tests, and being segregated due to posing a danger to others were examined as factors that influence security reclassification decisions to varying degrees.

Single events, which can alone result in an increase in security reclassification, might also account for unexplained variance for this sample. For example, a vicious assault on a staff member or another offender, or participation in activities which threaten the safety of the institution are events which can, alone, result in an increase in security reclassification.

Assaults Causing Serious Harm during the Review Period

Assaulting others, resulting in serious harm, is indicative of non-compliance and poor institutional adjustment. Women who had been reclassified to a higher level of security were over three times more likely to have committed an assault causing serious harm during the security review period (7.58% versus 2.39%).

Instigating Serious Disruptions during the Review Period

Offenders who instigated serious disruptions within the institution were also more likely to be reclassified to a higher security level upon review. Women who instigated disruptions were twelve times more likely to be reclassified to a higher level of security than those with no such history (6.06% versus 0.48%).

Escape Attempts during the Review Period

Also, illustrative of non-compliance with institutional rules, escape attempts during the security review period were found to effect reclassification outcome. Women reclassified to higher levels of security were three times more likely to have attempted to escape during the review period than other women (4.6% versus 1.4%).

Number of Positive Urinalysis Tests during Review Period

Positive urinalysis test charges are the result of an offender's consumption of a narcotic substance and the subsequent testing of that offender for traces of narcotics. Similar to serious institutional offences with convictions, this variable can be viewed as a form of non-compliance and institutional maladjustment that can be objectively measured.

While a urinalysis test is indicative of an offender's substance abuse problems, and thus a need for treatment, it also illustrates an inmate's willingness to depart from institutional regulations and become involved in offender sub-cultures. In the present sample, 27% of the women reclassified to higher levels of security had at least one positive urinalysis test during the review period, compared to only 8.5% of women not placed in higher levels of security.

Segregation and Danger to Others

The segregation of an offender due to the threat that she may pose to institutional staff or other offenders is an important indicator of institutional adjustment. In this sample, more than half of the women reclassified to higher levels of security (53%) had been placed in segregation for being a danger to others during the review period, whereas only 22% of the women not increased had been placed in segregation.

Factors Not Predictive of Reclassification to Higher Levels of Security

In general, dissidents of CSC's existing classification system (Canadian Association of the Elizabeth Fry Societies, 1998; Hannah-Moffat, 1999; Hannah-Moffat & Shaw, 2000) have not only criticized the approach to classification taken by the Service, but have also suggested that inappropriate and prejudicial factors are used to determine women's security classification. For example, being of Aboriginal ancestry and possessing psychological or psychiatric concerns have been suggested as factors overly considered in the reclassification process.

Aboriginal Status

Critics often state that the process of risk assessment is "racialized" (Hannah-Moffat & Shaw, 2000). While it is evident that Aboriginal women have been systemically overrepresented in the criminal justice system, some critics would suggest that the outcome of reclassification decisions are, to some degree, dependent on the ethnicity of the offender. For the present sample, however, there was no evidence to support this assumption (see Table 9). Even though Aboriginal women were over-sampled for this study, there was no difference in the rate of Aboriginal and non-Aboriginal women reclassified to higher levels of security (25% versus 24%). Although the Aboriginal women in the present sample were initially classified at higher levels of security8, their rate of increase was similar to that of non-Aboriginal women.

Table 9: Ethnicity by Security Reclassification Outcome
Ethnicity Non-Aboriginal (%) Aboriginal (%)
Decreased or Unchanged (n = 209) 155(76.3) 54 (75.0)
Increased Security Level (n = 66) 48 (23.7) 18 (25.0)
Total (N = 275) 209 (100) 66 (100)


8 A current study (Verbrugge & Wichmann, 2000, under review) suggests that Aboriginal women are initially classified at higher security due to the nature of their admitting offences.
Psychological and Psychiatric Concerns

Critics have also noted an apparent "medicalization" of women in society, including those in the criminal justice system (Hannah-Moffat & Shaw, 2000). If such medicalization of women offenders occurred systemically in correctional institutions, evidence of this would most likely exist at the level of security reclassification. This suggestion may then lead to the reclassification of women with psychological or psychiatric concerns to higher security levels.

The data suggest that a slightly higher percentage of women for whom psychological and psychiatric concerns were noted were reclassified to higher levels of security (35.6% versus 21.7%; see Table 10). However, a chi-square test revealed that these differences were not significant. Further, these differences may have been due to some third variable. For example, these women may also have displayed other institutional adjustment difficulties (institutional charges, etc.).

Table 10: Psychological and Psychiatric Concerns during Review Period by Security Reclassification Outcome
Security Level Decision No Concerns
Noted (%)
Concerns
Noted (%)
Decreased or Unchanged (n = 209) 180 (78.3) 29 (64.4)
Increased Security Level (n = 66) 50 (21.7) 16 (35.6)
Total (N = 275) 230 (100) 45 (100)

Regardless of any effect that psychological concerns may have on classification, with the current implementation of the Intensive Intervention Strategy in the women's regional facilities, psychological concerns are not, in the future, expected to be predictive of reclassification to higher levels of security. The Intensive Intervention Strategy brings more comprehensive mental health services into the regional women's facilities. This allows women offenders to address their psychological concerns while remaining in the same facility and same security level.

DISCUSSION

Subjective and Objective Factors in Security Reclassification

The security review process entails the consideration of multiple variables in ultimately deciding the most appropriate level of security required for each offender. As indicated by the results of the analyses, there is a definite interplay between subjective and objective factors in determining security reclassification. While three of the five most significant variables were objective, and the remaining two variables were subjective, it is clear that the interplay of both subjective and objective factors is integral to the security review and reclassification process.

The importance of subjective assessments by institutional staff who work with the women should not be underestimated. Their opinions or comments can often provide detailed information not available in institutional files, particularly given that the staff have continual direct experience and knowledge of the characteristics of the offender in question. Further, it is recognized that past strategies, which have been efficacious in meeting the legal requirement of placement into discrete levels of classification involved the use of multiple subjective factors. However, recent initiatives within the Service are leading the development of classification systems that are empirically supported and objectively based (e.g., gender-specific Security Reclassification Scales).

In the interest of addressing concerns of the critics, and more importantly, to ensure fair and equitable treatment of all offenders, it is necessary to identify the factors that affect placement into higher levels of security. Findings from this study revealed that both subjective and objective factors form the framework of these decisions. Five factors (uncooperative attitude, incurring a number of serious institutional charges with convictions, no or limited correctional plan and program motivation, possession or distribution of contraband, and a history of UAL) were clearly predictive of the outcome of reclassification decisions to increase women's security levels. Several additional objective variables (e.g., assaults, escape or escape attempts) also play a role in these decisions, though their base rates of occurrence were too low to support their inclusion in a regression model.

CONCLUSION

While it is evident that subjectivity exists in reclassification strategies, new initiatives attempt to limit the degree of subjectivity in this process and emphasize standardization and objectivity. In response to the need for a national standardized reclassification scale for women offenders, the Correctional Service of Canada is in the process of developing the Security Reclassification Scale for Women (SRSW), which is currently being re-validated as to its effectiveness in the reclassification process. The SRSW uses limited subjectivity in reclassification by relying primarily on objective and clearly measurable factors. By providing clearly articulated standards, women will be aware of how decisions are made to increase security levels, which, in turn, may have a positive impact on their motivations and determinations to reduce their security levels.

In summary, there will always be a degree of subjectivity inherent in the classification process. Subjectivity is an inevitable result in the human service industry. However, due to the importance of the reclassification and its numerous implications for women, the degree of subjectivity can be minimized by moving away from dependence on professional judgement and toward a more equitable, standardized and women specific assessment process. The objectivity of the SRSW as a reclassification measure, and the imminent implementation of the Intensive Intervention Strategy at women's facilities, enables the reclassification of women offenders to become a more objective, specific and informed process, while still remaining sensitive to the resources and needs unique to women offenders.

REFERENCES

Blanchette, K. (1997). Risk and need among federally-sentenced women offenders: a comparison of minimum-, medium-, and maximum-security inmates. Ottawa, ON: Correctional Service of Canada.

Blanchette, K. (1997). An examination of medium- and maximum-security federallysentenced women offenders. Ottawa, ON: Correctional Service of Canada. Canadian Association of the Elizabeth Fry Societies. (CAEFS, 1998). Position Paper Regarding the Classification and Carceral Placement of Women Classified as Maximum Security Prisoners. Ottawa, ON: Unpublished report.

Correctional Service of Canada. (1990). Creating Choices: Report to the Task Force on Federally Sentenced Women. Ottawa, ON: Correctional service of Canada.

Dowden, C., & Blanchette, K. (1999). An investigation into the characteristics of substance-abusing women offenders: risk, need and post-release outcome. Ottawa, ON: Correctional Service of Canada.

Hannah-Moffat, K. (1999). "Moral agent or actuarial subject: Risk and Canadian women's imprisonment". Theoretical Criminology, 3(1), 71-94.

Hannah-Moffat, K., & Shaw, M. (2000). 'Taking risks': gender, diversity, security classification and risk/needs assessment with Canadian federally sentenced women. Ottawa, ON: Status of Women Canada.

Luciani, F. (1997). The security classification of offenders: development of security reclassification protocols. Ottawa, ON: Correctional Service of Canada.

McDonagh, D. (1999). Federally Sentenced Women Maximum Security Interview Project: "Not Letting The Time Do You". Ottawa, ON: Correctional Service of Canada.

Verbrugge, P., & Wichmann, C. (2000, under review). Intake security classification of Aboriginal women offenders. Ottawa, ON: Correctional Service of Canada.

APPENDIX A: CANDIDATE PREDICTOR VARIABLES

Predictor

  1. Compliance with Institutional Rules and Regulations-Review
  2. Number of Serious Offences w/ Conviction - Review
  3. Assaults Causing Serious Harm - Ever
  4. Assaults Causing Serious Harm - Review
  5. Instigated Serious Disruption - Review
  6. Number of Minor Offences w/ Conviction - Review
  7. Number of Recorded Incidents - Review
  8. Number of Refuse Urinalyses - Review
  9. Number of Positive Urinalyses - Review
  10. Source Distribution of Contraband - Review
  11. New Charges - Review
  12. Involuntary Segregation: Danger to Others (# Times) - Review
  13. UAL from UTA/ Community Supervision- Ever
  14. Non-Violent Escape Attempt - Ever
  15. Non-Violent Escape Attempt - Review
  16. Detention Referral-Current Sentence
  17. Number of Suicide Attempts - Review
  18. Number of Self-Injury - Review
  19. Preventative Security Concern - Review
  20. Overall Case Needs - Current Sentence
  21. CRS Institutional Adjustment Group- Current Sentence
  22. CRS Security Risk Group- Current Sentence
  23. Psychological or Psychiatric Concerns Noted - Current Sentence
  24. Currently on Psychiatric Medications
  25. Substance Abuse Problem Rating - Review
  26. Correctional Plan - Program Motivation (1st Priority)- Review
  27. Correctional Plan - Program Progress (1st Priority)- Review
  28. Overall Criminal Risk - Current Sentence
  29. Quality of Interpersonal Relationships-Current Sentence
  30. Overall Marital/Family Adjustment-Current Sentence
  31. How often receives visits from Family/Community-Current Sentence
  32. Marital Status-Review
  33. Maintains Regular Contact with Family-Current Sentence
  34. Number of Successful ETA: Pers. Development - Review
  35. Number of Successful ETA: Total - Review
  36. Number of Successful ETA: Family Related - Review
  37. Number of Successful UTA: Family Related - Review
  38. Number of PFV - Review
  39. Pay Grade - Most recent