FORUM on Corrections Research

Detention: Is it meeting its goal?

At their statutory release date, federal inmates in Canada who are not released on full parole are generally released to serve the final one third of their sentence under supervision in the community. Some offenders, however, are deemed too dangerous for release and are detained in custody (known as detention) for a period of time beyond their statutory release date.

A provision of the Corrections and Conditional Release Act (1992) authorizes the detention² until the end of their sentence of inmates who are likely to commit a new violent offence or a serious drug offence. The main purpose of this provision is to increase the protection of society by maintaining the most dangerous offenders in custody. Detention cannot, however, extend past the end of an inmate's sentence. Since evidence suggests that longer sentences do not decrease recidivism,³ the main benefit of detention is to keep the offender incarcerated until the end of the sentence.

Who's referred?

Predicting, with a high degree of accuracy, who is likely to commit a violent offence is an extremely difficult task. However, case management officers and the National Parole Board have been required, by law, to make such predictions for the purpose of detention. These predictions must be made from the information they have available before the statutory release (SR) date of the offender.

The Corrections and Conditional Release Act (1992) contains very specific procedures which must be followed when an offender is detained in custody beyond the statutory release date. These procedures suggest that lawmakers wanted to ensure that detention was reserved only for the most serious offenders, as detention increases the limits on freedom imposed by the courts.

The Correctional Service of Canada refers offenders to the National Parole Board for detention consideration. The cases of inmates who have committed violent offences or drug offences are automatically reviewed to determine if detention should be considered.⁴ Other offenders, who may represent a threat to society on release, may also be reviewed. Referrals to the National Parole Board are made if it is determined that an offender, if released, is likely to commit another violent or serious drug offence before the end of their sentence. The National Parole Board decides whether or not to detain an inmate, after which annual reviews are held to determine if the offender should be released.

This article reviews the results of a recent study⁵ which addresses two fundamental questions: are the highest risk offenders referred and detained, and are the offenders who are referred and detained the most likely to recidivate?

To answer the first question, the study investigated criminal history factors. Time to readmission and time to a new offence determined if the detained offenders were the most likely to recidivate.

For offenders under supervision, readmission is a very useful measure of recidivism since an offender can be readmitted for a new offence and/or for failing to meet the conditions of the release. Additionally, the readmission can occur very quickly, making it a much more sensitive measure of recidivism than a new offence. However, as there is a considerable time lag associated with court proceedings, the readmission rates for offenders under supervision cannot be compared to those for offenders who have completed their sentences. Therefore, a second measure of recidivism was used: time to new offence or sentence. Since all offenders, whether under supervision or past the end of their sentence, are processed by the court in the same manner, the time to a new sentence is comparable for offenders being supervised and offenders who have completed their sentence.

Comparisons in the study are made between the following groups:

• offenders referred for detention but released without being detained (referred but released);
• offenders detained for a period of time but released before the end of their sentence (detained then released);
• offenders detained to the end of their sentence (detained to end of sentence);
• offenders released at the statutory release date with no previous releases
  (SR only);
• offenders released at the statutory release date, but who had a previous release (SR and other); and
• offenders released on full parole (FP).

Number of referrals

The study used all releases from custody and all referrals for detention from fiscal years 1989-90 to 1993-94. During this period, there were 1,115 referrals for detention and 917 offenders (82%) were detained for some period after the statutory release date. Follow-up data were available for 862 of the referred offenders who reached the end of their sentence within the period of the study. Of these, 62% were detained to the end of their sentence, 15% were detained and then released, and 23% were not detained.

Table 1 shows the eventual outcome for all the referrals and the percentage of cases with a two-year follow-up. Detention cases in which the sentence was not complete (column D) are not included in the analyses which follow. The two-year follow-up cases are those for which readmission and reconviction data are available.

Table 1

Release Outcome for Referral Cases
	A Refferred but released	B Detained then released	C Detained to end of sentence	D Detained but sentence incomplete*	Total
Total cases	198	131	533	253	1,115
With two-year folllow-up	164	92	235	2	493
Percentage with two-year follow-up	83	70	44	1	44
Notes:* includes those cases still detained at the conclusion of the study period and a few released and subsequently detained

It is interesting to note that the number of offenders referred for detention has been increasing from year to year. Since only offenders who are held in custody until their statutory release date are considered for detention, this increase was calculated as a percentage of the total number of statutory release cases for each year. The results, presented in Figure 1, show a steady increase in the percentage of referrals, from 4% in 1989-90 to 14% in 1993-94.

Figure 1

Risk factors for referred offenders

To determine if the offenders being referred for detention were the highest risk offenders, three factors normally associated with higher risk behaviour were studied: the level of criminal activity, the type of offence committed and the length of the sentence. The results for referral cases are then compared with those for other types of releases. In addition, the referrals are compared across three possible outcomes: released, detained and then released, and detained to end of sentence.

Of the offenders released at the SR date, over 50% had six or more offences for which they served time in a federal prison. However, only 35% of the referrals were offenders who had six or more offences (see Table 2).

Table 2

Level of Criminal Activity (Number of Offences) by Type of Release (Percentage)
Type of release					Distribution of refferal cases
Number of previous federal offences	FP	SR and other	SR only	Referral	Referred but released	Detained then released	Detained to end of sentence	Detained but sentence in- complete*
One or two	41	16	22	31	31	30	32	30
Three to five	27	26	28	34	30	33	36	34
Six or more	32	58	49	35	38	37	33	36
Number of cases	10, 624	6, 207	5, 654	1,115	198	131	533	253
Notes: FP full parole; SR and other - statutory release, but with previous day parole or full parole; SR only-statutory release without a previous release on the current sentence; Refferal - reffered for detention: * includes those cases still detained at the conclusion of the study period and a few released and subsequently detained.

Offenders referred for detention were more likely to be serving longer sentences than offenders not referred (see Table 3). More specifically, 37% of referred offenders were serving sentences of more than six years; only 20% of offenders released on full parole or statutory release were serving sentences of this length. Of all offenders detained to the end of their sentence, 27% were serving sentences of six years or more, while almost half (49%) were serving sentences of four years or less.

Table 3

Level of Criminal Activity (Number of Offences) by Type of Release (Percentage)
Type of release					Distribution of refferal cases
Number of previous federal offences	FP	SR and other	SR only	Referral	Referred but released	Detained then released	Detained to end of sentence	Detained but sentence incomplete*
Four years or less	58	58	60	38	43	25	49	20
More than four years to six years	21	18	20	25	21	30	24	27
More than six years	27	24	20	37	36	45	27	53
Number of cases	10,624	6,207	5,654	1,115	198	131	533	253
Notes: FP full parole; SR and other - statutory release, but with previous day parole or full parole; SR only-statutory release without a previous release on the current sentence; Refferal - reffered for detention: * includes those cases still detained at the conclusion of the study period and a few released and subsequently detained.

The types of offences for which offenders were serving sentences are presented in Table 4. Referred offenders were more likely to be serving a sentence that included a sex offence (59%) than the offenders not referred (19% or less). Offenders referred but released were less likely to have committed a sexual offence than offenders who were detained (44% versus 60%).

Table 4

Offenders with at least One Previous Conviction (Percentage)
Type of release					Distribution of refferal cases
Number of previous federal offences	FP	SR and other	SR only	Referral	Referred but released	Detained then released	Detained to end of sentence	Detained but sentence incomplete*
Sex	10	10	19	59	44	60	62	64
Assault	13	24	28	42	45	38	42	41
Robbery	24	42	38	22	25	24	18	24
Manslaughter	5	3	3	9	8	12	9	7
Other	9	12	10	6	9	5	5	4
Drug	28	13	11	3	4	5	1	4
Number of Cases	10, 624	6, 207	5, 654	1,115	198	131	533	253
Notes: FP full parole; SR and other - statutory release, but with previous day parole or full parole; SR only-statutory release without a previous release on the current sentence; Refferal - reffered for detention: * includes those cases still detained at the conclusion of the study period and a few released and subsequently detained.

Approximately 25% of the offenders released at their statutory release date had committed an assault. This compares with 42% of offenders referred for detention and 42% of detained offenders.

Offenders charged with robbery and drug offences were less likely to be referred, given the large proportion of offenders with these convictions. Approximately 40% of offenders released at their statutory release date were convicted of robbery while only 22% of offenders referred for detention had been convicted of robbery. Finally, while about 12% of offenders released at their statutory release date were serving sentences for drug offences, only 3% of the referrals were serving sentences for a similar offence. However, it must be remembered that only offenders likely to commit a serious drug offence after release may be detained, and many of the offenders charged with drug offences would not meet this criteria.

Regional differences

Figure 2 compares the number of referrals to the number of offenders who would have been released at their statutory release date in each region. The results indicate that this percentage has been increasing in all regions. The smallest increase has been in Quebec where only about 6% of offenders to be released on statutory release were referred. However, in most other regions, the percentage has risen from approximately 5% in 1989-90 to over 15% in 1993-94.

Figure 2

Aboriginal offenders

During the five years included in the study, approximately 25% of offenders referred for detention
were Aboriginal offenders. Since Aboriginal offenders account for about 11% of
the offender population, they seem to be overrepresented in the referral group. Additional analyses are needed to determine the reason for this difference.

Gender differences

Only 10 female offenders were referred for detention during the five years of the study. This represents about 1.6% of female offender releases during the period, which is considerably below the 4.8% of males referred for detention. Five of the 10 referrals were detained to the end their sentence. Frequencies in the other referral categories are too low to report.

Readmission

Readmission occurs either as a result of a revocation of a conditional release, commission of a new offence or a combination of the two. Revocation of a conditional release can only occur during a period of supervision, before the end of the sentence.

Figure 3 shows the percentage of offenders readmitted within two years of release. Three groups (the referred but released, detained then released and statutory release groups) have readmission rates which are relatively similar, between 43% and 49%. The parole group has a significantly lower readmission rate of about 24%. These results suggest that offenders referred for detention but released, and offenders detained then released, are as likely to be returned to federal custody as are offenders released at their statutory release date. If referred offenders were truly the most serious offenders, it would be expected that they would return at a higher rate than those released at the statutory release date.

Figure 3

New offence

New convictions that resulted in federal prison sentences were also used to measure recidivism. Offenders were followed for 24 months after their release from prison to determine if they had been convicted of a new offence.

Figure 4 shows the time to a new conviction⁶ for the six study groups within two years of release. Of the offenders released at their statutory release date, about 35% were sentenced for a new offence. Approximately 28% of offenders who were referred but released were convicted of a new offence within two years of their release. Of the offenders who were detained (either detained to the end of their sentence, or detained and then released), less than 20% were convicted of a new offence within the two-year period after their release. This rate was lower than for offenders released on full parole.

Figure 4

Overall, the offenders most likely to be convicted of a new offence within two years of their release were those released at their statutory release date without a referral for detention. Offenders who were detained were about as likely, or less likely, to be sentenced for a new offence as offenders released on full parole.

Discussion

While the number and rate of referrals and detained offenders have been increasing, it is not clear whether the highest risk offenders, those most likely to commit a new offence, are being identified by the current selection process.

The results indicate that offenders referred for detention are more likely to have longer sentences, but less criminal activity in their past. Referred offenders are more likely to have been convicted of a sexual offence or an assault. Offenders who are detained to the end of their sentence have a lower number of previous federal sentences, are more likely to be serving a sentence of four years or less and are most likely to have been convicted of a sexual offence or an assault when compared to offenders released at their statutory release date.

The recidivism results seem to support the observation that the offenders who are detained are not the highest risk. In terms of readmission, those referred for detention seem to be returned to prison at almost the same rate as offenders released at their statutory release date.

Results for new offences indicate that offenders released at their statutory release date are more likely to be convicted of a new offence than are offenders who were referred. In fact, detained offenders seem to be committing new offences at about the same rate as those released on full parole. While it is tempting to conclude that the increased incarceration associated with the detention was responsible for the reduction in recidivism, other research⁷ has shown this to be false.

The purpose of the detention provision of the Corrections and Conditional Release Act was to increase public safety by keeping a small number of high risk offenders in prison longer. It could be argued that this goal is being achieved: some offenders are kept in prison longer, on average 415 days, and 16% of these offenders would likely have committed a new offence within two years of being released if they had been in the community. However, approximately 35% of offenders released at their statutory release date commit a new offence within two years of release, and 18% commit a violent or drug offence. Therefore, it would appear that the selection process for detention has not resulted in the highest risk offenders being detained.

It is obvious that, in exceptional cases, offenders should be detained in custody after their statutory release date to ensure public safety. However, more work is needed to identify these offenders successfully and ensure that only a small number are detained. Alternatively, it might be argued that it is impossible to predict who is most likely to commit a serious offence beyond what we are currently capable of, given the high risk characteristics of offenders held in custody until their statutory release date. If this is the case, then detention, for all but the most serious cases, may not achieve its intended goal.

---

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

2. Detention was first introduced in 1986.

3. L.L. Motiuk, R.L Belcourt and J. Bonta, Managing High Risk Offenders: A Post-detention Follow-up, Research Report R-39
(Ottawa: Correctional Service of Canada, 1995).

4. Violent offences are specified in Schedule 1 of the Act and drug offences are specified in Schedule 2. These schedules are like appendices to the Act and simply contain lists of offences.

5. Additional details and analyses may be found in the report by
B.A. Grant, Inmates Referred for Detention (1989-90 to 1993-94): A Comparative Analysis, Research Report R-45 (Ottawa: Correctional Service of Canada, 1996).

6. The date of sentencing was used as an approximation of the date of conviction since the conviction date was not available in the database.

7. Motiuk, Belcourt and Bonta, Managing High Risk Offenders: A Post-detention Follow-up.