The Crown Files Research Project: A Study of Dangerous Offenders

James Bonta Andrew Harris
Solicitor General Canada Carleton University

Ivan Zinger Debbie Carriere
Carleton University Carleton University

Report No. 1996-01
Cat. No. JS4-1/1996-1
ISBN 0-662-62250-2

Acknowledgments

Numerous individuals facilitated this research project. We are deeply indebted to them, for without their assistance this project would have been impossible. Although we risk forgetting some, we would like to formally thank the many individuals who helped us along the way. Our apologies to those whom we may have forgotten.

For their assistance with the penitentiary file reviews we thank: Jim Blacker, Sonya Chupik, Pierre Couturier, John Duggan, Janice Grant, Ken Hazard, Irving Kulik, Wagdy Loza, Bob Lusk, Douglas McGregor, Larry Motiuk, Ken Payne, John Stonoski, Bob Watkins, Lois Wilde, and the Records Clerks, Case Management and other professional staff from the various penitentiaries.

In the survey of Crown prosecutors we thank Marie-Josée Bettez, Peter Ewert, Richard Isaac, Ian Scott, Murray Segal, and the prosecutors who gave of their time in order to participate in the interviews.

A number of people provided comments on earlier drafts of this report. We would like to thank the following for their helpful suggestions: Mary Campbell, Bob Cormier, Peter DeVink, Andrew Graham, Karl Hanson, Charles Haskell, Michel Larivière, Vanda Simanis, Carson Smiley, Nancy Stableforth, Dan Thompson, Suzanne Wallace-Capretta, Dave Whellams and Richard Zubrycki.

Finally, we would to thank Kimberly Fever for her excellent work in preparing this document for publication.

Table of Contents

Executive Summary

In response to one of the recommendations of the Federal/Provincial/Territorial Task Force on High Risk Violent Offenders (January, 1994), a research project was undertaken to provide empirical data that would assist Crown attorneys in the application of the Dangerous Offender provisions. The penitentiary and court files of 64 Dangerous Offenders (DOs) from Ontario and British Columbia were reviewed in order to construct a comprehensive description of this group of offenders. The DOs were then compared to another high risk violent group of 34 Detention Failures (DFs).

In the second part of the research project, 21 Crown attorneys (including one from Quebec) with experience in the application of the Dangerous Offender provisions were interviewed. They were asked to give their views on the strengths and obstacles related to the provisions, an assessment of other models for dealing with high risk violent offenders, and to evaluate a list of predictors of violent behaviour.

A Summary of the Major Findings

1. The majority of DOs are sex offenders

2. Antisocial personality features are frequent among DOs

3. Considerable information is available on DOs

4. There is considerable agreement between the two provinces

5. Crowns generally appear to target high risk violent offenders

6. 67% of Crowns were satisfied with the definition of a Serious Personal
Injury Offence (SPIO)

7. Crowns were generally satisfied with the present statutory criteria

8. Crown attorneys overwhelmingly endorse the DO legislation.

9. Operational Issues

1. DO applications require much more work than a normal criminal sentencing proceeding, and Crowns reported that they received little workload adjustment to deal with the application.
   
2. Approval for a DO application could be reduced to a lower and non-political level.
   
3. When first faced with a DO application, 52% of the Crown attorneys did not know what information was needed. The majority of prosecutors expressed a need for clearer policies and procedures on this matter. B.C. has recently instituted new policies and procedures; Ontario and Quebec are in the process of doing the same.

10. Crown's knowledge of predictors can be more empirically-based

11. Application of the findings to the screening of DOs

Introduction

High risk violent offenders and the danger they pose to public safety are ongoing public concerns. Governments have used many different methods to minimize the risk posed by such offenders. One approach has been to apply indeterminate sentences. Historically, Canada used Criminal Sexual Psychopath legislation (1948), then the Dangerous Sexual Offender (1960) law and in 1977, the Dangerous Offender provisions under Part XXIV of the Criminal Code of Canada. The earlier laws focused exclusively on sexual offenders and although they were criticized on a number of grounds, one of the more serious was the failure to target nonsexual violent offenders (Canadian Committee on Corrections; Ouimet Report, 1969). Thus, one goal of the Dangerous Offender (DO) provisions was to extend the Dangerous Sexual Offender (DSO) law to include nonsexual offenders who pose a serious risk to society.

More recently, community interest in managing violent offenders has intensified. Critics have suggested that the present DO provisions are inadequate and that new approaches must be taken to deal with high risk violent offenders. The few research studies that have examined the DO provisions found little change in the type of offenders being targeted compared to the earlier DSO legislation. For example, after the repeal of the Dangerous Sexual Offender provisions and during the period from 1977 to 1985, 78% of dangerous offenders were convicted of a sexual offence (Jakimiec, Porporino, Addario & Webster, 1986). In 1992, 90% of Dangerous Offenders had a history of a sexual offence (Pepino, 1993).

In January 1995, the Federal/Provincial/Territorial Task Force on High Risk Violent Offenders recommended to Ministers that a research project be undertaken to collect information on the characteristics of DOs and the factors that may have influenced a DO application. The goals of the research were to:

1) provide a comprehensive description of DOs;

2) identify information requirements and gaps;

3) assess the validity of the group targeted for DO application;

4) describe the Crown's evaluation of existing DO legislation and obstacles to its use; and

5) assess prosecutors' knowledge of indicators of high risk violent behaviour.

The information gathered may be helpful for a variety of purposes including the development of guidelines for identifying cases for DO applications.

The present report describes the results of that project. The report is comprised of two parts. Part 1 describes the review of penitentiary and Crown files which provide a description of Dangerous Offenders. Part 2 presents the results of an interview-based survey of Crown attorneys who have successfully prosecuted Dangerous Offenders. The survey results speak to the evaluation of the DO legislation and process by experienced prosecutors.

Part 1: The File Review

James Bontar Andrew Harris

Ivan Zinger Debbie Carriere

Introduction

There have been few systematic descriptions of the DO population. The earliest report was a preliminary unpublished study of the first 32 offenders declared Dangerous Offenders up to March, 1983 (Solicitor General Canada, 1983). Subsequently, Jakimiec et al. (1986) described 50 offenders found to be Dangerous Offenders between 1977 and 1985. Their review however, was limited to penitentiary files and stressed the incarceration experience. Motiuk and Seguin (1992) provided a statistical profile of 121 DOs incarcerated in federal penitentiaries in 1992. This was a "snapshot" of DOs at the time and although it captured the entire DO population, the variables used to describe these offenders focused on criminal history and factors of specific interest to the Correctional Service of Canada (e.g., geographical distribution of offenders, security levels, etc.).

The present project reviewed both penitentiary and court files and collected an extensive range of information on DOs from Ontario and B.C. As a result, the research afforded a detailed and comprehensive picture of the characteristics of DOs from these two provinces. Moreover, despite being restricted to Ontario and British Columbia, the present sample was larger than the sample of DOs described by Jakimiec et al. (1986).

Initially, the research project was to examine the files of successfully and unsuccessfully prosecuted DOs. However, after discussions with senior officials from Ontario and British Columbia we learned that unsuccessful applications were relatively infrequent and an attempt to collect information on unsuccessful applications would leave us with too few cases for analysis. Hence, another group of offenders was selected as a comparison for the DO files.

The comparison group consisted of federal offenders who were detained until warrant expiry, released, and subsequently re-offended violently. These Detention Failures (DFs) represent a group of offenders who, at least in retrospect, perhaps could have been designated DOs. The DF group permitted an evaluation of the validity of the dangerous offender process. That is, if offenders designated DOs have a high probability to re-offend violently then they might be expected to share many characteristics with the DFs. Further, any differences between the groups may inform prosecutors as to risk factors that should be considered in a DO application.

Method

Subjects

On May 3, 1995 there were 146 DOs in Canada. All but one were male with 67 held in Ontario and 37 in B.C. Two groups of federal offenders were selected for detailed file review. The first group comprised 64 offenders held under Part XXIV of the Criminal Code of Canada (C.C.C.), the Dangerous Offender provisions. From this Dangerous Offender group, 32 had been declared Dangerous Offenders in Ontario (47.8% of all DOs in the province) and 32 were from British Columbia (86.5% of all DOs in B.C). The present sample of 64 DOs represents 43.8% of the total DO population in Canada at the time of data collection.

The second group of high risk violent offenders consisted of 34 "Detention Failures"(DFs). A federal inmate who is judged likely to commit an offence causing death or serious harm prior to expiration of sentence may be detained until warrant expiry under the provisions of the Corrections and Conditional Release Act (1992). A Detention Failure was defined as an inmate who had been detained until expiration of sentence and who subsequently recidivated with a violent offence following release. Violent offenses included homicide, assault, robbery and sexual crimes.

The names of DOs were provided by the respective provincial Crown offices and subjects were selected in reverse chronological order. Offenders in this group were designated Dangerous Offenders between March, 1979 and March, 1995. Detention Failures were drawn from Ontario DFs described in an earlier study of high risk offenders (Motiuk, Belcourt & Bonta, 1995).

The Coding of Variables

A file coding manual was constructed consisting of variables drawn from a review of the research literature. Not only were basic demographic and criminal history information coded but also a number of variables suggested by research to correlate with violent behaviour. The initial coding manual was reviewed for its completeness and relevance and subsequently pilot tested on five files from Kingston Penitentiary. Revisions were made and a final pilot testing of 11 files from the Ontario Crown Office in Toronto and Kingston Penitentiary was conducted. The purpose of the pilot testing was to evaluate the availability and accessibility of the information.

Further revisions were completed and the final result was a coding manual consisting of 245 simple and composite primary variables. For 102 (42%) of these variables, the source of the information was recorded (e.g., Psychological assessments, Police reports, trial transcripts). In addition, whether the information came onto the file before or after the Dangerous Offender or Detention determinations was also recorded.

As noted earlier, we coded for variables which have been suggested by previous research to be related to recidivism (general and violent). In addition, we collected information on a small sub-set of miscellaneous, exploratory variables. A description of the 22 areas that were coded is presented in Appendix A. Appendix B shows the codes for the data sources and Appendix C presents the exploratory variables.

Measures

The coded variables provided sufficient information to score the Psychopathy Checklist - Revised (PCL-R). Drawing upon Cleckley's (1982) clinical description of the "psychopath", Hare (1985, 1991) has developed an objective diagnostic instrument called the PCL-R. The psychopath is a form of personality disorder typified by a lack of guilt and remorse, unresponsiveness to normal social controls and behaviour patterns that are generally antisocial. While not a risk prediction instrument in the strict sense of the term, this 20 item checklist has been shown to correlate with both general recidivism (Hart, Kropp, & Hare, 1988; Serin, Peters, & Barbaree, 1990) and violent re-offending (Harris, Rice, & Cormier, 1991).

In addition to the PCL-R, scores from the Statistical Information on Recidivism (SIR) scale were coded. The SIR was developed by Nuffield (1982) as an aid to parole decision-making in Canada. This fifteen-item scale is largely comprised of criminal history variables with fourteen of the fifteen items as static predictors. It is the policy of the Correctional Service of Canada (CSC) to conduct a risk assessment of all non-Aboriginal, male inmates using the SIR scale. The results reported here are taken directly from the SIR record forms on file.

Procedure

File coders were two Ph.D. candidates and one Masters candidate in Forensic/Correctional Psychology. This team of coders reviewed files from Crown Offices, CSC Regional offices, a federal records centre and eight Federal Correctional Institutions in Ontario and British Columbia. All files were coded independently. At the federal institutions the following files were reviewed: sentence administration, visits and correspondence, case management, psychology, discipline and dissociation, and preventive security. The data were analyzed using SPSS/PC+ (Norusis, 1990) and SPSS for Windows (Norusis, 1993).

Inter-rater Reliability. Eleven cases were coded by a third rater to assess inter-rater reliability. Inter-rater agreement among the three raters on the coded variables averaged 83.3%. The individual item agreement ranged from 100% to 54%. The lowest inter-rater agreement was for two items in the companions and friends section. Section inter-rater reliabilities ranged from 94.5% for Index Offence variables to 59% for the companions and friends section. The companions and friends section proved to have not only inadequate data but the information found in the files was highly subjective. As a result, data from the companions and friends section is not reported. When this section of five questions is removed from the inter-rater reliability calculations, the average inter-rater agreement increases to 85.2%.

Results

There were 64 Dangerous Offender (DO) files reviewed. Although the project examined many variables indicative of violent re-offending, it was also important to ask whether offender files provide sufficient information at the time of the index offence to permit the Crown to make a Dangerous Offender application. The coding scheme noted whether the information was available prior to the DO sentencing date or whether the information became available after the decision.

The results are presented in four sections. The first deals with a description of the DOs. The second section examines the point in time when critical information became available in the files, the third compares DOs from Ontario and B.C., and the last section compares the DOs and the DFs.

I. Description of Dangerous Offenders

Table 1 presents a general summary of personal-demographic information on the DOs. DOs scored in the "average" range of intelligence and their average age was 34.4 years. Nearly all (95.2%) were Caucasian; there were three Aboriginal offenders and there was no racial information on one subject. In the general penitentiary population, 70.6% of the population is Caucasian (Correctional Service of Canada, 1994). Approximately two-thirds had completed some high school and nearly half were single.

Table 1. Personal-Demographics of Dangerous Offenders at Index Offence¹

^{1 numbers vary due to insufficient file information}

The Index Offence. Part XXIV of the C.C.C. sets out a number of criteria for a designation of Dangerous Offender. The person must have committed a "serious personal injury offence" (SPIO) and be judged likely to cause future serious harm. Additional criteria considered in a DO application may include sexual offenses or offenses of a "brutal" nature. Information on the index offence as it pertains to Part XXIV along with the Court's reasons for a DO finding are presented in Table 2.

Table 2. The Index Offence for DOs

¹ Sex offenders only

As expected, the majority of DOs were convicted of a sexual offence. There were only 5 non-sexual cases. Over half of the DOs were convicted pedophiles with an average of 3.2 known victims (the number of victims ranged from one to 23). Over half of the victims were under the age of 16. The trend for the sexual Dangerous Offenders to prefer young victims was evident when we examined the age of the victims in more detail and found that 43.1% of the victims were under the age of 13. However, sample sizes were quite small and did not show statistically significant differences.

Brutality is one of the criteria for a DO designation. The Courts cited brutality as a factor in 8 of the 44 DO cases (18.2%) for which we had information. It was never cited as the sole criterion. Using file information, the brutality of the index offence was measured in numerous ways. One measure involved a rating of the index offence on a scale of 0 to 5. A score of 0 indicated no evidence of brutality and 5 indicated extreme physical assault (e.g., torture, 10 or more stab wounds). The mean score for DOs was 1.2. Nearly 70% of DO cases showed some evidence of brutality. Examining only cases of brutality that exhibited extreme physical violence (ignoring excessive threats), 19.6% of the DOs showed extreme physical signs of brutality.

Another measure of offence seriousness was the degree of victim physical injury. Information from police, court documents and victim impact statements were scored on a scale ranging from 0 (no evidence of victim injury) to 6 (death) and produced an average score of 2.5. None of the DOs committed an offence that resulted in death. Many crimes resulting in death (e.g., homicide) have mandatory lengthy sentences and as such, a DO application may be unnecessary. Approximately 62% of the DOs injured their victims physically.

The nature of the Index offence is but one consideration in a DO application. The Criminal Code provides a number of additional factors including a failure to control sexual impulses and repetitive/persistent aggressive behaviour. We now turn to these elements.

Criminal History and Persistence of Criminal Behaviour. An extensive criminal history is seen as a sign of likely re-offending. The Court cited a pattern of persistent and repetitive behaviour in 44% of DO cases and a failure to restrain behaviour in 62% of DO declarations (the reasons given by the Court are not independent; more than one factor can be cited for a case). The review of files did show that DOs have, on average, extensive criminal histories (Table 3). However, official histories of violent behaviour were not as frequent as we expected. The majority of DOs did not have a history of assault or violent sexual offenses.

Table 3. The Criminal Histories of DOs

Failure to Control Sexual Behaviour. Fifty-nine of the 64 DOs (92.2%) were convicted of a sex offence. However, not all sex offenders are declared DOs. Therefore, we examined in detail the sexual histories of the Dangerous Offenders whose index offence was sexual in nature in order to gain a better idea of the uniqueness of this group.

The DOs (sex offenders only) showed early sexual offending behaviours. For the 59 dangerous sex offenders, 96.6% showed evidence of forcible sexual activity prior to the age of 16 years. Information from CSC files revealed that the DOs admitted many sexual offenses for which they

were not charged, an average of 27.2 offenses. However, the range of undetected sexual offenses was quite large (0 to 201) which skews the average.

To summarize our first major finding:

The majority of DOs are sex offenders

Antisocial Personality. The social science and legal literatures on the assessment of dangerousness assign considerable importance to a constellation of personality characteristics referred to as antisocial personality. Impulsiveness, egocentrism, lack of empathy and thrill seeking are some of these characteristics. For many, antisocial personality is synonymous with a failure to control one's behaviour, a blatant disregard for the welfare of others and a high propensity to continue in an antisocial lifestyle. The assessment of antisocial personality characteristics and their relevance to a declaration of DO status is usually provided through the testimony of psychiatrists during the proceedings.

A number of approaches can be taken to the assessment of an antisocial personality. In this project, estimates of the incidence of antisocial personality were calculated in three ways. First, we searched for statements of a diagnosis of an Antisocial Personality Disorder (APD) in psychological, court and penitentiary files. A second, but more stringent approach, was to match the information in the files to the diagnostic criteria for Antisocial Personality Disorder described in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; American Psychiatric Association, 1994). Third, the file information was used to score the Psychopathy Checklist (PCL-R; Hare, 1991). The PCL-R provides the most stringent assessment of a personality disorder although some may argue that Clinical Psychopathy may measure a construct somewhat different from the DSM-IV criteria (Hare, 1991). DSM-IV criteria are dependent upon behavioral evidence of antisocial behaviour whereas the PCL-R also taps personality characteristics (e.g., shallow affect, irresponsibility).

A psychiatric diagnosis of APD was found in 72.9% of DO files. When the more stringent DSM-IV criteria were used, the rate of APD fell to 54%. These rates are similar to those reported in forensic/correctional populations. Sufficient information was available in 48 DO files to calculate PCL-R scores. Following Hare's (1991) recommended cut-off score of 30 for a diagnosis of psychopathy, 39.6% (19) of the DOs were classified as psychopaths. This base rate of psychopathy is almost twice as high as the rate typically found in offender populations (15 to 25%; Hare, 1991). The rate may actually be higher if the assessment was conducted with the benefit of interviews. Wong (1988) has noted that a file-based scoring of the PCL-R may yield a conservative estimate of psychopathy.

In closing this section, it is noteworthy that DOs, as a group, appear to be relatively free from a major mental illness characterized by extreme moods and poor contact with reality. Among DOs, 8.5% were diagnosed with a schizophrenic disorder. This finding suggests that offenders who behave in a violent manner and who have a serious mental illness may be dealt with under civil commitment procedures.

Antisocial personality features are frequent among DOs

II. The Availability of Information

Past inquiries into violent victimization have observed that critical information is frequently missing or unavailable. In order to empirically address this observation, the type of information available to Crowns in DO applications and when this information became available was coded. Not all the information in the project could be coded as to the point in time when it became available. However, a limited number of items were assessed as to whether the information was available prior to the DO sentencing decision. Table 5 summarizes this information. On the whole, considerable information was available to Crown attorneys prior to the Dangerous Offender finding.

Table 5. Availability of Information Prior to DO Decision (%)

¹ Sex offenders only

Considerable information is available on DOs

The finding that there is considerable information available on DOs does not necessarily mean that it is the right information needed to bring forward a successful DO application. As we will discuss later in the report, predictors of sexual recidivism form an important constellation that should be considered. Unfortunately, we were unable to adequately assess the availability of this type of information.

III. Ontario and B.C. Similarities

Soon after the implementation of the DO provisions, it became apparent that the provisions were not uniformly applied across the country. Ontario has the largest number of DOs but B.C. the largest per capita. Quebec has had only one case declared a Dangerous Offender. This does not mean that Quebec has no dangerous offenders, but that the province may be dealing with them in a different manner (e.g., civil commitment through provincial mental health legislation). Uniformity of application involves two questions. First, are the DO provisions being applied to similar cases? And second, are the provisions being applied to all cases to which they ought to be applied?

The first question was answered by comparing the DOs from the provinces of Ontario and British Columbia. The second question was not directly addressed by this project, but comparisons with

the Detention Failure group (next section) provide some indications of the appropriateness of cases declared DOs.

Very few differences were found between DOs from Ontario and DOs from B.C. Table 6 shows only the analysis for the Index offence. In terms of personal demographics there were no differences between the DOs from the two provinces. They were similar in marital status, race, grade completed and employment status. Similarly, we found no differences in the other variables analyzed (e.g., alcohol/drug dependency, diagnosis of antisocial disorder or child molesting). \

Table 6. Comparison of DO Cases from Ontario and B.C. on the Index Offence(%)

¹ Sex offenders only
ns = nonsignificant

In summary, we could not find any substantial degree of difference in the characteristics of the offence or the offender from the two provinces. The Courts appear to deal with similar cases in a uniform manner, at least in Ontario and B.C.

There is considerable agreement between the two provinces

IV. Comparing Dangerous Offenders with Detention
Failures

In the previous section on regional similarities, the question was raised as to whether all cases eligible for a DO application were being identified. To properly answer this question would have required a very different research methodology, but the question raises the general issue, are the offenders labelled as DOs actually high risk violent offenders?

An attempt to answer this question was made by comparing the DOs to a known group of high risk violent offenders, namely, 34 Detention Failures (DFs). The DFs are the only offenders who actually committed another violent offence after they were identified as potentially violent. If the DOs appear similar to the DFs, then we are led to conclude that high risk violent offenders are being appropriately targeted in Ontario and British Columbia. Any differences found between DOs and DFs may provide useful information as to the indicators of violent re-offending which could then be used in Dangerous Offender applications.

Table 7 presents a general summary of personal-demographic information on the DOs and DFs. This table is similar to Table 1 but now includes the DFs. Compared to the DFs, DOs were older and more likely to be Caucasian. This latter difference is true for the entire DO population where approximately 85% of DOs are Caucasian (Motiuk & Seguin, 1992). However, the proportion of Caucasians among the DOs is more exaggerated in our study due to the absence of sampling from the Prairie region. On all other personal-demographic variables there were no statistically significant differences.

Table 7. Personal-Demographics of DOs and DFs at Index Offence¹

Table 8 compares the DOs and the DFs with respect to the index offence. The most striking difference was that DOs were almost three times more likely to have been convicted of a sexual offence (92.2% vs 35.3%). There were only 5 non-sexual cases among the DOs. When the DOs were compared to DFs who were convicted of a sexual offence, the DOs averaged more known victims (3.2 vs 1.5). There were no other significant differences with respect to the sexual crimes of DOs and DFs.

Table 8. The Index Offence for DOs and DFs.

¹ Sex offenders only
² Score range from 0 (no injury) to 6 (death)

The only other differences found between the groups were the incidence of substance use and evidence of brutality during the index offence. For DFs, substance abuse was more common in both the offender and the victim. One interpretation of this finding may be that intoxication, particularly in the victim, could be viewed as a factor that may partly work against a DO application.

Nearly 70% of DO cases showed some evidence of brutality compared to 48.4% of DFs. However, if we examine only cases of brutality that exhibited extreme physical violence (ignoring excessive threats), then a slightly different picture emerges. The DFs, two of whom committed an index offense resulting in death, inflicted more serious physical signs of brutality than did DOs (35.5% vs 19.6%). However, the differences were not statistically significant. Lower rates of serious physical brutality in the DO sample may stem from the fact that many of the DOs were pedophiles who used nonphysical methods of coercion.

An examination of the criminal histories of DOs and DFs showed many similarities. The two groups differed only on their history of failure on probation/parole with almost all the DFs, not surprisingly, having a history of failure while on conditional release (Table 9).

Table 9. The Criminal Histories of DOs and DFs.

¹ The seriousness of criminal history was based on the Cormier-Lang Criminal History Scale (Webster et al., 1994).

Finally, we calculated the seriousness of the offender's criminal history using the Cormier-Lang Criminal History Scale (Webster, Harris, Rice, Cormier & Quinsey, 1994). This scale is a revision of the Akman and Normandeau (1968) criminal history seriousness scale. Various offenses receive ratings (1 for a B&E to 28 for murder) and the ratings are multiplied by the lifetime number of offenses. We found no statistically significant differences between DOs and DFs in the seriousness of their criminal histories.

Another possible method of validating the DO applications is to assess objectively their level of risk. In the early 1980s, the Ministry of the Solicitor General developed an objective, actuarial risk scale: the Statistical Information on Recidivism or SIR scale (Nuffield, 1982). Based upon file information and offender interviews, offenders are scored on 15 variables and a total score calculated. Consequently, they can be situated within one of five risk categories. There is considerable evidence indicating that scores on the SIR are indicative of future recidivism, both general and violent (Bonta, Hann, Harman & Cormier, 1996; Bonta & Hanson, 1995; Nuffield, 1982).

Scores on the SIR ranged from -18 to 24 (lower scores indicate higher risk to recidivate). The two groups showed no statistical difference in mean scores (-5.3 vs -1.6, t= 1.85, df= 88, p< .07).

The risk levels of DOs, as measured by the SIR, were more evenly distributed among the prognostic categories than were the risk levels for DFs. For example, 43.6% (n = 27) of the DOs scored within the Good and Very Good risk categories compared to 25% (n = 7) of the DFs. Unexpectedly, 13 DOs were assessed by the SIR as "Very Good" risks. However, all of them were sex offenders (as was the one DF in the Very Good risk category). The SIR is recognized as

problematic with sex offenders and CSC's Research Unit has taken corrective steps to improve the scale's application. Until revisions to the SIR are complete, the use of the scale is not recommended for DOs owing to the high proportion of sex offenders.

We examined in more detail the sexual histories of the Dangerous Offenders and Detention Failures who were convicted of a sexual crime as the index offence. It was expected that the sexual Dangerous Offenders would have a more extensive history of sexual deviance. Both groups showed early sexual offending behaviours. All 12 of the DFs who committed sexual crimes, for which we had information, had prior to the age of 16 forced a victim into sexual activity . For the 59 Dangerous Sex Offenders, 96.6% also showed evidence of early forcible sexual activity. The general sex offender research also shows that early sex offending is an important predictor of sexual recidivism.

Although both groups began their sexual crimes relatively early in life, they still showed some important differences. The DOs admitted to many more sexual offenses for which they were not charged compared to the DFs (27.2 vs 0.82, t = 2.61, p < .01). Furthermore, the DOs had more adult female victims (2.8 vs 1.2, t = 2.77, p < .05) and more female child victims (2.6 vs 0.9, t = 2.80, p < .01). This may however, be an artifact of age. As DOs are older than DFs, they would have had more time to build an extensive criminal record.

Finally, no matter how antisocial personality was measured, there were no differences between the DOs and the DFs. A psychiatric diagnosis of APD was found in 72.9% of DO files and 73.1% of DF files. Using DSM-IV, 54% of DOs met the diagnostic criteria for APD and for the DFs, the rate was 64.7% (X ² = 1.04, df = 1, ns). There were no significant between group differences when the PCL-R was used to assess Clinical Psychopathy. The average score on the PCL-R for the DOs was 27.6 and 27.0 for the DFs (t = .50, df = 80) with 39.6% (19) of the DOs classified as psychopaths compared to 32.4% (11) of the DFs (X ² = .45, df =1, ns).

The one clinical-personality difference found was that DFs were more likely to be diagnosed with a schizophrenic disorder (26.9% vs 8.5%, X ² = 5.07, p <.05).

Crowns generally appear to target high risk violent offenders

Summary of Part 1

In the file review of 64 DOs from Ontario and B.C., we found that 92.2% of the DOs were convicted of a sexual offence. Moreover, the Court gave a "failure to control sexual impulse" as the reason for the DO declaration in 73.3% of the cases. In practice, the present DO provisions still appear to target the same group of offenders selected by the previous Dangerous Sexual Offender legislation.

The majority of the DOs could also be described as having a personality disorder but the extent of the disorder varied with the type of assessment used. When relying on DSM-IV criteria and psychiatric judgement, the incidence of an Antisocial Personality Disorder appears no different from that found in general offender populations. However, the incidence of Clinical Psychopathy as measured by the PCL-R was almost double the rate found in forensic and correctional populations.

Both Ontario and B.C. appear to identify similar cases. Whether these cases are actually high risk violent offenders was partially answered by comparing DOs to DFs. In general, the two groups were similar suggesting that DOs comprise a high risk violent group. The one major difference between the groups was that there were fewer sex offenders among the DFs (35.3% of the DFs were convicted of a sexual offence). However, the sex offenders among the DFs were very similar to the DO sex offenders.

Part 2: Crown Attorney Interviews

Ivan Zinger James Bonta

Introduction

Part 2 of the Crown Files Research Project describes the results from the Crown attorney interviews. The interview survey was designed to identify the factors considered by Crown attorneys when initiating a DO application and to identify some of the obstacles in processing such an application. In addition, the prosecutors were asked to give their opinions on various proposals for dealing with high risk violent offenders and to rank indicators of risk for future violence.

Part 2 of the report is divided into six sections: (a) a methodology section which outlines the interview procedures and describes the respondents, (b) a description of the Crown attorneys' views regarding the current DO legislation, (c) a description of the prosecutors' views with regard to procedural issues, (d) a summary of the Crowns' reactions towards possible reforms, (e) a ranking of the predictors of violent recidivism by Crown prosecutors, and (f) a brief summary.

A. Method

Twenty-one Crown attorneys were interviewed. Eleven were from Ontario, nine from British Columbia, and one from Québec. Only one attorney was interviewed in Québec because only one offender had been declared a DO in that province. All of the prosecutors were selected because they had direct experience with DO applications. Semi-structured interviews were conducted with each of the participants (Appendix D shows the interview form).

Names of the Crown attorneys were obtained from the Criminal Law Division, Ontario Ministry of the Attorney General, the Office of the Assistant Deputy Minister, Ministry of the Attorney General of British Columbia, and the Direction générale des affaires criminelles, Ministère de la justice du Québec. Eighteen Crown attorneys were contacted and one declined to participate due to trial commitments. An additional four prosecutors were referred by the Crown attorneys who were initially contacted. Taken together, the Crowns had successfully processed 32 DO applications, and were involved in nine ongoing applications. The majority of the successful applications were recent: 76% were processed between 1992 and 1995.

The characteristics of the Crowns were as follows:

• 71% were males and 29 % were females
• 71% were assistant Crowns and 29% were senior Crowns
• The average number of years of experience as a Crown attorney was 14

Considering the overall average years of experience, DO applications were obviously assigned to experienced Crown attorneys.

B. Views on the Dangerous Offender Legislation

(a) Characteristics of the offence and offender

The Report of the Federal/Provincial/Territorial Task Force on High Risk Violent Offenders suggested that Crown attorneys be asked the following questions: (1) what were the characteristics and circumstances of the offence that prompted a DO application? and (2) what were the characteristics of the offender that initiated a DO application? The responses were diverse with relatively few common responses. Table 10 sets out the most frequent responses to the question on the characteristics of the offence that led to a DO application, and Table 11 indicates the responses regarding the characteristics of the offender.

Table 10. Offence Characteristics and Circumstances initiating a DO application

Characteristic/Circumstance	Frequency	%
Sexual offence	18	(87)
Child victim	10	(48)
Violent offence	6	(29)
Reoffended while on parole	5	(24)
Repetitiveness/number of victims	5	(24)
Brutal nature of the attack	4	(19)
Inability of Parole Board to keep dangerous offenders incarcerated	4	(19)

Table 11. Characteristics of the Offenders initiating a DO application

Table 10 shows that Crowns initiated DO applications primarily in instances of sexual offenses and offenses against children. With 18 (87%) of the prosecutors stating that the commission of a sexual offence led them to initiate a DO application, it is hardly surprising that we found from the file review that 92.2% of the DOs were convicted of a sexual offence.

Approximately one-quarter of the prosecutors stated that the DO provisions could be used to remedy perceived deficient parole practices. For example, 5 (24%) Crowns viewed a failure on parole as triggering a DO application and four Crown attorneys (19%) expressed reservations about the Parole Board's ability to keep potentially dangerous offenders incarcerated.

Crowns gave particular importance to two offender-based factors: criminal record and psychiatric diagnoses (Table 11). Both of the variables are regarded by the research as important in the assessment of violent offenders. More specifically, the presence of a psychiatric diagnosis of Antisocial Personality Disorder (APD) was seen to favour a DO application. Recall from the file review, that up to 73% of the DOs met the criteria for APD. Nearly two-thirds of the Crown attorneys believed the presence of such a diagnosis indicated that the accused posed a threat to the public. Offender risk assessment research supports the view that an antisocial personality is indeed a risk factor.

(b) Serious personal injury offence and severe psychological damage

Section 752 of the Canadian Criminal Code, R.S.C. 1985, c. C-46 defines a Serious Personal Injury Offence (hereinafter SPIO) as: (a) an indictable offence punishable by at least 10 years imprisonment, other than high treason and first and second degree murder, involving the use or attempted use of violence against another person, or involving conduct endangering or likely to endanger the life or safety of another person or inflicting or likely to inflict severe psychological damage upon another person; or (b) an offence or attempt tocommit sexual assault, sexual assault with a weapon, threats to a third party or causing bodily harm, or aggravated sexual assault.

The majority of Crown attorneys (67%) were generally satisfied with the current definition of SPIO. However, half (52%) found the 10 year minimum requirement to be inadequate. Several Crowns suggested that the following additional offenses, which currently provide for sentences of under ten years, should be included under the definition:

• Uttering threats
• Sexual interference
• Stalking (Criminal Harassment)
• Common assault
• Attempt to assault with a weapon

Concern about the exclusion of some of the above offenses centred around the problem of waiting until, for example, a threat had been acted upon, before taking action. It is unclear whether the prosecutors would like to see the minimum requirement lowered to five years or whether the maximum penalty for the above-listed offenses should be raised to 10 years. Twenty-nine per cent of the prosecutors suggested that a definition of SPIO should not be restricted to specific offenses and should explicitly include all sexual offenses and offenses against children.

While three-quarters of the Crown attorneys believed the term "severe psychological damage" should be included in the definition of a SPIO, almost half (43%) commented on its inadequacy. A third of the Crowns stated the requirement is unnecessary in all sexual offenses and offenses against children and that psychological harm should be presumed in such cases.

From the file review, approximately half the DOs were sex offenders who had victimized children. As expected, prosecuting these offenders presents a number of challenges. Five (24%) Crown attorneys said that it is difficult to establish severe psychological damage when children are victims because evidence of psychological harm may not emerge until later in their lives. As well, some victims may not be able to articulate or foresee what psychological problems could arise. Moreover, the need to prove this criterion may compel reluctant prosecutors to ask fragile victims to testify and open up their psychological histories. It was suggested that expert testimony on what damage can be expected should suffice.

67% of Crowns were satisfied with definition of a SPIO

(c) The four statutory criteria

In order to obtain a DO declaration, evidence must be brought before the Courts to establish one of the following:

1. A pattern of repetitive behaviour and a likelihood of causing death or injury to other persons, or inflicting severe psychological damage, through failure to restrain behaviour;
2. A pattern of persistent aggressive behaviour, showing a substantial degree of indifference respecting the reasonably foreseeable consequences to other persons;
3. Any behaviour that is of such brutal nature as to compel the conclusion that the behaviour in the future is unlikely to be inhibited by normal standards of behaviour restraint; or
4. A likelihood of causing injury, pain or other evil to other persons through failure to control sexual impulses.

When the Crown attorneys were asked if they believed that their colleagues were generally knowledgeable about the legislative criteria, the majority (75%) said no. However, when asked if they thought that they were capable of identifying cases for a DO application, 71% said yes.

The prosecutors were asked if they had any difficulties with the above-mentioned statutory criteria. The majority (72%) were content with the existing criteria. The following are specific comments by the attorneys who noted some problems with the legislative criteria.

(i) Criteria 1 and 2: Repetitive and Persistent Behaviour

By definition, a pattern is repetitive and persistent, thus making the wording redundant and the phrases "showing a failure to restrain his behaviour" and "showing a substantial degree of indifference" are superfluous. The second criterion may not be necessary because it is implicitly expressed in the first. Finally, there are difficulties in proving a pattern when an offender has a very versatile offence history or when a lengthy period of time has elapsed between the offenses.

(ii) Criterion 3: Brutality

The file review findings were consistent with the views of the Crown attorneys: this criterion is infrequently used. Only 18.2% of actual DO decisions cited brutality despite the fact that 69.6% of DO cases showed signs of brutality (see Table 8). The infrequency of this citation in light of the actual incidence of brutality may be due to two factors. First, brutality may be subsumed under other criteria (i.e., failure to control sexual impulse and a pattern of aggressive behaviour). Second, the term "brutality" may be difficult to define. A number of prosecutors found the term too subjective.

(iii) Criterion 4: Failure to Control Sexual Impulse

This subsection forms a key part in almost all DO applications with 92.2% of DOs in our sample convicted of a sexual offence. One specific concern for the Crown attorneys had to do with proving a likelihood of injury, pain or evil. Some prosecutors commented that showing proof may unduly subject victims to re-victimization by requiring them to testify.

Crowns were generally satisfied with the present statutory criteria

(d) Mandatory testimony of psychiatrists

The DO hearing requires the testimony of at least two psychiatrists, one of whom is nominated by the defense. The Crowns were asked whether the mandatory requirement for two psychiatrists to testify was necessary for establishing (1) whether the offender is likely to recidivate violently (dangerousness), and (2) whether the offender is treatable.

On the issue of dangerousness, 57% of Crown attorneys believed the requirement was necessary. The percentage agreement increased to 76% on the issue of offender treatability. These prosecutors stressed the seriousness of the DO application process, the exceptional nature of the disposition, and the need for an adversarial-based process where each party has an opportunity to call its own experts.

The dissenting Crown attorneys maintained that the need for expert testimony on these two issues should be left entirely to the Crown's discretion. Four prosecutors (19%) stated that demonstrating dangerousness is largely self-evident and another four felt that the mandatory testimony should be broadened to include psychologists. A number of Crown attorneys pointed to the problem of uncooperative offenders. They argued that the defense often has an unfair advantage when evidence from a Crown-appointed psychiatrist is based only on a review of an offender's correctional files whereas a defense-appointed psychiatrist has had the opportunity to interview the offender.

Four Crown attorneys provided the following technical comments:

1. provide guidelines as to the content of psychiatric testimony and include them in the C.C.C.;
2. section 756 of the C.C.C., which empowers the Court to direct an offender to attend for psychiatric observation, runs counter to frequent instructions by their counsel not to participate in the assessment; and
3. there is no provision to limit the time taken for the nomination of the defense psychiatrist thereby resulting in potential and needless delays in the proceedings.

(e) The consent process

A DO application requires the consent of the Attorney General of the Province. After the consent has been obtained, and after proper notice, a DO hearing is held. Failure by the Crown to give notice to the offender that the Crown intends to bring a DO application prior to his or her plea does not offend the principles of fundamental justice [R. v. L. (1987), 37 C.C.C. (3d) 1 (S.C.C.); R. v. B.(A.J.) (1987), 34 C.C.C. (3d) 249 (Nfld.C.A.)]. From discussions with officials, this consent process appears to insure that applications are successful. The success rate for applications where consent was obtained was 97%.

The majority of Crowns (76%) felt that the consent requirement was necessary for a variety of reasons: (1) as a safeguard against abuse, (2) to maintain consistency and uniformity, (3) because of the seriousness of the consequences for the offender, and (4) to ensure the quality of application through a process that maximizes scrutiny. However, even among the prosecutors who favoured the consent procedure, 50% said the consent of a Regional Director or the Office of the Assistant Deputy Minister, as the case may be, should be sufficient. Some Crown attorneys expressed reservations about the fact that the ultimate decision is made by a political entity, the Provincial Attorney General, who does not necessarily have a legal background.

The few Crown attorneys who stated that the consent requirement was unnecessary raised two general arguments against it. First, they asserted the process is intimidating and deters applications because it is expensive and time consuming. Second, these Crowns suggested that the need for consent in DO applications was anomalous given the absence of a consent requirement in prosecutions for 1st, or 2nd degree murder, where there are similar serious consequences for the accused.

C. Procedural Issues

(a) Timing of the decision to proceed with a DO application

Crown attorneys were asked when in the process they considered making a DO application. The majority (76%) stated that the decision to proceed with the application was made shortly after the arrest. This finding highlights the importance of the police in providing information needed to determine whether a DO application is warranted. The present study did not review what level of information is required and provided at the arrest stage but this appears to deserve further study.

(b) The role of the media

In response to questions about media coverage on serious crimes, 95% of Crowns stated that media coverage did not play a role in the decision to proceed with a DO application. As one prosecutor commented, initially the media's information on an accused is limited to the alleged offence. Since the decision to proceed by way of a DO application is generally made shortly after arrest, the media may not have sufficient time in which to exert an influence. Some prosecutors stated however, that media coverage may influence the consent process because the final decision rests with a political entity, the Attorney General of the Province.

(c) Availability of information

Despite reporting that the decision to proceed with a DO application begins soon after arrest, the majority of Crowns (62%) felt that there was inadequate information upon arrest to determine adequately whether a DO application is warranted. It appears that the initial review of the offence and criminal history alert Crowns to the possibility of a DO application and this triggers a more extensive collection of information.

The Crown attorneys were asked to rank, in order of importance, eight sources of information needed when contemplating a DO application. Table 12 lists the average ranking (1 being the most important source of information and 8 being the least important).

Table 12. The Ranking of Information

Table 12 shows penitentiary files and victim impact statements ranked low. The low ranking of penitentiary files may be due to some uncertainty among Crowns as to what these files actually contain. From our experience with these files, they are an invaluable source of information. They

include, among other things, police and pre-sentence reports, psychological reports, and whether treatment was accepted or refused. Efforts to educate Crowns as to the content of penitentiary files should be considered. With respect to victim impact statements, their relative lack of importance may simply be the result of prosecutors' reluctance to request fragile victims to testify in open Court.

The Crown attorneys were then asked: (1) how readily available are the various sources of information listed in Table 12, and (2) to comment on the difficulties they faced when attempting to access these sources. Table 13 shows the average level of difficulty in accessing these sources. The sources were rated as follows: 1) easily available, 2) difficult but available, and 3) virtually impossible to access.

Table 13. Accessibility of Information

Notable is the finding that the Crown attorneys rated penitentiary files as easily accessible. This is puzzling given that these files were also rated as not very important sources of information. Perhaps if prosecutors were educated as to the content of penitentiary files, the value placed upon this source of information would shift substantially.

In general, Crowns reported few problems accessing various sources of information. Only access to mental health and court documents posed any degree of difficulty. More specific comments on the various obstacles with the sources of information are:

1. For some Crowns, adult criminal records provide insufficient information on the content of prior convictions. These Crowns wanted additional detail in order to establish a persistent pattern and identify prior victims. It was suggested that filing certified copies of transcripts of prior convictions should be sufficient to prove the content of prior convictions. Nineteen
   percent of the prosecutors noted that prior convictions are not always listed on CPIC nor are they up to date.
2. Penitentiary files could be helpful in addressing the concern about the lack of information on prior convictions. For offenders with a penitentiary history, easily accessible and detailed criminal histories are available. Five Crowns (24%) did note however, that the requirement of a subpoena was time consuming and files are often produced on the day of the hearing leaving little time for review. They suggested that prosecutors be allowed to make photocopies for their own records. However, CSC has agreements with all provinces allowing the sharing of information found in penitentiary files with Crowns. Full access to files are provided and photocopies are permitted. Once again, education of Crowns on the legal and practical ability to access CSC files should be considered.
3. Even though, in most instances, Young Offender criminal records are destroyed after the statutory five year period, Crown attorneys did not show much concern about the loss of this information. A few said that some Young Offender records are not up to date and if the
   information was needed in a particular case more complete records can be found in probation offices.
4. A third of the prosecutors encountered difficulties in obtaining victim impact statements either because of difficulties in locating victims or convincing young and traumatized victims to testify. One Crown attorney said that parents' victim impact statements about their children's experiences should be admitted into evidence.
5. The common practice of destroying police reports anywhere between two to five years after a conviction was a major frustration. More than half (57%) of the Crowns said the destruction of records was undesirable and several Crown attorneys suggested legislation requiring the storage of original documents. Once again, correctional files can be used to partially alleviate this problem. However, it should be noted that CSC is also faced with similar difficulties in information retrieval and therefore, some penitentiary files will lack information that is outside of the control of CSC.
6. Psychiatric and psychological reports were found by 48% of the Crown attorneys as difficult to access because of issues of privacy. The difficulties concerned the need for warrants or subpoenas to access the report, if accessible at all.

As with police records, the unavailability of Court decisions and transcripts of prior convictions was a concern for many. 71% of prosecutors said they had faced situations where they were unable to access prior Court decisions or prior transcripts. Court documents, including decisions and transcripts, are sometimes lost or systematically destroyed.

(d) Cooperation

Crown attorneys were asked if at the time of their first DO application they were aware of the information needed to obtain consent from the Provincial Attorney General. Eleven (52%) responded that they were unaware and 17 (86%) reported seeking advice as to the evidence and paperwork needed to proceed with the application. The majority sought advice from other Crowns who had experience with a DO application, followed by Regional Directors, and head offices.

Several Crown attorneys said that a standard precedent and a checklist of items needed to complete the consent request would be useful. Ontario, British Columbia, and Québec have different documents and procedures designed to assist attorneys. In Ontario, the current Crown Policy Manual does not contain any references to DO legislation and procedures but policy guidelines have been drafted and awaiting amendments to violent offender legislation. An informal checklist issued by the Crown Law Office - Criminal (Ontario) was mentioned and used by three Ontario prosecutors.

In B.C., the Crown Counsel Policy Manual contains a short guideline outlining procedures and a list of criteria to help establish when a DO application is warranted. Currently, the manual is being revised to include case law, clearer criteria, and a policy statement to encourage prosecutors to initiate applications. In Québec, the Crown attorneys' policy manual does not contain any references to DO legislation or procedures. The Ministère de la justice is currently developing policy guidelines and formal consent procedures.

Crown attorneys were asked to rate the level of cooperation they received from key actors in DO applications. Their answers were divided into one of three categories: very cooperative, somewhat cooperative, or not cooperative. The categories receive a rating of 1, 2, or 3 respectively. Table 14 presents the average level of perceived cooperation.

Table 14. Average Level of Cooperation from Key Actors

Table 14 shows that Crowns were generally satisfied with the cooperation they received from various individuals. The following suggestions for improvements were offered:

1. Enhanced police resources to facilitate the extensive police investigation needed in a DO application.
2. Sufficient time to prepare the case (caseload demands will be described more fully in the next section).
3. 19% of Crowns requested quicker responses from their head offices.
4. Although Crown attorneys described the cooperation of psychiatrists and psychologists as good, many reiterated their earlier comments about the release of information.
5. Crowns perceived victims and witnesses as quite cooperative, but raised again the issues of victim location and victim trauma.

(e) Caseload while processing a DO application

All Crowns agreed that compared to a normal criminal sentencing, a DO application involves three times or more the amount of work. When asked whether their caseload remained the same when they were involved in a DO application, 67% said yes. The majority of Crowns (71%) would have preferred a caseload reduction or assistance from a junior prosecutor.

(f) Plea bargaining practices

Several Crown attorneys (48%) reported a total of 19 cases where they contemplated proceeding with a DO application, but abandoned the idea. In most instances, they dropped the idea prior to requesting the consent of the Provincial Attorney General. A sampling of the reasons given for not pursuing a DO application were:

• Fragile victim
• Index offence not a SPIO (common assault and threatening)
• Offender would have refused a guilty plea if the Crown proceeded by way of a DO application and instead agreed to 2nd degree murder
• Prior offence very old, problem finding records
• Proving facts of prior convictions difficult
• Children involved, did not wish to prolong proceedings
• Two serious offenses separated by a long period of time

71% of Crown attorneys said they would consider plea bargaining as a viable option if evidentiary problems existed. Assistant Crowns were more likely to consider plea bargaining than Senior Crowns (83% versus 33%).

(g) Miscellaneous technical suggestions

Comments on specific sections of the C.C.C. were as follows:

• If a trial judge is unable to preside at the DO hearing, transcripts are not admissible [R. V. Canning, [1966] 4 C.C.C. 379, 49 C.R. 13 (B.C.C.A.)]. Transcripts should be admissible in such instances to avoid having victims testify again at sentencing.

• Parole eligibility after three years discourages some Crowns from proceeding with a DO application. With the eligibility period starting from the time the offender is brought into custody [section 761(1) of the C.C.C.], offenders may be eligible for parole shortly after being declared a Dangerous Offender.

• Section 755(4) of the C.C.C. should not restrict the number of experts allowed to testify.

• Legislation should be introduced to disallow severance of cases when a DO application is initiated. Or alternatively, prosecutors should be allowed to adjourn the hearing until all outstanding charges have been addressed.

• Sub-paragraph 672.21(3), which allows certain protected statements to be introduced as evidence, should be amended to make these statements also admissible in DO applications. It should be noted that these statements will be admissible in the yet-to-be proclaimed dangerous mentally disordered accused legislation.

D. Views on Proposed Reforms

The present DO legislation was perceived very positively by Crown attorneys. All of the attorneys agreed that the provisions are useful, and 95% stated that without DO legislation normal sentencing would not be capable of dealing with high risk violent offenders.

Crown attorneys were asked to rate the desirability and the feasibility of current models being examined by the federal Department of Justice. Table 15 lists the four different models that were presented to the Crown attorneys.

Table 15. Four Models for Dealing with High Risk Violent Offenders

Table 16 shows the average level of desirability and feasibility expressed in respect of each model. Desirability was rated as follows: desirable, somewhat desirable, or undesirable (rated 1, 2, or 3 respectively). Feasibility was rated as follows: feasible, may be feasible, or not feasible (rated 1, 2, or 3 respectively).

Table 16. Average Ratings of the Four Models

Overall, the average desirability ratings show the most support for the contingent sentencing model. The results also reflect the findings from the Department of Justice's more general consultations with provincial Ministries of the Attorney General. Interestingly, the contingent sentencing model is the closest to the present DO legislation. The least desirable model was the "back end" DO application model. Most attorneys preferred that sentencing be done at the "front end", immediately after a conviction. Many commented on Model 2, the one year opportunity model. The following comments may also be applicable to Model 1:

• In guilty plea situations, it would be unfair for the Crown to reopen the case.

• Crowns may be tempted to include in plea negotiations the offer that a DO application will not be sought within the critical year. This would then defeat the purpose of such legislative initiative.

• Sentencing provides a sense of finality and this model would minimize this idea.

• Some asked what could be so enlightening in the first year after sentencing.

• Trial and sentencing go together, and where delays occur between the two, the judge may forget the details.

• Under such a model, offenders may refuse assessment or treatment for fear of revealing information which could be used in the DO application.

Most of the comments regarding Model 3 were aimed at the desirability of what is perceived mainly as a "back end" option. Crown attorneys from B.C. showed higher disapproval ratings of this model than prosecutors from Ontario (2.22 versus 1.36). This model is based on Washington state's Sexual Predator legislation which has recently been declared unconstitutional by the U.S. Supreme Court. Legal challenges to a similar law in Canada could be anticipated.

E. Knowledge of Predictors

From the hundreds of cases that are presented to Crown attorneys for prosecution in any given year, decisions must be made as to which cases require special prosecution. Many times the decisions are guided by judgements as to the seriousness of the offence and the likelihood of further serious harm. In few cases, are these judgements as important as in a DO application. And in all cases, the judgement to prosecute an offender by way of a DO application depends upon some knowledge of the predictors of violent re-offending.

Crown attorneys were asked how they would rate a list of predictors of violent recidivism (see Appendix D, Q30). They rated 37 possible predictors selected from a review of the research literature on a 10 point scale, 10 being most important and 1 being not important. Table 17 shows the average rating for each predictor.

Table 17. Ranking of the Predictors of Recidivism

In general, the Crown attorneys' ratings of the various predictors were congruent with the research literature on the prediction of criminal behaviour. Factors such as criminal history (rated #4), psychopathy and antisocial personality disorder (#1, #2) and antisocial attitudes (#12) are some of the best predictors of general recidivism (the reason for underscoring "general" will soon become apparent). Personal distress factors (#19, #30), class (#35) and race (#37) are poor predictors. One factor that the Crowns failed to give sufficient emphasis is social support for crime (#25). This factor is probably one of the best predictors of criminal conduct.

Although the ranking of predictors is reflective of the research literature on general recidivism, the majority of DOs are sex offenders and the intent of the legislation is to prevent further violent behaviour. The predictors of general recidivism may not necessarily be the same as the predictors of sexual or violent recidivism. Table 18 compares the ten top ranked predictors listed by the Crown attorneys with the ten best predictors of sexual recidivism identified by a meta-analytic review of the sex offender literature (Hanson & Bussière, 1995).

Table 18. The 10 Highest Ranked Predictors: Crown Ranking and Sex Offender Research

To simplify the table, diagnoses of Clinical Psychopathy and Antisocial Personality Disorder were combined. The list identified by the research should be approached with caution as some of the predictors are based on relatively few studies. For the predictors nominated by Crown attorneys, Impulsivity and Use of Weapon, there were insufficient research studies. Nevertheless, examination of the two rankings can be instructive.

Three factors deemed highly predictive by the Crown showed weak empirical relationships to sexual recidivism. From the Hanson and Bussière (1995) study, lack of remorse, assault/violent history and brutality (measured by injury to victim) were relatively insignificant predictors of sexual recidivism. However, a history of assault/violent and brutality meet the legislative criteria for a DO application.

It is also noteworthy which variables the Crown attorneys did not consider important when the research literature suggests otherwise. These variables are a diagnosis of a severe psychiatric disorder, a negative relationship with the mother, prior criminal history and age. As noted in Part 1 of this report, a diagnosis of a severe mental disorder may work against a DO application and it is possible that these offenders are found unfit to stand trial or not criminally responsible on account of a mental disorder.

The single best, research nominated predictor of sexual recidivism is a phallometrically assessed preference for children. It is important to differentiate those child molesters who display a sexual preference for children and other child molesters. Many child molesters do not show deviant

sexual preferences even though they have committed deviant sexual acts against children. It is the subgroup of child molesters who sexually prefer children who pose the higher risk for sexually re-offending and phallometric assessments provide the best means for identifying this subgroup of offenders.

It would have been desirable to compare the rankings given by the Crown attorneys with an empirical list of the predictors of violent recidivism, but such a summary currently does not exist. Webster et al. (1994) provide a tentative list of the predictors of violent re-offending but we are hesitant to draw from their list given that it is based upon a sample of mentally disordered offenders. However, a few remarks may be in order.

From the Webster et al. (1994) list, the degree of victim injury was found to be negatively related to violent recidivism. That is, brutality may actually be associated with a lower likelihood of re-offending violently. In our own findings, physical injury of the victim was not as frequent among the DOs as we expected. These results appear counter-intuitive. Why would the degree of victim injury not be related to violent recidivism? A possible explanation is that severe physical injury often results from highly charged emotional interactions. These interactions are likely in situations where the offender knows the victim and often result in "crimes of passion". Anecdotal evidence suggests that Dangerous Offenders may not fit this pattern in that the victim is more likely to be a stranger. Although we know of no research that directly addresses this issue, there is the possibility that Crowns and the Courts may deal more harshly, by way of a DO application, with offenders who prey on strangers.

Returning to the Webster et al. (1994) predictors of violent recidivism, The best predictor was a diagnosis of psychopathy/APD. This is consistent with the Crowns' rankings and the predictors of sexual recidivism. The next best predictor was early school maladjustment, a factor that does not appear on the list of predictors of sexual recidivism and which was ranked #28 by the Crowns (Table 17).

F. Summary

One of our major findings from the survey of Crown prosecutors was that most Crowns are generally satisfied with the existing DO provisions. Certainly many suggestions for improving this part of the C.C.C. were given, but on the whole the consensus was that there was little need for substantially new legislation.

Crown attorneys face a variety of obstacles when initiating and processing DO applications and this suggests a need to fine-tune the DO process. It should be noted that many of these changes may also improve the normal sentencing process. All the individual issues raised by the

prosecutors warrant careful consideration given the potential consequences in failing to successfully process a DO application. Some may not survive Charter challenges, others may be too expensive to implement, but in general, most can be addressed with relatively minor changes to the C.C.C., the Canada Evidence Act, R.S.C. 1985, c. C-5, and Provincial procedural practices.

The Crown attorneys were not generally supportive of the various options that have been proposed as alternatives to the current DO provisions. "Front end" sentencing alternatives and simple streamlining of current sentencing practices are more likely to be favoured by Crown attorneys.

The information provided in this report can be used to remove some obstacles faced by Crown attorneys who initiate and process DO applications. One important implication from the findings is for the development of a flagging system to identify offenders for a DO application. Tables 8, 9, and 18 could be helpful in developing a list of criteria needed to identify possible DOs. Since all three provinces surveyed are currently developing policy guidelines for their own Crown policy manuals, it could be useful to share their guidelines with a view to promoting consistency across the various jurisdictions in Canada.

General Summary and Conclusions

The Crown Files Research Project reviewed the files of 64 DOs, representing nearly 44% of all DOs in Canada, and another 34 high risk violent offenders. The project also conducted systematic interviews with 21 Crown attorneys who had experience with DO applications. Together the results give a description of a high risk violent population of offenders and offer suggestions to improve the application of the Dangerous Offender provisions.

In general, Crown prosecutors from Ontario and B.C. appear to target violent high risk offenders. However, the Dangerous Offenders sampled in the present study were predominately sex offenders. Nonsexual, violent offenders were infrequently dealt with by way of a DO application. Comparing the present legislation with the Dangerous Sexual Offender (1960) laws, the impact of the current legislation on identifying nonsexual violent offenders appears minimal. We say this with hesitation because it is possible that DOs from the other provinces may not share similar profiles to the DOs found in this project. For example, Aboriginal offenders who are declared DOs may have characteristics different from the non-Aboriginal DO population.

The majority of the Crowns stated that the decision to proceed by way of DO application is made shortly after arrest. For the most part, there appears to be sufficient information available to

proceed. Yet, the prosecutors also expressed frustration over the unavailability of information from the police and the courts suggesting that a number of cases may be missed in the screening process.

Missing potential DO applicants has serious consequences. The present findings, along with the results from recent research on the prediction of sexual reoffending (e.g., Hanson & Bussière, 1995), provide an opportunity to develop and apply these risk factors in the identification of offenders for DO application. A pilot project with Crown offices for the screening of high risk

violent offenders may be beneficial in this regard. Crowns can be provided assistance in gathering the appropriate information on some of the important risk factors (e.g., Antisocial Personality Disorder, victim preference, etc.) needed to make a decision on a DO application. Educating Crowns on the information available in penitentiary files and providing protocols for easier access to these files would form key elements of such a project.

Finally, the majority of the Crowns were satisfied with Part XXIV of the C.C.C. A fundamentally new model for dealing with high risk violent offenders was not deemed necessary. Rather, some clarification in language and procedures would be sufficient along with guidelines for using the DO provisions of the Criminal Code.

References

Appendix A

Sections of the Coding Manual

1. Demographics
2. Important Dates
3. Index Offence
4. Alcohol/Drug Problems
5. Education/Employment
6. Emotional/Personal
7. Institutional
8. Participation/Performance
9. Financial
10. Family/Marital
11. Accommodation/Leisure/Recreation
12. Companions/Friends
13. Attitudes/Orientation
14. Early Childhood
15. Conduct Disorder
16. Anti-Social Personality Disorder
17. PCL-R Items (as coded from file information)
18. PCL-R Items (as copied from offender file)
19. SIR Scale (as copied from offender file)
20. Sex Offenses
21. Media Coverage
22. Criminal History

Appendix B

Codes for the Type of Report

1) Psychology Report

2) Psychiatrist Report

3) Social Worker Report

4) Case Manager Report

5) Intake Assessment

6) Police Report

7) Criminal Profile Report

8) Correctional Plan

9) Final Treatment Report

10) Penitentiary Placement Report

11) Penitentiary File (other)

12) Force Field Analysis

22) Crown Brief

23) Lawyer Notes

24) Consulting Psychiatrist DEFENSE

25) Consulting Psychiatrist CROWN

26) Pre-sentence Report

27) Judge's Decision

28) Crown Memorandum

29) Court Transcript

32) External Psychology Report

33) External Psychiatrist Report

34) Children's Aid Society Report

35) Social Work Report (Childhood)

36) Parole Board Report

41) Discipline and Dissociation File

42) Visits and Correspondence File

43) Sentence Administration File

44) Case Management File

Appendix C

The Exploratory Variables

1. Degree of brutality evident and whether weapons were used in the index offence
3. Whether the perpetrator pleaded guilty to his index offence
5. Treatment attendance and outcome
7. Children's Aid Society and reform/training school histories
9. Fetishes and specifics of preferences for sexual violence
11. Indication of sexual offenses for which the perpetrator had not been charged
13. Total number of adult male and female victims (lifetime)
    Total number of child male and female victims (lifetime)
15. Number of juvenile offenses and convictions
17. Reasons given by the Court for the Dangerous Offender finding
19. Amount of print media coverage as measured in square centimetres
21. Number of aliases
23. Physical deformities
25. Victim intoxication at the time of the index offence
27. Social links to gangs/organized crime
29. Reports of childhood physical abuse and neglect

Appendix D

Crown Attorney Interview Form

A. Basic Information:

First/Last

Year/month/day of coding

Subject Number

Provinces

Gender

Position

Seniority--Year called to Bar

Total number of D.O. applications involved

Current number of D.O. applications

D.O. success rate

D.O. contemplated but dropped

Date of last D.O. application

B. General:

Q1: What were the characteristics of the offence that led you to initiate a D.O. application?

Q2: What were the characteristics of the accused that prompted you to initiate a D.O. application?

Q3: At what point in time did you consider making the application?

Q4: Did the media coverage play a role in your decision to make a D.O. application?

Q5: Do you think the media coverage plays a role in the decision of other Crown attorneys to make D.O. applications?

C. Legislation:

Q6: In your own words, what constitutes a "Serious Personal Injury Offense" (SPIO)?

Q7: Are you satisfied with the statutory definition of SPIO?

Q8: Do you find the 10 years minimum requirement adequate, or would you like to see it changed?

Q9: Should "Severe psychological damage" be included in the definition of SPIO?

Q10: Do you have any problems with the following statutory criteria?

Q11: In your opinion, is the mandatory requirement for two psychiatrists to testified necessary for establishing whether the offender is likely to violently recidivate (dangerousness)? 

For establishing whether the offender is treatable?

Q12: Do you think the consent requirement is necessary?

Q13: In general, do you think that Crown attorneys are knowledgable about the legislative criteria?

Q14: In general, do you think that Crown attorneys are capable of identifying the cases where a D.O. application is warranted?

Q15: In your opinion, what changes are needed to improve the legislative criteria?

D. Procedure:

Q16: In your opinion, is there enough information available for prosecutors to allow them to make accurate identification of cases where a D.O. application is warranted?

Q17: If you were contemplating a D.O. application, what information would be more important to you?

Rank order in order of importance:

1. Adult criminal record
2. Young offender criminal record
3. Police reports
4. Victim impact statements
5. Court decisions/transcripts of priors
6. Psychiatric reports
7. Psychological reports
8. Penitentiary files

Q18: How readily available is the above-mentioned information?

Easily available =1

Difficult but available =2

Virtually or Impossible to access =3

Q19: At the time you initiated your first application, were you aware of the information needed to obtain consent for a D.O. application?

Q20: Did you seek advice as to the required evidence and paperwork needed to proceed with an application?

Q21: What level of cooperation/support did the following persons provided you?

Very cooperative =1

Somewhat cooperative =2

Not cooperative =3

If 2 or 3: How could cooperation be enhanced?

Q22: Compared to a normal sentencing hearing, how much more work is involved for a D.O. application?

Q23: Did your caseload remain the same when you were involved in a D.O. application?

If yes, explain what arrangements were made?

If no, what arrangement should have been made?

Q24: Did you ever encounter evidential problems (i.e. in sexual offenses cases) which prevented you from initiating a D.O. application?

Q25: If evidential problems are present, do you consider plea bargaining as a viable option?

Q26: If the accused seems to meets the D.O. criteria, but is willing to enter a guilty plea and agree to a 15 years penitentiary term, would you or would other Crown attorneys consider jeopardizing the plea bargain and proceed with a D.O. application?

E. Proposals for Reforms:

Q27: Do you think the D.O. provisions are useful?

Q28: Do you think normal sentencing is capable of dealing with high risk offender adequately, without the necessity of relying on D.O. legislation?

Q29: Would you consider the following options (1) desirable, and (2) feasible?

Desirable =1

Somewhat desirable =2

Undesirable =3

Feasible =1

Maybe feasible =2

Not feasible =3

• If there are reasonable grounds to believe that an offender is likely to recidivate at the end of his/her sentence, a D.O. application could be initiated in the last year of the offender's sentence.

• Same as above, but within one year of sentencing.

• Prosecutors could apply to a Court for a finding that probable cause exists to believe that a person is a sexually violent predator (person who suffers from a mental or personality disorder). Upon such finding, the person would be civilly committed.

• Offenders found guilty of serious personal injury offenses where there is evidence of mental abnormality, sexual behavioral disorder or sexual motivation in the offence would be sentenced to an indeterminate sentence subject to a review at a point in time determined by the Court.

F. Knowledge of Predictors:

Q30: On a 10 point scale, 10 being most important and 1 being not important, how would you rate the following factors as predictors of violent recidivism?