Correctional Service Canada
www.csc-scc.gc.ca
Common menu bar links
Research Reports
This Web page has been archived on the Web.
Inmates Referred for Detention (1989-90 to 1993-94) A Comparative Analysis
Brian A. Grant
Research Division
Correctional Research and Development
Correctional Service of Canada
July, 1996
Table of Contents
Chapter 1: Introduction
Chapter 2: Method
Chapter 3: Changes Over Time
Chapter 4: Previous Criminal Record
Chapter 5: Recidivism
Chapter 6: Analyse
Executive Summary
The purpose of the report is to provide detailed analyses of the detention provisions of the Corrections and Conditional Release Act (CCRA) in preparation for a legislated parliamentary review of the Act. The study addresses three main questions:
1. Are the right people referred for detention? That is, are the highest risk violent offenders referred for detention? It is not clear that the highest risk offenders are being referred for detention. However, the offenders who are referred have committed violent offences. The recidivism data indicate that those referred are no more likely to be readmitted, or to commit a new offence, than many of those released without being referred for detention.
2. Are the right people being detained? That is, are the people detained the highest risk offenders? The results suggest that offenders who are detained are not higher risk than offenders referred and not detained. For example, while detained offenders are more likely to have had a conviction for a sexual offence, they are less likely to have had convictions for robbery and drug offences. In addition, detained offenders are more likely to have been given shorter sentences by the courts and to have fewer previous convictions for which they served their sentence in a federal prison. While some detained offenders have a high risk of re-offending, others are less likely to reoffend than those who are released at their SR date.
3. Are those who have been detained the most likely to reoffend after release and are they the most likely to reoffend violently? The data suggest that offenders who are released at their statutory release date, including those who are referred but not detained, are more likely to be readmitted than those who are detained to the end of their sentence. It also appears that those who are not detained are more likely to reoffend within two years of their release, and are equally likely to reoffend with a violent offence (another scheduled offence) as those who are detained.
The results also indicate that for both those referred (excluding those detained to the end of their sentence) and those released at their statutory release date the readmission rate is very high, approximately 50%, within two years of release. Readmission includes technical violations of release conditions.
Presented in the report are descriptions of the changes in the number and types of referred and detained offenders, offence histories and recidivism rates. The study also looks at gender, regional and racial differences in referral and detention patterns.
The main findings of the study are summarized below.
1. The number of referrals, as a percentage of all releases, has been increasing over the five year period of the study and the number of referrals, as a percentage of all potential statutory releases, has been increasing even more dramatically. While those referred are more likely to have longer sentences, 6 years or more, than other inmates released, those detained to the end of their sentence are less likely to have longer sentences.
2. Offenders detained to the end of their sentence tend to be older than those released after serving two thirds of their sentence (statutory release), and those referred and subsequently released. Female offenders are less likely to be referred for detention, but once referred are equally likely, compared to male inmates, to be detained until the end of their sentence.
3. Aboriginal offenders are over-represented in the referral group, but given that they are referred, they appear in the same distribution as Non-Aboriginals across the four sub-groups used to study referral cases (referred and released, detained and then released, detained to the end of their sentence, and other).
4. There are large regional differences in the percentage of cases referred. Quebec has the lowest referral rate, while other regions refer more than would be expected given their proportion of the inmate population. Year-to-year increases in the percentage of releases referred are consistent across all regions.
5. Offenders referred for detention have a greater number of scheduled offences than other offenders, and this number has been constant over the five years of the study. This would suggest that the increase in the number of referrals is not the result of the number of convictions offenders have for scheduled offences. Drug offences are a relatively small percentage of the total number of scheduled offences and have not been increasing over the five years of the study. The observed increase in the number of referred cases cannot be attributed to the addition of Schedule 2 to the CCRA, nor can it be attributed to a change in the number of drug offences.
6. Referral cases are more likely to have been convicted of scheduled sexual and assault offences as well as non-scheduled violent offences than other offenders. They are less likely to have been convicted of robbery offences. Amongst those referred, offenders who were detained were more likely to have been convicted of a sexual offence than those who were released. This suggests that detention referrals are targeting sex offenders. Based on the number of previous criminal offences, offenders referred for detention have a lower level of criminal activity than those released at their statutory release date, and have a level of criminal activity more similar to those released on full parole.
7. There appears to be an increase in the variety of violent offences for both referred offenders and those released on full parole and statutory release over the five year period of the study. This could account for some of the increase in referrals, although the result is inconsistent with respect to other findings.
8. Aboriginal offenders, who are referred, are less likely than Non-aboriginal offenders to have been convicted of sexual offences and drug offences and are more likely to have been convicted of assault and manslaughter.
9. The pattern of readmissions is similar for the statutory release and the referred groups who were subsequently released before the end of their sentence. The group detained to the end of their sentence had the lowest readmission rate, but this may reflect the need for court action prior to readmission, rather than a difference between this group and the others. Time to a new offence is shortest for those released at their statutory release date and those referred and released. Those who were detained had re-offence rates lower than those released on full parole. On average, the incapacitation effect of detention lasts only about one year, the period the offender remains in prison.
10. The readmission and re-offence rate was lowest for those over 35 years of age. Given that the average age of the referral group is 32, and the average age of those detained to the end of their sentence is 33, it is likely that many of these offenders have a lower risk of re-offending than would be expected given their offences.
11. Only 16% of those detained to the end of their sentence, and 12% of those detained and then released, had a new offence of any type within two years of release. Only 11% committed a scheduled offence as their first new offence, while 15% did commit a new scheduled offence sometime after their release (given a minimum of two years following release).
12. Aboriginal offenders have higher readmission rates than Non-aboriginal offenders, but these differences are reduced when new offence is used as the recidivism measure. These results suggest that Aboriginal offenders may be more likely to have their conditional release revoked, even though they are not more likely to commit a new offence. Additional investigations may be needed to explain the observed differences.
13. Regional differences are evident for both readmission rates and new offences within two years of release. Patterns of readmissions across referral groups and regions are consistent with the general patterns observed. Those detained have lower re-admission and re-offence rates. Striking is the relatively higher re-offence rate in the Atlantic region for full parole releases and those detained to the end of their sentence.
Overall the results suggest that it is very difficult to discriminate within the high risk group of offenders who are held until their statutory release date. Revised methods of selecting inmates for detention based on objective assessment of risk may be more effective at selecting the most serious offenders, rather than relying on a set offences as specified in Schedules 1 and 2 of the CCRA.
By taking account of incarceration costs it is estimated in the report that over one million dollars a year could be saved if most referral cases were released to the community rather than being detained. It is argued that the high readmission rate for these types of offenders would ensure that the safety of the community is protected, while allowing those who can remain crime free to do so.
Finally, additional research would assist in the development of better selection criteria for those most likely to commit new offenses. For example, using objective risk measures would be more effective than the current list of offenses specified in the Schedules of the CCRA. Unfortunately, the mandate for this project does not include the development of improved selection tools.
Acknowledgments
A research report takes considerable time and effort from people with a variety of different skills. Key to producing accurate report data, and to the presentation of the data, are staff who provide data development and data analytic support. For this report, Ray Belcourt provided the initial data bases from which the information on detention and other cases were obtained. His efforts creating workable data bases from the Offender Management System (OMS) and in helping us to ensure the integrity of those data bases was the first step in making the study possible. Pierre Couturier from the National Parole Board (NPB) provided the information on NPB decisions for the referral cases. Linda Lefebvre developed the software to extract pre-release and post-release data from the different data bases, and spent countless hours testing the data programs to ensure we had accurate information. Mike Muirhead provided data analytical support. He had to deal with the many permutations and combinations of variables which were used in achieving the analyses presented in the report. Many preliminary and associated tables have been produced to make possible a single summary number presented in the report. Mike was responsible for doing this work. His efforts provided me with the time I needed to write the report.
Chapter 1:
Introduction
One of the principles stated in The Corrections and Conditional Release Act (Bill-C36) is that the Service use the least restrictive measures consistent with the protection of the public...' (Sec 4(b)). To adhere to this principle, efforts are made to provide programming to offenders so they may be released safely from a prison setting to community supervision at the earliest possible date in their sentence. Early release is usually achieved through the use of parole for lower risk offenders, but higher risk may not qualify for parole. Offenders who do not qualify for parole are normally released to community supervision at their statutory release date which is reached after serving two-thirds of their sentence in custody. However, for a small group of offenders extraordinary measures are needed to ensure the protection of society, and release may be delayed beyond the statutory release date.
If it is believed that an offender is likely to commit a violent offence, or a serious drug offence before the expiry of the court imposed sentence then the Correctional Service of Canada (CSC) may apply to the National Parole Board (NPB) for permission to detain the offender, that is, to not release the offender on the statutory release date. Subsequent reviews may be held to determine if the offender should be released prior to the expiration of the sentence. However, CSC must release the offender from custody at the end of the court sentence, even if there is a concern that the offender may commit another serious crime.
The Corrections and Conditional Release Act specifies the conditions under which an inmate can be detained (held in prison beyond the statutory release date). In summary, these require that the offender has committed one of the offences listed in Schedule 1 or Schedule 2 of the Act. Schedule 1 provides a list of violent offences and Schedule 2 provides a list of drug offences. These Schedules are presented in Appendix A of the report. The offences listed in the Schedules to the Act are generally referred to as scheduled offences, which simply means they are listed in the Schedules attached to the Act. If the offender has committed a scheduled offence, and if there is an expectation that the offender will commit a violent offence or a serious drug offence, prior to the expiration of the sentence, then the offender may be detained by order of the National Parole Board. In exceptional circumstances, offenders, who have committed offences that are not listed in the Schedules attached to the Act, may also be detained.
It is recognized that detaining someone beyond their statutory release date is a serious restriction on the legislated right to be released to community supervision after having served two-thirds of the sentence. However, the restriction may be justified because of concerns, in these special cases, that a serious violent or drug offence will occur, and the need to ensure the protection of society.
Given the seriousness of the restriction it should be possible to demonstrate that detention achieves its goal of protecting society from further offences. Detention does provide additional protection for society during the period of detention, but the offender must be released eventually. The question then is, does the period of detention reduce or increase the likelihood of another offence, and if another offence occurs, does it occur more quickly following release at the end of the sentence when there is no supervision available, as opposed to during the period of supervision which follows a conditional release at the statutory release date.
The analyses in the report are designed to:
1. Provide an overview of the number of detained cases.
The number of offenders detained will be compared across
- other forms of release (full parole and statutory release)
- outcome of the detention hearing
- over the five years from 1989-90 to 1993-94
- across the five administrative regions of CSC
- between Aboriginal and Non-aboriginal offenders.
1. Compare the previous criminal history of referred and detained offenders.
2. Compare the recidivism rates of detained and other groups of offenders to determine if detention had an effect on the recidivism rate and the time to recidivate.
The number of females who were detained over the five year period was very small. While the results from female offenders are included in the data presented, no special analyses of female offenders could be performed.
The analyses should provide answers to the following questions:
1. Are the right people referred for detention? That is, are the highest risk violent offenders referred for detention?
2. Are the right people being detained? That is, are the people detained the highest risk offenders?
3. Are those who have been detained the most likely to reoffend after release and are they the most likely to reoffend violently?
Outline of the report
The report is divided into five sections; methodology, an overview of offenders who are referred and detained relative to those who are released, a description of the previous criminal history of all offenders, presentation of recidivism results, and discussion. For readers who are less concerned about the technical aspects of how the data were obtained and the implications of the methodology, it may be possible to skip the methodology section and move directly to the results of the analyses presented in Chapters 3, 4 and 5. The overview presented in Chapter 3 provides a description of the number of referral releases relative to other releases, sentence length, age, racial and regional distribution of offenders released on full parole, statutory release, and those who were referred for detention. Chapter 4 presents information on the previous criminal history of offenders. Recidivism results, presented in Chapter 5, include information on both readmissions to federal institutions and re-offence rates. The discussion brings the results together in summary form and includes a discussion of the implications of results for the three questions posed above.
Chapter 2
Method
Data for the study were created using three data bases (admissions, releases and offences) obtained from the Offender Management System (OMS) maintained by the Correctional Service of Canada (CSC). In addition, the National Parole Board provided a data set containing decisions for all offenders referred for detention from 1989-90 to 1993-94. From these data bases a combined data base was constructed.
The final data base contains information on all offenders released during the period 1989-90 to 1993-94. It includes information on their previous federal criminal history, the date and type of release within the study period, readmissions after release and offences committed subsequent to release from prison.
It was intended that changes in referral patterns and detention outcomes would be tracked over time. Therefore, release groups or cohorts were created for each fiscal year. The first release within the study period (1989-90 to 1993-94) places the offender in a cohort for the year of that release and they remain in that cohort. If an offender was released in 1989 he/she would be in the 1989-90 cohort and subsequent admissions and releases would not change their cohort. Offenders shown as being released in 1993 are those who had their first release during the study period in 1993.
One of the effects of this is that the number of releases per year, shown in the report, is less than that presented in the general release statistics presented by CSC because for the purposes of the study, an offender is only counted as a release once in the five year period. However, many offenders are released more than once. These subsequent releases are included in the follow-up data. Another effect of this is an apparent decline in the number of statutory releases over the years since these cases are most likely to have multiple releases over the study period, but only their first release is counted. However, because cases are being tracked over time, and it would be inappropriate to double count the same offender, only one release can be counted.
Offenders are referred to the National Parole Board for detention by the Correctional Service of Canada. Those referred are called referral cases and once referred, the National Parole Board must make a decision to release or detain the offender. Even if a referral case is not detained, they remain in the referral group for the purposes of the study. It should be noted that those referred represent a small percentage, generally less than 5%, of those released in any year.
Since referred offenders do not have release dates similar to other offenders, some definitions are needed to allow comparisons. Offenders who are detained do not have a release date until they reach the end of their sentence or until the National Parole Board decides that the risk to society is low enough to allow them to be released. Therefore, for offenders who were referred, an index date was created. This date corresponds to the date on which the National Parole Board decided they should either be detained or released. This date will be less than six months before the statutory release (SR) date, and is normally two to four months before the SR date. In some cases the date will be closer to the SR date, but these are exceptions. The index date allows for a comparison between those released and those detained to determine the incapacitation effect of detention. For some analyses with referral cases the true release date is used; this is the date the offender was released from prison, either following an order from the NPB, or at the end of their sentence. Which time period is being used, index date or release date is specified in the analyses. For offenders released on full parole or statutory release (without having been referred), the index date and the release date are the same.
To reduce the complexity of the data development, a number of compromises were made to create the final data set. The offences committed prior to release or referral are all offences committed in the offenders' federal criminal career. Therefore, older offenders will likely have more offences than younger offenders. In addition, in analyses presenting offence information prior to release, the offences included are not just those for the current admission or current sentence, although this may be the case for many of the offenders.
The data base includes information on the first re-admission to a federal prison after release. Re-admission may be for a new offence, but for most offenders, readmission is for a technical violation related to their conditional release. However, for those detained to the end of their sentence, there is no conditional release and therefore no technical violations. Therefore, readmission rates will be lower for those detained to the end of their sentence than for those released on some form of conditional release, either full parole or statutory release.
In addition, the type and number of offences for which the offender was sentenced at the first sentencing date after release from prison and subsequent sentencing are also recorded. These data will be used to measure time to recidivism and offence types. Ideally, we would measure time to a new offence using the date the offence was committed . Unfortunately, this information is not routinely collected in the electronic data systems being used. Therefore a substitute measure was needed. The only consistently available date with offences is the sentence date and this was used. The sentence date may be anywhere from one to eighteen months after the offence date and is dependent on the seriousness of the crime and the ability of the court system to process cases.
The follow-up data do not include information on offences which did not result in the person serving federal prison time. These offences are not included in the current Correctional Service of Canada data bases, and must be obtained from outside agencies. In most cases, these will be offences which resulted in fine, probation, or jail sentences less than two years and offenders with these types of sentences fall under provincial jurisdiction.
Analysis groups
Four main groups were defined for the analyses as described below; the first three are based on the first type of release during the five year period of the study. Any referral in the five year period places the offender in the referral group. Referral cases are always counted in the year in which there is a referral, and previous releases are ignored. The referral cases are sub-divided into four groups based on the outcome of the detention process and their subsequent release history. The eight groups are defined below along with the short form used to refer to them in tables and charts:
- Full parole (FP)
- Statutory release following another release, either full parole or day parole (SR & other)
- Statutory release only (SR only)
- Referred for detention (Referral). Includes anyone who had a detention hearing by the NPB.
- Referred for detention, but not detained (Referred, but released). This includes offenders granted one chance statutory release and offenders granted statutory release with a residency requirement, if the residency requirement was lifted before the end of the sentence.
- Detained to the end of sentence, that is, detained until warrant expiry date (WED) ( Detained to WED). This group includes those granted statutory release with a residency requirement, if the residency requirement remained in effect to the end of the sentence. Only one such case existed in the data.
- Detained, then released before WED (Detained then released)
- Other (Other). Consists mostly of those who were still detained during the entire follow-up period, and a very small number of referral cases who were released and then subsequently detained. This latter group was too small to make analyses possible.
When reviewing the data tables the reader will frequently find that the total number of cases varies from table to table. This occurs because attempts are made to keep as many cases in each analysis as possible and some cases have missing information for different factors or variables. Therefore, depending on which factor or variable is being investigated the number of cases will vary. To reduce the confusion, readers should consider using the percentages which are usually presented with the raw numbers. The raw numbers are useful in determining if percentages relate to a large or small number of cases (10% of 10 cases is only one case), but are not directly useful for making comparisons across groups.
Samples
The total number of cases in each analysis group, across the five year period, is presented in Table 1. Also shown in the table is the percentage of cases that could be followed for a two year period after release. Cases that could not be followed for two years are those who were released after March 31, 1993. For cases that were detained, the percentage that could be followed for two years was lower than those who were released on SR or full parole. For example, while approximately 80% of the cases who were released without detention were followed for two years and 70% of those who were detained and released had a two year follow-up, only 44% of those who were detained to the end of their sentence had a two year follow-up.
Table 2-1: Number of cases in each group and percentage of cases with two year follow-up.
| Sample | Type of releases | Distribution of referral cases | ||||||
|---|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referral | Referred but released | Detained then released | Detained to WED | Other1 | |
| Total | 10,624 | 6,207 | 5,654 | 1,115 | 198 | 131 | 533 | 253 |
| With two year follow-up | 8,322 | 5,284 | 4,782 | 491 | 164 | 92 | 235 | 2 |
| Percentage with two year follow-up | 78.33 | 85.13 | 84.58 | 44.03 | 82.83 | 70.23 | 44.09 | 0.79 |
1Other includes those cases still detained at the conclusion of the study period, and a few cases who were released and subsequently detained.
Analysis
All analyses were conducted using the Statistical Analysis System (SAS, Version 6.10) and Microsoft Excel (Version 5.0). In general, statistical tests are not performed on the data because the study captures all cases in the Correctional Service. As such, the results are presented for the entire population and there is no need to determine if parameter estimates are reliable, as is the function of statistical analyses. Where a difference is observed it is a true difference, not influenced by sampling error, and readers may determine if it is a meaningful difference from their perspective.
Chapter 3:
Changes over time
The number of offenders referred for detention, as a percentage of all releases, has been increasing steadily from 1989-90 to 1993-94. Results presented in Table 3-1 and Figure 3-1 present the number and percentage of releases by year, for the past five years. As can be seen in Figure 3-1 the percentage of releases that are referred for detention has been increasing while the percentage of cases released on statutory release (SR) has been declining. Referrals have increased from 3% to 6% of all releases.
Table 3-1: Type of release by year
| Type of release | 1989-90 | 1990-91 | 1991-92 | 1992-93 | 1993-94 |
|---|---|---|---|---|---|
| Full Parole N | 1,920 | 2,037 | 2,139 | 2,384 | 2,405 |
| % | 36.3 | 41.8 | 45.4 | 49.7 | 50.6 |
| SR-Other release N | 1,619 | 1,357 | 1,233 | 1,168 | 990 |
| % | 30.6 | 27.9 | 26.2 | 24.4 | 20.8 |
| SR-first release N | 1,608 | 1,254 | 1,108 | 1,025 | 1,062 |
| % | 30.4 | 25.8 | 23.5 | 21.4 | 22.4 |
| Referral N | 149 | 221 | 232 | 219 | 294 |
| % | 2.8 | 4.5 | 4.9 | 4.6 | 6.2 |
| Total | 5,296 | 4,869 | 4,712 | 4,796 | 4,751 |
The results in Table 3-1 and Figure 3-1 also indicate that full parole releases have increased while release on SR has decreased. Part of the reason for this is that SR cases are more likely to be multiple term offenders, but are only counted once in these data.
Figure 3-1: Type of release by year
Referral for detention is reserved only for cases with the highest risk of reoffending violently. As a result it may be more appropriate to compare changes in the number of referral cases with changes in the number of SR cases, rather than comparing the number of referrals with all types of releases including parole. SR cases are also high risk cases since many of them have either failed on a previous release, or been considered too high a risk for release. Figure 3-2 presents the percentage of potential SR releases (all SR releases plus referrals) that were referred for detention. The data indicate that there is a steady increase in the percentage over the five years starting with 4% in 1989-90 and increasing to 12% in 1993-94.
Figure 3-2: Percentage of potential SR releases referred for detention
The increase in the number of referrals could be the result of changes in the offence histories of offenders during this period and the result of the increased number of offences for which detention was an option. In 1992, the CCRA added a small number of offences to Schedule 1 and created Schedule 2 (drug offences) making detention possible for a larger number of offences. However, analyses presented later in the report indicate that the average number of scheduled offences has not changed for those referred for detention over the five year period of the study.
Length of sentence
Those referred for detention are more likely to have longer sentences than those released on FP or SR as shown in Table 3-2. The majority, 63%, of those referred for detention have a sentence of six years or less, while approximately 80% of those released on full parole and SR have sentences of this length. Within the group of offenders who were referred for detention, those detained to the end of their sentence were less likely to have a sentence of more than 6 years (27%) than those released at their SR date (36%) and those released after a period of detention (45%). Put another way, almost 50% of those detained to the end of their sentence had sentences of four years or less. These results indicate that of offenders referred for detention, the ones eventually detained to the end of their sentence had received shorter sentences than those who were released.
Table 3-2: Length of sentence by type of release.
| Sample | Type of releases | Distribution of referral cases | ||||||
|---|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referral | Referred but released | Detained then released | Detained to WED | Other1 | |
| Four years or less | 57.55 | 58.04 | 60.31 | 38.39 | 42.93 | 25.19 | 48.59 | 20.16 |
| More than four years to six years | 20.76 | 17.90 | 19.77 | 24.84 | 20.71 | 29.77 | 24.20 | 26.88 |
| More than six years | 21.69 | 24.05 | 19.92 | 36.77 | 36.36 | 45.04 | 27.20 | 52.96 |
| Number of cases | 10,624 | 6,207 | 5,654 | 1,115 | 198 | 131 | 533 | 253 |
1Other includes those cases still detained at the conclusion of the study period, and a few cases who were released and subsequently detained.
Age at Admission
The mean age of offenders at admission for the index release is presented in Table 3-3. The data suggest that those released on SR, either as their first release or as a subsequent release, are younger than those released on full parole or those referred for detention. While those released on SR (either with previous releases or as a first release) have a mean age of less than 30 years, those in the other groups have a mean age of 32 years. The oldest group are those detained to WED, with a mean age of almost 33 years. Older offenders tend to have a lower risk for re-offending and generally, as offenders age, their criminality decreases.
Table 3-3: Average admission age for each release group.
| Sample | Type of releases | Distribution of referral cases | ||||||
|---|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referral | Referred but released | Detained then released | Detained to WED | Other1 | |
| Mean age (average) | 32.54 | 29.15 | 30.07 | 32.41 | 32.72 | 30.65 | 32.88 | 32.09 |
| Age range | 16 - 81 | 17 - 69 | 17 - 78 | 17 - 72 | 19 - 72 | 17 - 67 | 17 - 68 | 18 - 69 |
| Number of cases | 10,481 | 6,179 | 5,630 | 1,115 | 198 | 131 | 533 | 253 |
1Other includes those cases still detained at the conclusion of the study period, and a few cases who were released and subsequently detained.
Female offenders
Female offenders are a very small percentage of the total inmate population, accounting for less than 3% overall. Of the 1,115 inmates referred for detention over the five year period included in the study, only 10 were female. While 4.8% of males were referred for detention, only 1.6% of female releases were referred. Half of these (5) were detained to the end of their sentence, which is similar to the 48% of male referrals detained to the end of their sentence. Those not detained to the end of their sentence fall into the other 3 referral subgroups in frequencies too low to report. Since the number of cases is very small it is not possible to present analyses specifically for female offenders.
Aboriginal Peoples
Table 3-4 presents the percentage of Aboriginal offenders in each release group by year, along with the percentage of Aboriginal offenders amongst all releases for each year. During the five years of the study, Aboriginal offenders represented between 10% and 12% of all releases. However, they only accounted for between 5% and 8% of all full parole releases. Within the group which were given statutory release, following another release, Aboriginal offenders were about equally represented compared to their percentage of all releases. Aboriginal offenders are slightly over-represented in the group given statutory release with no previous releases (14% to 18%) and are greatly over-represented in the group that was referred for detention. Amongst those offenders who were referred for detention, between 22% and 28% were Aboriginal offenders, almost double their representation in the total release population. There is no clear trend in these percentages across the years suggesting that over the period from 1989-90 to 1993-94, the factors which influenced selection did not change.
Table 3-4: Percentage of offenders who are Aboriginal for each type of release, by fiscal year.
| Year | SR - other release | SR - no other release | Referral | All releases | |
|---|---|---|---|---|---|
| 1989-90 | 6.42 | 12.17 | 17.64 | 27.52 | 12.16 |
| 1990-91 | 5.47 | 8.96 | 16.03 | 26.70 | 10.11 |
| 1991-92 | 7.22 | 13.22 | 15.60 | 21.55 | 11.46 |
| 1992-93 | 7.56 | 11.44 | 14.52 | 26.94 | 10.84 |
| 1993-94 | 7.89 | 14.75 | 14.22 | 28.91 | 12.00 |
| All years | 6.96 | 11.93 | 15.87 | 26.37 | 11.3 |
Of those offenders who were referred and then released, 24% were Aboriginal. Of offenders who were detained, 23% of those later released were Aboriginal offenders and 28% of those detained to the end of their sentence were Aboriginal offenders. Of those in the other category, 27% were Aboriginal offenders. These results suggest a consistent distribution across the different referral groups, and that the over representation of Aboriginal offenders in the referral group does not extend to any one of the referral sub-groups. For example, given the group of offenders who were detained to the end of their sentence, Aboriginal offenders were represented at about the same level as those who were referred.
Regional differences
The percentage of each type of release within each region, is presented in Table 3-5. These results indicate, that across regions the number of cases referred varies from a low of 2% in the Quebec region to a high of 7% in the Pacific region.
The last two columns of the table show the percentage of all referral cases presented in each region as compared to their percentage of all releases. If referrals were occurring at the same rate as releases, the percentages in these two columns would be similar. Where the percentage is lower in the referrals column it indicates that referrals are under-represented in that region. The results suggest that Atlantic and Quebec regions have fewer referrals than would be expected given the number of releases they have, while Ontario, Prairie and Pacific regions over-refer.
Table 3-5: Percentage of release types within regions.
| Region | Full Parole | SR- other releases | SR -no other releases | Referral | Number of releases by region | Percentage of releases by region | Percentage of referrals by region |
|---|---|---|---|---|---|---|---|
| Atlantic | 57.89 | 22.00 | 16.92 | 3.19 | 2,541 | 10.79 | 7.26 |
| Quebec | 48.68 | 32.77 | 16.32 | 2.23 | 7,073 | 30.04 | 14.17 |
| Ontario | 44.76 | 21.32 | 28.12 | 5.79 | 6,251 | 26.55 | 32.47 |
| Prairie | 39.53 | 25.44 | 28.31 | 3.72 | 4,941 | 20.98 | 27.78 |
| Pacific | 34.23 | 26.17 | 32.96 | 6.64 | 2,740 | 11.64 | 16.32 |
| Total | 45.03 | 26.26 | 23.97 | 4.77 | 23,546 | 100 | 100 |
The observed differences may be partially due to the differential use of full parole within each region. For example, the Atlantic and Quebec regions have the highest percentage of full parole releases. To take this into account, the results may be looked at in terms of the number of referral cases relative to the number of potential SR releases, since most SR releases are potential detention releases. These results are summarized in Figure 3-3 which suggest less disparity across all regions, except Quebec. Excluding Quebec region, between 15% and 18% of SR releases were referred for detention in 1993-94. In Quebec, only 6% of SR releases were referred for detention in that year.
Figure 3-3 also shows the change in percentage of referral cases relative to all potential SR cases, across the five year period of the study. Potential SR cases include those who were released on SR and those who were referred. The results in the figure indicate there was an increase in the relative percentage of cases referred across all regions, although the increase is less dramatic in Quebec. In a number of the regions referrals went from about 5% of all potential SR cases to more than 15% over the five year period.
Figure 3-3: Percentage of referral cases relative to potential SR cases (all SR cases and those referred).
Summary
The following points summarize the results from this chapter:
- The number of referrals, as a percentage of all releases, has been increasing over the five year period of the study. The number of referrals, as a percentage of all potential SR releases, has been increasing even more dramatically.
- Those referred are more likely to have longer sentences, 6 years or more, than other inmates released, but within the referral group, those detained to the end of their sentence are less likely to have longer sentences.
- Offenders detained to WED tend to be older than those released on SR, and those referred and subsequently released.
- Female offenders are less likely to be referred for detention, but once referred are equally likely, compared to male inmates, to be detained until the end of their sentence.
- Aboriginal offenders are over-represented in the referral group, but given that they are referred, they appear in the same distribution as Non-Aboriginals across the four sub-groups used to study referral cases (referred and released, detained and then released, detained to the end of their sentence, and other).
- There are large regional differences in the percentage of cases referred. Quebec has the lowest referral rate, while other regions refer more than would be expected given their proportion of the inmate population. Year to year increases in the percentage of releases referred are consistent across all regions.
Chapter 4 :
Previous Criminal History
Previous criminal history is a useful predictor of future criminality. If the offenders who are referred are the highest risk offenders, one would expect them to have the largest number of offences and the widest variety of offences. In addition, it would be expected that they would have a higher frequency of more serious offences.Changes over time
As noted earlier, in 1992 the CCRA added new offences to Schedule 1 and created a new Schedule 2 for drug offences. This creates a problem when trying to study the effects of the legislation over time since the inclusion criteria, those who could be referred, changed. In order to reduce the impact of these changes on the analyses presented here, offences committed prior to the CCRA have been coded using the Schedule 1 and 2 offence list created in 1992.
The increase in the number of referrals could be the result of an increase in the number of scheduled offences committed by each offender or the proportion of offenders committing scheduled offences. A change in the number of offences, or the number of offenders committing scheduled offences, would be reflected in a change in the mean number of scheduled offences. Figure 4-1 addresses this issue by showing that there has been no statistically reliable change in the average number of scheduled offences for offenders who have been referred (F(4,1110)=.44, ns). Offenders referred for detention have an average of 2.6 scheduled offences with 70% of them having more than one scheduled offence.
The three comparison groups (full parole, SR-first release and SR-other release), on the other hand, did show changes over the five years. Those released on full parole had an average of 2 scheduled offences in 1990-91 and 1992-93, but this declined to 1.8 in 1993-94. Those released on SR after a full parole or day parole release had a steady increase in the number of scheduled offences from 1.94 to 2.27 in 1992-93; there was a slight decline the following year. Those released on SR for the first time within their sentence showed a steady increase in the average number of scheduled offences until 1993-94 when there was a slight decline.
Figure 4-1: Changes in the average number of scheduled offences 1989-90 to 1993-94.
Overall, these results demonstrate that the average number of scheduled offences was highest for those who were referred for detention, but that the number did not change over the five year period. For the comparison groups, those released on SR showed a general increase in the average number of scheduled offences while those released on parole showed a decrease. These results suggest that the level of criminality may be increasing for those released on SR. However, the data also suggest that the detention referral process is selecting those offenders with the largest number of scheduled offences.
Figure 4-2 shows the average number of Schedule 1 and 2 offences for referral cases from 1989-90 to 1993-94. Although prior to CCRA Schedule 2 offences were not identified as offences for which an offender could be detained, it is possible to identify these offences for statistical purposes as shown here. It can be seen from the figure that Schedule 2 offences contribute minimally to the average number of scheduled offences for referred offenders. In addition, the proportion of drug offences has varied across the years with no consistent pattern. These results suggest that drug offences did not contribute significantly to the increase in referral cases.
Figure 4-2: Average number of Schedule 1 and Schedule 2 offences for referral cases
Types of offences
Table 4-1 presents the percentage of offenders with at least one previous conviction for each of nine offence types by type of release. Within this table it is possible to compare the group of offenders who were referred for detention with the other types of releases, and the referral group is subdivided into four groups based on their release status after the detention hearing.
Examination of the percentages in the table indicates that those convicted of sex offences, assault, manslaughter and violent non-scheduled offences are over-represented in the referral group. Of those referred for detention, 59% had at least one sex offence. This compares to 19% and 10% for the two SR groups (SR only and SR with other release) and 10% for the full parole group. Within the referred group, approximately 60% of those who were initially detained had at least one sex offence, while 44% of those not detained had a sex offence. Forty-two percent of those who were referred had been charged with an assault, and this compares to 28% for those released on SR only, 24% for those released on SR after another release and 13% for those released on parole. Those referred, but released were slightly more likely to have been convicted of an assault (45%) than those referred and then detained. Within the detained group, 42% of offenders who were detained to WED had an assault, while 38% of those detained, but later released, had an assault offence.
Table 4-1: Percentage of offenders with at least one conviction prior to release for selected offence types by release group.
| Type of offence | Type of release | Distribution of referral cases | ||||||
|---|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referral | Referred but released | Detained then released | Detained to WED | Other | |
| Scheduled | ||||||||
| Sex | 10.34 | 9.96 | 19.37 | 59.01 | 44.44 | 60.31 | 61.91 | 63.64 |
| Assault | 13.13 | 24.10 | 28.32 | 41.88 | 45.45 | 38.17 | 42.03 | 40.71 |
| Robbery | 24.03 | 41.65 | 38.47 | 21.61 | 25.25 | 24.43 | 18.39 | 24.11 |
| Manslaughter | 5.00 | 3.50 | 3.24 | 8.79 | 7.58 | 12.21 | 9.19 | 7.11 |
| Other | 9.13 | 11.55 | 10.03 | 5.65 | 8.59 | 4.58 | 5.44 | 4.35 |
| Drug | 27.91 | 13.44 | 10.68 | 2.60 | 3.54 | 4.58 | 1.31 | 3.56 |
| Non-scheduled | ||||||||
| Murder | 4.28 | 0.00 | 0.00 | 0 | 0 | 0 | 0 | 0 |
| Violent1 | 11.57 | 8.62 | 11.12 | 21.88 | 17.68 | 22.90 | 23.45 | 21.34 |
| Non-violent | 83.01 | 90.93 | 83.14 | 62.33 | 63.13 | 67.18 | 61.73 | 60.47 |
| Number of Cases | 10,624 | 6,207 | 5,654 | 1,115 | 198 | 131 | 533 | 253 |
Note: FP - Full parole; SR & other - Statutory release,
but with previous day parole or full parole;
SR only - Statutory release without a previous release on the current sentence;
Referral - referred for detention.
1 Violent non-scheduled offences are mainly offences which were processed through the courts as a summary conviction, rather than by indictment. In general, this means the offences are less serious.
Of the referrals, 9% had been convicted of manslaughter while for other release groups, the percentage ranged from 3% to 5%. Those who were detained were more likely to have been convicted of manslaughter than those released after referral.
Twenty-two percent of the referral group had also been convicted of a non-scheduled violent offence as compared to about 10% of the other release groups. While 62% of the referral group had other convictions for non-violent non-scheduled offences this was much higher, over 80%, for the other release groups. These results indicate that referral cases are involved in more violent offending.
Within the group who were referred, there was less likely to be a conviction for robbery and drug offences than for those with other types of releases. Of those referred, only 3% had a conviction for a drug offence, while for those given other types of releases the percentage convicted of a drug offence varied from 11% for SR only releases to 28% for full parole releases.
These results suggest that referrals are most likely to occur for those convicted of sexual offences and assault offences, as well as other non-scheduled violent offences.
Table 4-2 presents a comparison of the types of previous offences by type of release for the years 1989-90 and 1993-94. The purpose of the table is to show how the offence types for different types of release have changed over the years. In 1993-94 a higher percentage of the referral cases had at least one sex offence, assault offence, drug offence or violent, non-scheduled offence, than in 1989-90. However, the table also shows that the percentage of offenders with these offences has increased for the other types of release. These results suggest that, while the referral cases in 1993-94 are more violent than in the earlier fiscal year, the same pattern is evident for the other release groups.
Table 4-2: Percentage of offenders with at least one offence by type of release for 1989-90 and 1993-94
| Type of offence | 1989-90 | 1993-94 | ||||||
|---|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referral | FP | SR & other | SR only | Referral | |
| Scheduled | ||||||||
| Sex | 9.92 | 8.61 | 16.36 | 54.36 | 9.74 | 13.34 | 23.74 | 58.50 |
| Assault | 13.10 | 24.08 | 26.97 | 40.27 | 11.48 | 25.81 | 30.16 | 45.92 |
| Robbery | 26.10 | 40.61 | 38.66 | 18.79 | 19.35 | 44.58 | 39.68 | 21.43 |
| Manslaughter | 6.26 | 3.93 | 3.28 | 10.74 | 3.61 | 3.47 | 3.10 | 7.14 |
| Other | 11.27 | 12.16 | 9.67 | 5.37 | 6.96 | 11.82 | 10.55 | 4.76 |
| Drug | 24.79 | 12.23 | 8.21 | 2.01 | 29.13 | 15.73 | 11.35 | 4.08 |
| Non-scheduled | ||||||||
| Murder | 4.07 | 0.006 | 0.00 | 0.00 | 3.65 | 0.00 | 0.00 | 0.00 |
| Violent1 | 11.53 | 6.74 | 10.93 | 19.46 | 11.39 | 10.95 | 13.53 | 23.81 |
| Non-violent | 81.73 | 91.83 | 84.08 | 58.39 | 85.35 | 88.61 | 78.44 | 62.24 |
| Number of Cases | 1,916 | 1,603 | 1,583 | 149 | 2,300 | 922 | 872 | 294 |
Note: FP - Full parole; SR & other - Statutory release,
but with previous day parole or full parole;
SR only - Statutory release without a previous release on the current sentence;
Referral - referred for detention.
1 Violent non-scheduled offences are mainly offences which were processed through the courts as a summary conviction, rather than by indictment. In general, this means the offences are less serious.
Level of criminal activity
The level of criminal activity can be estimated by the number of all offences for which an offender was convicted. Earlier, the average number of scheduled offences was presented, but in this section we are looking at all types of offences for which the offender received a sentence to be served in federal prison. Level of criminal activity has been associated with risk of reoffending upon release.
Table 4-3 presents the level of criminal activity by type of release. Criminal activity is divided into one or two offences, three to five offences and six or more offences. Offences included in these counts are those for which sentence had to be served in a in a federal prison. The results indicate that while 50% or more of those released on SR had 6 or more offences, only 35% of those referred for detention had this level of criminal activity in their past. In addition, while 38% of those referred, but released, had six or more offences, only 33% of those detained to the end of their sentence fell into this category. Overall, those referred, and within the referred group, those detained to the end of their sentence, had previous criminal activity levels closer to those released on full parole than those released on SR.
Table 4-3: Level of criminal activity (number of offences) by type of release.
| Number of previous federal offences | Type of releases | Distribution of referral cases | ||||||
|---|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referral | Referred but released | Detained then released | Detained to WED | Other1 | |
| One or two | 40.77 | 16.08 | 22.32 | 31.03 | 31.31 | 29.77 | 31.52 | 30.43 |
| Three to five | 27.49 | 25.71 | 28.49 | 33.99 | 30.30 | 32.82 | 35.65 | 33.99 |
| Six or more | 31.74 | 58.21 | 49.19 | 34.98 | 38.38 | 37.40 | 32.83 | 35.57 |
| Number of cases | 10,624 | 6,207 | 5,654 | 1,115 | 198 | 131 | 533 | 253 |
1 Other includes those cases still detained at the conclusion of the study period, and a few cases who were released and subsequently detained.
Aboriginal offenders
Table 4-4 presents the types of offences that Aboriginal and Non-aboriginal offenders were convicted of prior to their release. These data indicate that Aboriginal offenders referred for detention are less likely (50%) than Non-aboriginal offenders (62%) to have
Table 4-4: Percentage of offenders with at least one conviction for each offence type by release group for Aboriginal and Non-aboriginal offenders.
| Type of offence | Type of release | Distribution of referral cases | |||||||
|---|---|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referral | Referred but released | Detained then released | Detained to WED | Other | ||
| Aboriginal offenders | |||||||||
| Scheduled | |||||||||
| Sex | 23.27 | 14.32 | 21.96 | 50.34 | 35.42 | 56.67 | 54.42 | 49.28 | |
| Assault | 26.66 | 39.05 | 37.35 | 57.82 | 70.83 | 46.67 | 55.10 | 59.42 | |
| Robbery | 19.62 | 36.89 | 33.22 | 25.17 | 31.25 | 23.33 | 20.41 | 31.88 | |
| Manslaughter | 12.45 | 10.54 | 7.47 | 15.65 | 10.42 | 23.33 | 16.33 | 14.49 | |
| Other | 5,95 | 7,57 | 7,36 | 6,80 | 14,58 | 6,67 | 6,80 | 1,45 | |
| Drug | 7.44 | 6.22 | 5.13 | 1.02 | 0 | 0 | 0 | 4.35 | |
| Non-scheduled | |||||||||
| Murder | 6,09 | 0,14 | 0 | 0 | 0 | 0 | 0 | 0 | |
| Violent | 14,48 | 6,62 | 6,42 | 14,97 | 10,42 | 16,67 | 17,69 | 11,59 | |
| Non-violent | 66.17 | 85.14 | 80.38 | 73.47 | 79.17 | 73.33 | 70.75 | 75.36 | |
| Number of Cases | 739 | 740 | 897 | 294 | 48 | 30 | 147 | 69 | |
| Non-Aboriginal offenders | |||||||||
| Scheduled | |||||||||
| Sex | 9.38 | 18.87 | 9.37 | 62.12 | 47.33 | 61.39 | 64.77 | 69.02 | |
| Assault | 12,12 | 26,61 | 22,09 | 36,18 | 37,33 | 35,64 | 37,05 | 33,70 | |
| Robbery | 24.35 | 39.84 | 42.28 | 20.34 | 23.33 | 24.75 | 17.62 | 21.20 | |
| Manslaughter | 4.43 | 2.44 | 2.54 | 6.33 | 6.67 | 8.91 | 6.48 | 4.35 | |
| Other | 9.37 | 10.54 | 12.08 | 5.24 | 6.67 | 3.96 | 4.92 | 5.43 | |
| Drug | 29.45 | 11.74 | 14.40 | 3.17 | 4.67 | 5.94 | 1.81 | 3.26 | |
| Non-scheduled | |||||||||
| Murder | 4.15 | 0.04 | 0 | 0 | 0 | 0 | 0 | 0 | |
| Violent | 11.34 | 12.03 | 8.89 | 24.36 | 20.00 | 24.75 | 25.65 | 25.00 | |
| Non-violent | 84.27 | 83.68 | 91.71 | 58.34 | 58.00 | 65.35 | 58.29 | 54.89 | |
| Number of Cases | 9,877 | 4,754 | 5,464 | 821 | 150 | 101 | 386 | 184 | |
been convicted of a sexual offence. On the other hand, Aboriginal offenders are more likely to have been convicted of an assault (58%) compared to Non-aboriginal offenders (35%). As was seen in the previous analysis, those who were detained (and either later released or held until the end of their sentence) were more likely to have been convicted of a sexual offence in both the Aboriginal and Non-aboriginal groups.
Aboriginal offenders were more likely to have been convicted of manslaughter (16%) than Non-aboriginal offenders (6%). However, Aboriginal offenders were less likely to have been convicted of a drug offence (1%) than Non-aboriginal offenders (3%). In terms of other release groups, Aboriginal offenders are more likely to have been convicted of sex offences, assault offences and manslaughter than non-Aboriginal offenders.
Summary
The following points summarize the findings in this chapter:
- Offenders referred for detention have a greater number of scheduled offences than other offenders, and this number has been constant over the five years of the study. This would suggest that the increase in the number of referrals is not the result of the number of convictions offenders have for scheduled offences.
- Drug offences are a relatively small percentage of the total number of scheduled offences and have not been increasing over the five years of the study. The observed increase in the number of referred cases cannot be attributed to the addition of Schedule 2 to the CCRA, nor can it be attributed to a change in the number of drug offences.
- Referral cases are more likely to have been convicted of scheduled sexual and assault offences as well as non-scheduled violent offences than other offenders. They are less likely to have been convicted of robbery offences. Amongst those referred, offenders who were detained were more likely to have been convicted of a sexual offence than those who were released. This suggests that detention referrals are targeting sex offenders.
- Based on the number of previous criminal offences, offenders referred for detention have a lower level of criminal activity than those released on SR, and have a level of criminal activity more similar to those released on full parole. This result suggests that, based on this risk factor, offenders referred for detention may have a lower risk to reoffend than those released on SR.
- There appears to be an increase in the variety of violent offences (more offenders committing different types of violent offences) for both referred offenders and those released on full parole and statutory release over the five year period of the study. This could account for some of the increase in referrals, although the result is inconsistent with that in number 1 above.
- Aboriginal offenders who are referred, are less likely to have been convicted of sexual offences and drug offences and are more likely to have been convicted of assault and manslaughter.
Chapter 5 :
Recidivism
At first glance, measuring recidivism appears relatively straight forward. For example, the number of people returning to prison is often thought of as a measure of recidivism. While conceptually accurate, this measure is of little practical value unless one considers:- the length of the follow-up period (one, two, five, or ten years, or life),
- the reasons for returning to prison (such as violations of conditional release or new sentence) and
- whether readmission rates are for those still under sentence (being supervised in the community) or for those who have completed their sentence.
To address these problems the study uses two measures of recidivism and these are presented below.
One measure of recidivism used in the study is readmission to federal corrections within a fixed period of time; one and two years is used in this study. This is a somewhat crude measure, but it is useful for a quick snapshot of what the return rates are. However, it does not distinguish between readmission for a technical violation of conditional release and readmission as a result of a new offence. In addition, it does not account for differences in the probability of readmission between someone under supervision in the community and someone who has passed the end of their sentence date. In most studies, we only need to follow offenders to the end of their sentence and so readmissions occur under similar circumstances for all offenders. However, in a study of detention we also have to be concerned about follow-up after the end of the sentence for those detained to the end of their sentence.
While under supervision both a technical violation of the conditional release, or arrest (without waiting for a conviction) can lead to readmission. In addition, almost any jail sentence obtained while under supervision, regardless of the nature of the crime, will result in a readmission to federal prison. However, for offenders who have completed their sentence, readmission to a federal prison can only occur after sentencing by the court, and then, only if the sentence is for two years or more. For those under supervision, readmission to a federal prison is more likely and occurs in a shorter time than for offenders who have completed their sentence. Therefore, readmission is a useful measure of recidivism for offenders being supervised in the community.
A second measure of recidivism is time to a new offence. The advantage of this measure is that it is equally useful for those who are and are not under supervision in the community. Unfortunately, current data systems do not provide the date on which new offences occur. Therefore, a substitute measure must be used. The most consistently available data is sentencing date and it was selected for use. While readmission is affected by being under supervision, sentencing date is not. All new offences must be processed by the courts and it is assumed that this process will require the same amount of time for both offenders under supervision and offenders who have completed their sentence. While time to sentencing may be influenced by factors associated with the court, and extraneous to the offender, it is assumed that these factors will influence both groups equally.
Time to new sentence might be a problem if the follow-up period was very short, but the current study uses a follow-up period of 24 months. It is expected that most sentencing occurs within six months of the offence, and almost all but the most serious offences have a sentencing date within one year of the offence. The effective follow-up period is therefore between 12 and 18 months.
While studying recidivism it is possible to measure the effect of incapacitation, that is, the effect of keeping the person in custody, rather than allowing time to be served in the community. Take for example the cases of two offenders who have the same sentence expiration date. One offender is released with 24 months of supervision to serve and the other is released at the end of the sentence after being detained. If both offenders commit a crime six months after the end of their sentence, the incapacitation will have had no effect because the crimes occurred at the same time. On the other hand, if they both commit a new offence within 6 months of release to the community, we can conclude that incapacitation had an effect, since one crime was delayed for 2 years while the person was incarcerated.
To analyze the effect of incapacitation, recidivism for those detained is measured from the date of detention being ordered (incapacitation), not from the date of release into the community; the incapacitation period is from the detention date to the release date. If the time to recidivism from the detention date is significantly longer than the time to recidivism from release to the community then it can be concluded that incapacitation had an effect on delaying recidivism.
Readmissions and new offences, throughout this study, refer to readmissions to federal prisons and new offences resulting in federal prison terms.
Readmissions
Time to readmission
Figure 5-1 presents the percentage of offenders who were readmitted after one and two years of follow-up. In addition, the percentages for readmissions and readmissions with a new offence and the number of cases for each group in the two year follow-up are presented in Table 5-1.
The full parole and the two statutory release (SR) groups follow the expected pattern with the lowest readmission rate for those released on full parole (24% after two years) and the highest rate of readmission for those released on SR without any other previous release (49%). This latter group would have served their entire sentence to the SR date without a release on either day parole or full parole. As such, they are viewed as a higher risk group of offenders. The group of offenders who were referred for detention, but who were released had readmission rates very similar to the SR groups (46%) as did those who were ordered detained initially, but who were subsequently ordered released (41%) by the National Parole Board prior to the end of their sentence date.
Figure 5-1: Percentage of releases readmitted within one and two years.
Table 5-1: Percentage of releases readmitted within two years and number of cases.
| Type | Type of releases | Distribution of referral cases | ||||||
|---|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referred but released | Detained then released | Detained to WED | Other1 | ||
| % readmitted within 2 years | 24.06 | 44.03 | 49.39 | 46.34 | 41.49 | 15.88 | 1.57 | |
| % readmitted with a new offence2 | 7.38 | 13.63 | 11.62 | 12.81 | 7.60 | 14.04 | 0 | |
| Number of releases | 8324 | 5285 | 4782 | 164 | 94 | 233 | 254 | |
Note: These results do not include cases where there was less than two years from the release date to the study end date, March 31, 1995.
1 Other includes those cases still detained at the
conclusion of the study period, and a few cases who were released
and subsequently detained.
2 Includes those readmitted under supervision with a new
offence and those readmitted through a warrant of committal after
their sentence expired.
The group which was detained to their WED had the lowest readmission rate at 16% after two years. However, the lower readmission rate for those detained to the end of their sentence does not mean these offenders are the least likely to reoffend. Readmission takes longer for this group because it occurs after sentencing by the court for a new offence, whereas for offenders in the other groups, who are being supervised, readmission can occur very quickly through revocation of the conditional release. It is estimated that the effective follow-up for those detained to the end of their sentence is 12 to 18 months (allowing 6 to 12 months for processing of new criminal charges), not 24 months as for those in the other groups who are being supervised. Results presented later, which use sentence date rather than readmission date, address more effectively the question of recidivism for those detained to the end of their sentence.
It can be seen that for all types of releases, the majority of readmissions occur within the first year after release. For the SR groups and those referred but released, only 15% of readmissions occurred during the second year of follow-up. Readmissions in the second year were 28% of all readmissions for those released on parole and were only 9% for those who were initially detained and then released.
For those who were detained to the end of their sentence, 45% of the readmissions occurred in the second year of follow-up. The higher percentage in the second year is probably partly the result of the court process leading to readmission which delays the readmission date.
Table 5-1 also presents the percentage of first readmissions that occurred following a new offence. These results indicate that, in general, the referred and SR groups had reoffence rates between 12% and 14%, with the exception of the detained and then released group of whom 8% were readmitted with a new offence on their first readmission. The results for this latter group are partly explained by the fact that 82% of the readmissions were for technical violations of their conditional releases as shown in Table 5-2 (which follows in the next section).
While these results indicate that those detained to the end of their sentence have a re-offence rate similar to the other groups it is likely that these data underestimate the number of readmissions for new offences for all groups except those detained to the end of their sentence which could only be readmitted with a new offence. For the other groups, only the first readmission is counted so a subsequent readmission with a new offence does not appear in the table and frequently a readmission for technical violation is recorded and not changed, when later information confirms that a new offence was committed.
Another way of looking at these data is presented in Figure 5-2 which shows the cumulative percentage of readmissions by month for a two year follow-up for each of the groups. Basically, the graph shows the number of months from release to readmission for those who could be followed for two years after release from prison. On the vertical axis is the cumulative percentage of cases within each group that have been readmitted and along the horizontal axis is the number of months to readmission.
Figure 5-2: Cumulative percentage readmitted by time to readmission from release date.
These results show readmission patterns which are very similar for the SR groups and the referred groups who were released either at their SR date (referred and released), or some time later following a subsequent review by the National Parole Board (Detained and then released). The readmission rate varies by 10% across these groups and the data indicate the following ranking, from most likely to be readmitted to least likely to be readmitted,
- SR only,
- referred and released,
- SR preceded by another release,
- detained and then released.
The steep slope on the curves during the first 12 months indicates, as was shown previously, that readmission is most likely to occur within the first 12 months and there is a decreasing rate of failure after one year. In addition, the steepest slope in the first few months after release is for the group that were detained and then released. Results for this group are shown with a broken line on the graph and this pattern of readmission suggests that the detention period simply delayed the onset of problems.
Those offenders released on full parole and those detained until the expiry of their sentence had the lowest rate of readmission over the two year period. It is interesting that the rate of readmission, as shown by the slope of the curve, is similar for both full parole releases and those detained until the end of their sentence.
The results suggest that the referred groups do not differ very much from those who received statutory release. Interestingly, the group which was detained and then released shows the steepest readmission curve suggesting that whatever effect there was of detaining them was quickly lost upon their release.
While the results for those detained to WED appear most dramatically different from the others, this may be due to the fact that readmission for this group can only occur after a new conviction. The court process, which can add months to the readmission date, makes comparisons of this group with the others, on this graph, problematic.
Reasons for readmission
The reasons for a readmission can be classified into three main categories:
- Warrant of committal; where an offender reached the end of the sentence and then committed a new crime for which they were sentenced.
- Conditional release failure for technical reasons; where, in the opinion of the parole supervisor and the National Parole Board, the offender has failed to meet the conditions associated with the release, or has shown a high probability of re-offending, or some aspect of the release has changed such as the loss of a halfway house bed or cancellation of a program.
- Conditional release failure with a new offence; where at the time of readmission the offender has been charged with a new offence. Generally, these cases are underestimated in the electronic data systems because information on new offences is not always available at the time of readmission.
Table 5-2 presents the percentage of offenders readmitted by type of admission and type of release. The first line of data in the table presents the percentage of each release type based on the number of readmissions. The second line of data, for each release type, is the percentage of each admission type for all releases, not just those who were readmitted. The data in the table show that of the readmissions, between 69% and 82% are for technical violations without a new offence and that between 13% and 24% are for new offences committed while on conditional release. An additional 5% to 9% are readmitted for a new offence after they have completed their sentence.
What is particularly interesting in the table is the consistency, across the types of release, in the percentage of offenders readmitted for technical violations of conditional release with a range from 69% to 76%, excluding the detained and then released group for which 82% were readmitted for a technical violation. Likewise, readmission with a new offence for those released on, or before their SR date, account for approximately 21% of readmissions (18% to 24%). The detained and then released group has the highest level of readmissions for technical violations (79%) and the lowest for readmission with a new offence (13%).
Table 5-2: Percentage of cases readmitted by type of admission and type of release after two year follow-up.
| Type of release | Warrant of committal | Conditional release-technical | Conditional release- new offence | Other admission type | Number of cases readmitted | Number of releases |
|---|---|---|---|---|---|---|
| Full parole | ||||||
| % of readmissions | 6.69 | 69.18 | 23.98 | 0.15 | 2002 | |
| % of releases | 1.61 | 16.64 | 5.77 | 0.04 | 8323 | |
| SR & other | ||||||
| % of readmissions | 9.33 | 68.89 | 21.62 | 0.17 | 2327 | |
| % of releases | 4.11 | 30.33 | 9.52 | 0.08 | 5285 | |
| SR only | ||||||
| % of readmissions | 5.67 | 76.33 | 17.87 | 0.13 | 2362 | |
| % of releases | 2.80 | 37.70 | 8.82 | 0.06 | 4782 | |
| Referred, but released | ||||||
| % of readmissions | 5.26 | 71.05 | 22.37 | 1.32 | 76 | |
| % of releases | 2.44 | 32.93 | 10.37 | 0.61 | 164 | |
| Detained then released | ||||||
| % of readmissions | 5.13 | 82.05 | 12.82 | 0 | 39 | |
| % of releases | 2.17 | 34.78 | 5.43 | 0 | 92 | |
| Detained to wed | ||||||
| % of readmissions | 89.19 | 0 | 0 | 10.81 | 37 | |
| % of releases | 14.04 | 0 | 0 | 1.70 | 235 | |
| Other1 | ||||||
| % of readmissions | 0 | 100.00 | 0 | 0 | 4 | |
| % of releases | 0 | 1.59 | 0 | 0 | 251 | |
| Total | 6847 | 19132 | ||||
Note: These results do not include cases where there was less than two years from the release date to the study end date, March 31, 1995.
1 Other includes those who were still detained and a small number of cases who were released and then detained.
When the percentages are calculated on the basis of all releases, not just readmissions, those released at, or after their SR date (the two SR groups, the referred and released group, and the detained then released group) have very similar results. Approximately 34% (32% to 38%) of these offenders are readmitted for a technical violation while approximately 9% (5% to 10%) are readmitted with a new offence while under supervision.
Offenders held to the end of their sentence cannot be returned for violations of conditional releases and therefore all of their readmissions are from the courts through a warrant of committal.
Age at admission
Figure 5-3 presents the readmission rates by age at admission for the release groups. The data clearly indicate that the readmissions rates are lower for older offenders which is consistent with other findings (Andrews & Bonta, 1994). For offenders who were referred the readmission rate is almost 50% lower for the group over 35 years of age than for the youngest group, under 25 years. This age effect is consistent across the referral sub-groups. Amongst those who were detained to WED, 11% of those over 35 years of age were readmitted within two years while 19% under 25 were readmitted.
Figure 5-3: Readmissions within two years by age at admission
The lower readmission rate for the older offenders suggests that they are lower risk offenders. Recall that the mean age for offenders who were referred was 32, suggesting that the risk for these offenders may have been lower than for other types of releases.
New offence
The readmission data, presented above, provides one measure of recidivism. However, there are problems with using readmission as discussed earlier. Readmission, while under supervision, will occur much faster than readmission after the end of the sentence date since, in the latter case, a court case must be held before the readmission can occur whereas in the former, readmission follows almost immediately after knowledge that an offence may have occurred or is likely to have occurred. The alternative to readmission is to use re-offence date.
Unfortunately, CSC data systems do not consistently record the date of new offences so it is necessary to select a substitute measure. The date on which an offender is sentenced represents the most accurate approximation to an offence date that is available. It is well known that the sentence date may occur some time after the offence, and may even be delayed after the conviction has been rendered. However, delays in arriving at the sentencing date should be consistent across all offenders and therefore the sentence date represents an acceptable alternative.
Time to new offence
The time to a new offence by type of release is shown in Figure 5-4. For all groups the cumulative percentage who commit a new offence is lower than for readmission, as expected. The curves on the graph show similar patterns for the two SR groups. The referred, but released group, has a curve similar to that of the SR groups, but it is lower.
Figure 5-4: Cumulative readmissions by time to new offence for release groups.
The full parole, detained, and detained then released groups all show similar slopes for their readmission curves with steady increases in the percentage of offenders with new sentences. Within these groups, full parole has the highest percentage of cases with new sentences, at 18%, while for those detained to WED, 16% had a new offence within two years, and 12% of those detained and released had a new offence. These results suggest, that in terms of new offences, those offenders in the detained and released group, and the detained to WED group were less likely to commit a new offence which resulted in a new federal sentence (these analyses do not consider provincial sentences).
It should be noted that those in the groups other than those detained to the end of their sentence, are likely to have spent some of the two year follow-up period incarcerated since between 25% and 50% were revoked while on conditional release and therefore spent at least some of the follow-up period incarcerated. In addition, those who were detained to WED may have been arrested and spent time incarcerated awaiting trial. Therefore, the data presented are for time to new offence, and do not represent time at risk to commit a new offence.
Incapacitation
Imprisonment can have a number of different functions. Among these are punishment, treatment, and incapacitation. Incapacitation refers to the fact that while in prison one cannot commit crimes within the general population, although it is still possible to commit crimes in the prison environment. While it is hoped that the increased prison time for those who are detained will lead to improved treatment of their problems, it is recognized that one of the main functions of detention is incapacitation. The offender presents such a high risk that it is considered unsafe to allow him or her into the community.
To determine if incapacitation has much benefit it is necessary to look at the time it takes offenders to commit a new crime from the point when they are released as compared to the period of detention. For the current study, 235 offenders were detained to the end of their sentence and had a two year follow-up period. Of these, 38 (16.2%) were sentenced for a new offence (scheduled and non-scheduled) within the two year follow-up period after release. If the follow-up is measured from the date of detention, rather than from the release date, we can get a measure of the effect of incapacitation.
Of the 38 detained offenders who committed a new offence, 16% (6) were re-sentenced within one year of their detention decision, 47% (18) were re-sentenced within two years and 89% (34) had been re-sentenced within three years of their detention decision. Given that most of these offences occurred within two years of the release date, and within three years of the detention decision it would appear that, on average, the incapacitation effect of detention prevented new offences for about one year. The effect is actually shorter than this since sentence date, not offence date, is being used to calculate the time lag. The effect would likely be reduced by about six months if offence date were used.
While a one year, or 12 month, incapacitation effect may seem short, it should be remembered that the average (median) detention period is only 415 days, or slightly less than 14 months. In addition, data presented earlier indicated that 50% of those detained to the end of their sentence had sentences of four years or less. For a sentence of four years, the maximum detention period possible is 16 months. Overall then, the incapacitation effect of detention is relatively short.
Age at Admission
As with readmissions, age is related to the percentage of cases with a new offence within two years as shown in Figure 5-5. Older offenders are almost half as likely to have committed another offence within two years than are younger offenders. For those detained to the end of their sentence, 11% of those over 35 years of age committed a new offence within two years and 20% under 25 committed a new offence within the same period. These results, combined with those for readmission, suggest that older offenders with a more extensive criminal history may be more likely to be detained, even though their age makes them a lower risk for re-offending. The increased number of offences for this group is, in part, due to their longer period of exposure to criminal activity simply because of their age. Correlational analyses indicate that age and criminal history were about equally related to new offending.
Figure 5-5: Percentage of offenders with new offence within two years by age
Type of offence
The preceding section presented time to new offence, for any type of offence. The following section breaks offences into scheduled and non-scheduled offences, and provides a further breakdown of scheduled offences into sexual offences, assaults, robbery, manslaughter, drug offences, and non-scheduled offences into violent and non-violent offences.
Since the rate of re-offending with a scheduled offence is considerably lower than for all offences the results here will appear different from those presented in the preceding graph which presented re-conviction for all offences. An additional difference between the previous graph and these data is that in this section offences are subdivided into groups which creates a different pattern of reoffending.
Table 5-3 presents the percentage of offenders who were convicted of at least one offence in each of the offence types at their first sentencing after release. The first line of percentages in the table shows the percentage for any scheduled offence. While among those released on full parole, only 6% committed a new scheduled offence within two years of release, 11% to 13% of those on statutory release and 11% of those referred for detention committed a new scheduled offence.
Within the referral group, 15% of those released and 11% of those detained to the end of their sentence committed new scheduled offences within two years. Only 4% of those who were detained, but subsequently released, committed a new scheduled offence within two years. Put another way, these results indicate that of the 235 inmates who were held in prison until the end of their sentence, only 26 of them committed another scheduled offence within two years of their release. The rate of reoffending is the same as for those who were released at their statutory release date.
Sex offences were most likely to have been committed by those referred, and among those detained to the end of their sentence, 6% committed at least one sex offence within two years of release. This compares to less than 1% of those released on full parole or statutory release. Approximately 4% of those released on statutory release and those referred were convicted of a scheduled assault offence. This finding is not surprising given 60% of those referred had been convicted of a sexual offence in the past while less than 20% of those not referred had committed a sexual offence.
Table 5-3: Percentage of offenders with a two year follow-up having at least one offence within two years at first sentencing after release.
| New offence | Type of release | Distribution of referral cases | |||||
|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referral | Referred but released | Detained then released | Detained to WED | |
| Scheduled | |||||||
| All offences | 5.64 | 11.01 | 12.80 | 11.00 | 14.67 | 4.35 | 11.06 |
| Sex | 0.31 | 0.74 | 0.90 | 4.48 | 4.27 | 2.17 | 5.53 |
| Assault | 1.99 | 3.69 | 5.33 | 3.87 | 4.88 | 1.09 | 4.26 |
| Robbery | 2.01 | 5.55 | 5.67 | 1.89 | 4.88 | 1.09 | 2.13 |
| Manslaughter | 0.05 | 0.11 | 0.13 | 0.00 | 0.00 | 0.00 | 0.00 |
| Other | 0.41 | 0.81 | 1.09 | 0.61 | 1.22 | 0.00 | 0.43 |
| Drug | 1.38 | 1.31 | 1.19 | 0.20 | 0.61 | 0.00 | 0.00 |
| Non-scheduled | |||||||
| Murder | 0.00 | 0.19 | 0.23 | 0.20 | 0.00 | 0.00 | 0.43 |
| Violent | 0.85 | 1.50 | 2.24 | 2.04 | 0.61 | 1.09 | 2.98 |
| Non-violent | 15.63 | 27.80 | 28.90 | 10.49 | 17.07 | 7.61 | 6.81 |
| Number of Cases | 8322 | 5284 | 4782 | 491 | 164 | 92 | 235 |
The data in Table 5-3 also indicate that referral cases had about the same level of readmissions with an assault conviction (4% for referrals vs. 4% to 5 % for SR releases), but a lower rate of readmission with a robbery offence (2% vs. 6%).
Table 5-4 presents the rates of readmission for the study groups for both first admissions and subsequent admissions. When all offences within the two year follow-up are included the SR only group has a 19% rate of reconviction for a scheduled offence, the SR with other releases has a 17% reconviction rate, while the reconviction rate for those referred is only 13%.
Table 5-4: Percentage of offenders with a two year follow-up having at least one scheduled offence at first and subsequent sentencing after release.
| Follow-up | Type of release | Distribution of referral cases | |||||
|---|---|---|---|---|---|---|---|
| FP | SR & other | SR only | Referral | Referred but released | Detained then released | Detained to WED | |
| First offence(s) | 5.64 | 11.01 | 12.80 | 11.00 | 14.67 | 4.35 | 11.06 |
| All offenses | 7.70 | 16.65 | 18.53 | 13.24 | 16.46 | 5.43 | 15.04 |
| Number of Cases | 8322 | 5284 | 4782 | 491 | 164 | 92 | 235 |
Aboriginal and Non-aboriginal Comparisons
Results presented in Figure 5-6 indicate that the readmission rate for Aboriginal offenders is significantly higher than for Non-aboriginal offenders. The data in the graph are for a two year follow-up. For most release groups Aboriginals offenders have a readmission rate more than 12 percentage points above the Non-aboriginal offenders. For those Aboriginal offenders who were referred, but subsequently released, the readmission rate is almost double that of Non-aboriginal offenders. However, for those detained to the end of their sentence, there is very little difference in the readmission rate for the two groups.
Figure 5-6: Percentage of Aboriginal and Non-aboriginal offenders readmitted within two years of release.
The percentage of Aboriginal and Non-aboriginal offenders who have a new offence (scheduled or non-scheduled) within two years of release is shown in Figure 5-7. For new offences the large differences between Aboriginal and Non-aboriginal offenders seen with readmissions, is reduced. For most groups the differences are relatively small. However, for those referred, but released, Aboriginal offenders are almost twice as likely to have a new offence within two years of release. Noticeable differences showing higher readmission rates for Aboriginal offenders are also evident for those released on full parole and those detained and then released.
Figure 5-7: Percentage of Aboriginal and Non-aboriginal offenders with at least one new offence (scheduled or non-scheduled) within two years of release.
The large differences between Aboriginal and Non-aboriginal offenders for readmission suggests that Aboriginal offenders may be more likely to be revoked while under supervision, and this effect is evident for regular releases and for those released after having been referred. The Aboriginal/Non-aboriginal differences are less evident when recidivism is measured using new offences, rather than readmission for any reason. Where supervision does not play a role in readmission, such as for offenders detained to the end of their sentence, the differences all but disappear.
Regional Comparisons
Readmission rates for full parole and statutory release are relatively consistent across the regions, except for the full parole readmissions in the Atlantic region which is substantially higher than for parole in the other regions. These results are shown in Figure 5-8. Readmissions from statutory releases are over 50% in the Prairie region. Overall, the readmission rate is lower for those referred for detention (30% in four of five regions) than for those released on their statutory release date. However, when the data for referrals are subdivided the results show that readmission rates are highest for those referred, but released, with rates reaching 60% in the Prairie region. As with most other results, the group detained to the end of their sentence have the lowest readmission rates, and the pattern across regions is similar to that observed for full parole releases.
Figure 5-8: Percentage readmitted within two years by region.
It should be remembered that readmission includes revocation of conditional release (full parole or statutory release) for technical violations as well as following the commission of a new offence. Results presented earlier indicated that revocations for technical violations account for the largest percentage of all readmissions. It should also be noted that all readmission for those detained to the end of their sentence result from a new offence.
Readmission rates across regions for new offences (scheduled or non-scheduled) within two years of release (measured using new sentences) are presented in Figure 5-9. The Atlantic region has the highest rate of new offences in the two year follow-up period for those referred for detention at just over 30% as compared to about 20% in the other regions. The re-offence rate is highest for those referred, but released. In the Atlantic, Ontario and Prairie regions, over 30% of the offenders in this group have a new offence within two years of release. This compares with a re-offence rate of less than 20% for those detained to the end of their sentence in all regions except Atlantic, where the re-offence rate approaches 50%.
Figure 5-9: Percentage of offenders with a new offence (scheduled or non-scheduled) within two years, by region.
These results suggest that of those referred for detention, the least likely to reoffend were detained to the end of their sentence, whereas those who were released were more likely to commit a new offence. As with other analyses, the group that were detained to the end of their sentence exhibit a similar pattern to those who were released on parole.
Summary
The following points summarize the findings in this chapter:
- The pattern of readmissions is similar for the SR and the referred groups who were subsequently released before WED. The detained to WED group had the lowest readmission rate, but this may reflect the need for court action prior to readmission, rather than a difference between this group and the others.
- Time to a new offence is shortest for those released on SR and those referred and released. Those who were detained had re-offence rates lower than those released on full parole.
- On average, the incapacitation effect of detention lasts only about one year, the period the offender remains in prison.
- Only 16% of those detained to the end of their sentence, and 12% of those detained and then released, had a new offence of any type two years after their release. Only 11% committed a scheduled offence as their first new offence, while 15% did commit a new scheduled offence sometime after their release (given a minimum of two years following release).
- The readmission and re-offence rate was lowest for those over 35 years of age. Given that the average age of the referral group is 32, and the average age of those detained to WED is 33, it is likely that some of these offenders have a lower risk of re-offending than would be expected given their offences.
- Aboriginal offenders have higher readmission rates than Non-aboriginal offenders, but these differences are reduced when new offence is used as the recidivism measure. These results suggest that Aboriginal offenders may be more likely to have their conditional release revoked, even though they are not more likely to commit a new offence. Additional investigations may be needed to explain the observed differences.
- Regional differences are evident for both readmission rates and new offences within two years of release. Patterns of readmissions across referral groups and regions are consistent with the general patterns observed. Those detained have lower readmission and re-offence rates. Striking is the relatively higher re-offence rate in the Atlantic region for full parole releases and those detained to WED.
Chapter 6:
Appendix
Overall the results suggest that inmates referred for detention are more likely to have committed serious violent offences. However, the detention referrals have actually committed fewer offences as compared to those given statutory release. Sexual offenders are more likely to be referred and to be detained than those who have committed other types of offences.Readmission and re-offence rates for those offenders who are detained to the end of their sentence suggest that they are less likely to reoffend within two years of release than those offenders who are released either at their statutory release date or after a period of detention, but prior to the expiration of their sentence. Some of those referred for detention, and in particular those who are detained to the end of their sentence, may actually be lower risk than those who are released. For example, those detained to the end of their sentence are older, have fewer criminal convictions requiring a federal sentence, and are less likely to be readmitted within two years for having committed a new offence.
The results demonstrate that those offenders released at, or after their statutory release date, and therefore deemed to be too high a risk for release on parole are a particularly problematic group. They have only a 50% chance of being successful while under supervision after statutory release, and they are likely to commit a new crime following their release. From this group, potential detainees who are likely to commit a new violent offence, must be selected.
There were 235 offenders kept in prison until the end of their sentence rather than being released into the community to serve the last one-third of their sentence. Of these, approximately 38, or 16% committed a new offence within two years after their eventual release. This rate of reconviction is lower than for those released on statutory release and it is about the same as for those who were referred for detention, but released. It is also lower than those who were released on full parole.
The Introduction presented three questions to be answered by the results from this study:
- Are the right people referred for detention? That is, are the highest risk violent offenders referred for detention?
- Are the right people being detained? That is, are the people detained the highest risk offenders?
- Are those who have been detained the most likely to reoffend after release and are they the most likely to reoffend violently?
The results from the study suggest the following responses:
- It is not clear that the highest risk offenders are being referred for detention. However, the offenders who are referred have committed violent offences. The recidivism data indicate that those referred are no more likely to be readmitted, or to commit a new offence, than many of those released without being referred for detention.
- The results suggest that offenders who are detained are not higher risk than offenders referred and not detained. For example, while detained offenders are more likely to have had a conviction for a sexual offence, they are less likely to have had convictions for robbery and drug offences. In addition, detained offenders are more likely to have been given shorter sentences by the courts and to have fewer previous convictions for which they served their sentence in a federal prison. While some detained offenders have a high risk of re-offending, others are less likely to reoffend than those who are released at their SR date.
- Are those who are detained the most likely to reoffend? The data suggest that offenders who are released at their statutory release date, including those who are referred but not detained, are more likely to be readmitted than those who are detained to the end of their sentence. It also appears that those who are not detained are more likely to reoffend within two years of their release, and are equally likely to reoffend with a violent offence (another scheduled offence) as those who are detained.
Some of the results suggest that lower risk offenders are detained , which is consistent with an early study (Motiuk, Belcourt and Bonta, 1994) which compared offenders using the SIR scale (General Statistical Information on Recidivism), a measure of risk of re-offending. These results suggest that the decisions to refer, and to detain, may be based on the nature of the offence rather than the risk of violent re-offending. This might explain why sexual offenders are more likely to be detained than those convicted of other violent offences.
The results presented in the study raise a number of questions which require investigation. While the study was able to characterize the types of offenders, types of offences and recidivism patterns, it could not deal with the nature of the initial offences nor the nature of subsequent offences. It may be that case management officers, in making a decision for referral, and the National Parole Board, in making a decision to detain, use other criteria that are valid for reasons other than statistical probability. However, the results of the study suggest that the decisions made are not achieving the goals of detaining the most serious offenders. An analysis of the other criteria being used is needed to more fully understand either their applicability in the decision process, or the inaccuracies which they introduce.
The results indicate that there is a need to examine the methods used in selecting cases for detention. Modifications to the current tools and systems used for selection should increase the accuracy of the selection of the highest risk offenders, particularly those most likely to commit a new violent offence after their release. Given the relatively high rate of reoffending and readmission among the group of offenders who remain in institutions until their statutory release date, specialized selection tools and procedures may be needed to take account of the unique characteristics of these offenders.
Given the similarity in re-offence rates between those detained and those released at their statutory release date, it may be more effective to release offenders under supervision and then revoke the conditional release when problems occur. The additional effort required for detention review does not seem to contribute to a reduction in the likelihood of re-offending. With this approach the low risk offenders would be released and the higher risk offenders who demonstrate an inability to adapt to community standards would be re-incarcerated.
While protection of society must remain a prime consideration, there is evidence to suggest that the increased incarceration costs of detention may not be achieving much benefit. If one considers the 235 offenders detained to the end of their sentence and assumes a readmission rate similar to the SR groups of about 50% in two years, then 118 of the detention offenders would return. However, 117 would remain in the community. Given an average detention period of 415 days and $46,000 per year for custody, the cost of maintaining these offenders in prison would be about $6.1million. It would cost about $1.13 million to supervise the offenders in the community. The net savings would be $4.77 million over five years accounted for in this study, or about $1 million per year.
Of concern in this analysis is the type of offence committed by those detained to the end of their sentence and those released at the statutory release date. Of the SR group and those detained to the end of their sentence an equal percentage, about 11%, committed a scheduled offence after release; this amounts to 26 of the 235 detained offenders. The detained offenders were more likely to commit a sexual offence, but less likely to commit a robbery. Given the difference in the size of these two groups (10,000 Vs 235) many more scheduled offences were committed by those released on SR. In addition, the SR group was four times more likely (28% Vs 7%) to commit a new non-violent offence than those detained to the end of their sentence.
A saving of $1 million per year accounts for only about half of those detained until the end of their sentence because the 235 cases were those who had sufficient follow-up period within the study. The potential savings would double if the more than 500 offenders detained until the end of their sentence were included, and would further increase if those detained and then released were included. In addition, the estimate assumes a readmission rate for detained offenders similar to the SR group. However, the data suggest a lower readmission rate for those detained.
Finally, additional research would assist in the development of better selection criteria for those most likely to commit new offences. For example, using objective risk measures might be more effective than the current list of offences specified in the Schedules of the CCRA.
Summary of Findings
- The number of referrals, as a percentage of all releases, has been increasing over the five year period of the study. The number of referrals, as a percentage of all potential statutory releases, has been increasing even more dramatically.
- While those referred are more likely to have longer sentences, 6 years or more, than other inmates released, but within the referral group, those detained to the end of their sentence are less likely to have longer sentences.
- Offenders detained to WED tend to be older than those released on SR, and those referred and subsequently released.
- Female offenders are less likely to be referred for detention, but once referred are equally likely, compared to male inmates, to be detained until the end of their sentence.
- Aboriginal offenders are over-represented in the referral group, but given that they are referred, they appear in the same distribution as Non-Aboriginals across the four sub-groups used to study referral cases (referred and released, detained and then released, detained to the end of their sentence, and other).
- There are large regional differences in the percentage of cases referred. Quebec has the lowest referral rate, while other regions refer more than would be expected given their proportion of the inmate population. Year to year increases in the percentage of releases referred are consistent across all regions.
- Offenders referred for detention have a greater number of scheduled offences than other offenders, and this number has been constant over the five years of the study. This would suggest that the increase in the number of referrals is not the result of the number of convictions offenders have for scheduled offences.
- Drug offences are a relatively small percentage of the total number of scheduled offences and have not been increasing over the five years of the study. The observed increase in the number of referred cases cannot be attributed to the addition of Schedule 2 to the CCRA, nor can it be attributed to a change in the number of drug offences.
- Referral cases are more likely to have been convicted of scheduled sexual and assault offences as well as non-scheduled violent offences than other offenders. They are less likely to have been convicted of robbery offences. Amongst those referred, offenders who were detained were more likely to have been convicted of a sexual offence than those who were released. This suggests that detention referrals are targeting sex offenders.
- Based on the number of previous criminal offences, offenders referred for detention have a lower level of criminal activity than those released on SR, and have a level of criminal activity more similar to those released on full parole. This result suggests that, based on this risk factor, offenders referred for detention may have a lower risk to reoffend than those released on SR.
- There appears to be an increase in the variety of violent offences (more offenders committing different types of violent offences) for both referred offenders and those released on full parole and statutory release over the five year period of the study. This could account for some of the increase in referrals, although the result is inconsistent with that in number 7 above.
- Aboriginal offenders who are referred are less likely to have been convicted of sexual offences and drug offences and are more likely to have been convicted of assault and manslaughter.
- The pattern of readmissions is similar for the SR and the referred groups who were subsequently released before WED. The detained to WED group had the lowest readmission rate, but this may reflect the need for court action prior to readmission, rather than a difference between this group and the others.
- Time to a new offence is shortest for those released on SR and those referred and released. Those who were detained had re-offence rates lower than those released on full parole.
- On average, the incapacitation effect of detention lasts only about one year, the period the offender remains in prison.
- Only 16% of those detained to the end of their sentence, and 12% of those detained and then released, had a new offence of any type two years after their release. Only 11% committed a scheduled offence as their first new offence, while 15% did commit a new scheduled offence sometime after their release (given a minimum of two years following release).
- The readmission and re-offence rate was lowest for those over 35 years of age. Given that the average age of the referral group is 32, and the average age of those detained to WED is 33, it is likely that some of these offenders have a lower risk of re-offending than would be expected given their offences.
- Aboriginal offenders have higher readmission rates than Non-aboriginal offenders, but these differences are reduced when new offence is used as the recidivism measure. These results suggest that Aboriginal offenders may be more likely to have their conditional release revoked, even though they are not more likely to commit a new offence. Additional investigations may be needed to explain the observed differences.
- Regional differences are evident for both readmission rates and new offences within two years of release. Patterns of readmissions across referral groups and regions are consistent with the general patterns observed. Those detained have lower re-admission and re-offence rates. Striking is the relatively higher re-offence rate in the Atlantic region for full parole releases and those detained to WED.
Appendix
Corrections and Conditional Release Act [1992, c. 20]
Schedule I
(Subsections 107(1), 125(1) and 126(1) and sections 129 and 130)
-
1. An offence under any of the following provisions of the Criminal
Code:
- (a) paragraph 81(2)(a) (causing injury with intent);
- (b) section 85 (use of firearm during commission of offence);
- (c) subsection 86(1) (pointing a firearm);
- (d) section 144 (prison breach);
- (e) section 151 (sexual interference);
- (f) section 152 (invitation to sexual touching);
- (g) section 153 (sexual exploitation);
- (h) section 155 (incest);
- (i) section 159 (anal intercourse);
- (j) section 160 (bestiality, compelling, in presence of or by child);
- (k) section 170 (parent or guardian procuring sexual activity by child);
- (l) section 171 (householder permitting sexual activity by or in presence of child);
- (m) section 172 (corrupting children);
- (n) subsection 212(2) (living off the avails of prostitution by a child);
- (o) subsection 212(4) (obtaining sexual services of a child);
- (p) section 236 (manslaughter);
- (q) section 239 (attempt to commit murder);
- (r) section 244 (causing bodily harm with intent);
- (s) section 246 (overcoming resistance to commission of offence);
- (t) section 266 (assault);
- (u) section 267 (assault with a weapon or causing bodily harm);
- (v) section 268 (aggravated assault);
- (w) section 269 (unlawfully causing bodily harm);
- (x) section 270 (assaulting a peace officer);
- (y) section 271 (sexual assault);
-
(z) section 272 (sexual assault with a weapon, threats to a
third party or causing bodily harm);
- (z.1) section 273 (aggravated sexual assault);
- (z.2) section 279 (kidnapping);
- (z.3) section 344 (robbery);
- (z.4) section 433 (arson - disregard for human life);
- (z.5) section 434.1 (arson - own property);
- (z.6) section 436 (arson by negligence); and
- (z.7) paragraph 465(1)(a) (conspiracy to commit murder).
-
2. An offence under any of the following provisions of the Criminal
Code, as they read immediately before July 1, 1990:
- (a) section 433 (arson);
- (b) section 434 (setting fire to other substance); and
- (c) section 436 (setting fire by negligence).
-
3. An offence under any of the following provisions of the Criminal
Code, chapter C-34 of the Revised Statutes of Canada, 1970, as they
read immediately before January 4, 1983:
- (a) section 144 (rape);
- (b) section 145 (attempt to commit rape);
- (c) section 149 (indecent assault on female);
- (d) section 156 (indecent assault on male);
- (e) section 245 (common assault); and
- (f) section 246 (assault with intent).
SCHEDULE II
(Subsections 107(1) and 125(1) and sections 129, 130 and 132)
1. An offence under any of the following provisions of the Narcotic Control Act:
- (a) section 4 (trafficking);
- (b) section 5 (importing and exporting);
- (c) section 6 (cultivation);
- (d) section 19.1 (possession of property obtained by certain offences); and
- (e) section 19.2 (laundering proceeds of certain offences).
2. An offence under any of the following provisions of the Food and Drugs Act:
- (a) section 39 (trafficking in controlled drug);
- (b) section 44.2 (possession of property obtained by trafficking in controlled drug);
- (c) section 44.3 (laundering proceeds of trafficking in controlled drug);
- (d) section 48 (trafficking in restricted drug);
- (e) section 50.2 (possession of property obtained by trafficking in restricted drug); and
- (f) section 50.3 (laundering proceeds of trafficking in restricted drug).
