Correctional program evaluation : offenders placed on probation or released from prison in fiscal year ...

              North Carolina
Sentencing and Policy Advisory Commission
Prepared By
Karen Calhoun Vicky Etheridge
Tamara Flinchum Ashleigh Gallagher, Ph.D.
Ginny Hevener David Lagos
Susan Katzenelson
Project Conducted in Conjunction with the
North Carolina Department of Correction
Submitted Pursuant to Session Law 1998-212, Section 16.18
April 15, 2010
CORRECTIONAL PROGRAM
EVALUATION:
OFFENDERS PLACED ON PROBATION
OR RELEASED FROM PRISON
IN FISCAL YEAR 2005/06
Project Conducted in Conjunction with the
North Carolina Department of Correction
North Carolina Sentencing and Policy Advisory Commission
P.O. Box 2472
Raleigh, NC 27602
(919) 890-1470
www.nccourts.org/courts/crs/councils/spac
The Honorable W. Erwin Spainhour Susan Katzenelson
Chairman Executive Director
CORRECTIONAL PROGRAM
EVALUATION:
OFFENDERS PLACED ON PROBATION
OR RELEASED FROM PRISON
IN FISCAL YEAR 2005/06
NC SENTENCING AND POLICY ADVISORY COMMISSION
MEMBERSHIP
Hon. W. Erwin Spainhour, Chairman
Superior Court Judge
Thomas Bennett
NC Victim Assistance Network
Hon. Stan Bingham
State Senator
Hon. Alice Bordsen
State Representative
Hon. Charles E. Brown
NC District Court Judges’ Association
Joseph B. Cheshire V
NC Advocates for Justice
Locke T. Clifford
NC Bar Association
Louise Davis
NC Community Sentencing Association
Judge Richard A. Elmore
NC Court of Appeals
Judge Robert C. Ervin
NC Conference of Superior Court Judges
Hon. Garry W. Frank
NC Conference of District Attorneys
Hon. Paul H. Gibson
NC Association of County Commissioners
William P. Hart
NC Attorney General’s Office
Secretary Linda W. Hayes
NC Department of Juvenile Justice and
Delinquency Prevention
Mary Y. “Larry” Hines
Private Citizen, Governor’s Appointee
Sheriff Shepard Jones
NC Sheriffs’ Association
Secretary Alvin W. Keller, Jr.
NC Department of Correction
Hon. Eleanor Kinnaird
State Senator
Hon. Jimmy L. Love, Sr.
State Representative
Hon. Floyd B. McKissick, Jr.
State Senator
Moe McKnight
NC Retail Merchants’ Association
Dr. Harvey Lee McMurray
Academic Member
Luther T. Moore
Lieutenant Governor’s Appointee
Judge Fred G. Morrison, Jr.
Justice Fellowship
Chief Frank Palombo
NC Association of Chiefs of Police
Anthony Rand
NC Post-Release Supervision & Parole
Commission
Rhonda Raney
NC Dept. of Crime Control & Public Safety
June Ray
NC Association of Clerks of Superior
Court
Billy J. Sanders
Commission Chairman’s Appointee
Hon. Timothy L. Spear
State Representative
NC SENTENCING AND POLICY ADVISORY COMMISSION
STAFF
Susan Katzenelson
Executive Director
John Madler
Associate Director for Policy, Staff Attorney
Ginny Hevener
Associate Director for Research
Karen Calhoun Tamara Flinchum
Senior Research & Policy Associate Senior Research & Policy Associate
Vacant David Lagos
Senior Research & Policy Associate Research & Policy Associate
Ashleigh Gallagher, Ph.D. Sara Perdue
Research & Policy Associate Research & Policy Associate
Vicky Etheridge
Administrative Assistant
TABLE OF CONTENTS
EXECUTIVE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i
I. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Defining Recidivism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Structured Sentencing and Recidivism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Comparison of Recidivism Rates for Previous Recidivism Studies . . . . . . . . . . . . . . . . . 4
Table 1.1: Rearrest Rates for North Carolina Offenders . . . . . . . . . . . . . . . . . . . 5
Research Design and Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Follow-up Period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Time at Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Outcome and Process Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Data Sources and Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Report Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
II. STATISTICAL PROFILE OF THE FY 2005/06 SAMPLE . . . . . . . . . . . . . .. . . . . . . . . . . . .9
Type of Punishment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Figure 2.1: Type of Punishment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Personal Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Criminal History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Table 2.1: Personal Characteristics by Type of Punishment . . . . . . . . . . . . . . . 12
Table 2.2: Prior Arrests by Type of Punishment . . . . . . . . . . . . . . . . . . . . . . . 13
Most Serious Current Conviction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Figure 2.2: Most Serious Current Conviction: Felonies Only (n=29,304) . . . . 14
Table 2.3: Most Serious Current Conviction by Type of Punishment . . . . . . . . 15
Figure 2.3: Most Serious Current Conviction: Misdemeanors Only (n=31,520) 16
Table 2.4: Offense Class for Most Serious Current Conviction by Type of
Punishment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Type of Prison Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Table 2.5: Offense Class of Most Serious Conviction for Prison Releases by Type
Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Offender Risk and Recidivism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Figure 2.4: Offender Risk Level by Type of Punishment . . . . . . . . . . . . . . . . . . 21
Table 2.6: Offender Risk Level for Prison Releases by Type of Entry . . . . . . . . 22
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
III. CRIMINAL JUSTICE OUTCOME MEASURES FOR THE FY 2005/06 SAMPLE . . . 23
Definition of the Follow-up Period and Time at Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Table 3.1: Percent at Risk and Average Time at Risk by Type of Punishment . . 24
Criminal Justice Outcome Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Table 3.2: Rearrest Rates by Type of Punishment . . . . . . . . . . . . . . . . . . . . . . . 25
Figure 3.1: Number of Rearrests for Offenders with Any Rearrest . . . . . . . . . . 26
Table 3.3: Rearrests Rates by Risk Level for Prison Releases by Type of Entry 27
Table 3.4: Rearrests by Type of Punishment and Crime Type . . . . . . . . . . . . . . 28
Figure 3.2: Rearrest Rates by Type of Offender Risk Level: Three-Year
Follow-Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Figure 3.3: Rearrest Rates by Type of Punishment and Risk Level: Three-Year
Follow-Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Table 3.5: Rearrests by Risk Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Table 3.6: Offender Risk Level and Rearrest Rates by Offense Class for Most
Serious Current Conviction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Table 3.7: Reconviction Rates by Type of Punishment . . . . . . . . . . . . . . . . . . . . 35
Table 3.8: Reconvictions by Type of Punishment and Crime Type . . . . . . . . . . . 36
Table 3.9: Reincarceration Rates by Type of Punishment . . . . . . . . . . . . . . . . . 38
Interim Outcome Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Table 3.10: Technical Revocation Rates by Type of Punishment . . . . . . . . . . . 40
Figure 3.4: Number of Infractions During Incarceration . . . . . . . . . . . . . . . . . 41
Table 3.11: Average Number of Infractions During Incarceration . . . . .. . . . . . 41
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Figure 3.5: Rearrest Rates: Three-Year Follow-Up . . . . . . . . . . . . . . . . . . . . . . 42
Figure 3.6: Criminal Justice Outcomes by Type of Punishment: Three-Year
Follow-Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Figure 3.7: Criminal Justice Outcomes by Risk Level: Three-Year Follow-Up 44
IV. MULTIVARIATE ANALYSIS OF OUTCOME MEASURES . . . . . . . . . . . . . . . . . . . . 45
Multivariate Analysis: What is a Regression Model? . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Regression Analysis: Recidivist Arrest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Table 4.1: Effect of Personal and Criminal Justice Factors on Recidivist Arrest 48
Regression Analysis: Recidivist Incarceration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Table 4.2: Effect of Personal and Criminal Justice Factors on Recidivist
Incarceration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
V. MENTAL HEALTH PROBLEMS AND PRISON RELEASES . . . . . . . . . . . . . . . . . . . . . 58
Definition of Mental Health Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Figure 5.1: Mental Health and Prisoners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Personal Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Table 5.1: Personal Characteristics by Mental Health Status of Prisoners . . . . .61
Most Serious Current Conviction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Figure 5.2: Current Offense Class by Mental Health Status . . . . . . . . . . . . . . . . 60
Figure 5.3: Average Months Served by Offense Class and Mental Health Status 62
Figure 5.4: Offender Risk Level by Mental Health Status . . . . . . . . . . . . . . . . . . 62
Infractions While Incarcerated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Figure 5.5: Number of Infractions during Incarceration by Mental Health
Status, All Prisoners (n=19,733) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Table 5.2: Average Number of Infractions During Incarceration by Mental Health
Status (For Prisoners with Infractions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Criminal Justice Outcome Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Table 5.3: Criminal Justice Outcome Measures by Mental Health Status During
the Three-Year Follow-Up Period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Table 5.4: Rearrest Rates by Mental Health Status and Type of Release During
the Three-Year Follow-Up Period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Table 5.5: Rearrest Rates by Mental Health Status and Offender Risk Level
During the Three-Year Follow-Up Period . . . . . . . . . . . . . . . . . . . . . . . . . 65
Rates of DSM-IV Diagnosed Disorders among the Entire Prison Sample . . . . . . . . . . . . 65
Figure 5.6: Rates of Diagnosed Mental Health Illnesses by Gender (n=19,733) 66
DSM-IV Diagnoses by Diagnostic Category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Table 5.6: Mental Health Diagnoses by Diagnostic Category: Prisoners with a
Mental Health Diagnosis (n=4,294) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Figure 5.7: Definitions of the Most Common Diagnosis Categories in the
FY 2005/06 Prison Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Figure 5.8: Rearrest Rates by DSM-IV Diagnosis Category: Three-Year
Follow-Up, Prisoners with a Mental Health Diagnosis (n=4,294) . . . . . . 70
The Co-Occurrence of Substance Abuse and Mental Health Problems . . . . . . . . . . . . . . 70
Figure 5.9: Prevalence of Substance Abuse and Mental Health Problems among
Prisoners (n=19,733) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Table 5.7: Rearrest Rates by Mental Health and Substance Abuse Problems . . . 71
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
VI. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
APPENDIX A: GLOSSARY OF MAJOR VARIABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
APPENDIX B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Appendix B-1: Individual Program and Correctional Supervision Summaries . . . . . . . . 89
Appendix B-2: Summary Information for Correctional Programs . . . . . . . . . . . . . . . . . 125
APPENDIX C: PUNISHMENT CHARTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126
APPENDIX D: MEASURING OFFENDER RISK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Prediction of Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Measuring Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Figure D.1: Variables Included in Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
APPENDIX E: MULTIVARIATE TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Table E.1: Multiple Regression Results of Personal and Criminal Justice Factors
on Prison Infractions: Prisoners FY 2005/06 (n=19,733) . . . . . . . . . . . 134
Table E.2: Effect of Personal and Criminal Justice Factors on Technical
Revocation: Probationers FY 2005/06 (n=41,091) . . . . . . . . . . . . . . . . . 135
i
EXECUTIVE SUMMARY
Introduction
In 1998, the North Carolina General Assembly directed the Sentencing and Policy
Advisory Commission to prepare biennial reports evaluating the effectiveness of the State’s
correctional programs (Session Law 1998-212, Section 16.18). This study constitutes the sixth
report in compliance with the directive and analyzes a sample of 60,824 offenders released from
prison or placed on probation in FY 2005/06 using a three-year follow-up period. Only offenders
sentenced under the Structured Sentencing Act (SSA) were included in the sample. The study
defines recidivism as rearrest, reconviction, and reincarceration. In addition, two interim outcome
measures were examined: 1) technical revocation of probation or post-release supervision for
offenders supervised in the community and 2) prison infractions during incarceration for prisoners
in the sample. This report also focuses on prisoners identified with a mental health problem.
Data Sources
Data for offenders in the sample were provided by the Department of Correction (DOC)
and the Department of Justice (DOJ).
Statistical Profile of the FY 2005/06 Sample
The sample of 60,824 offenders included 47.1% community probationers, 20.5%
intermediate probationers, 29.1% prisoners with no post-release supervision (no PRS), and 3.3%
prisoners with post-release supervision (PRS), all placed on probation or released from prison
during FY 2005/06. Overall, there were 41,091 probation entries and 19,733 prison releases in the
FY 2005/06 sample. Almost 79% of the offenders were male, 51.2% were black, 13.8% were
married, 39.4% had twelve or more years
of education, and 39.6% were identified
as having a substance abuse problem by
either a prison or probation assessment.
Their average age was 31.
Overall, 55.9% of community
punishment probationers, 80.7% of
intermediate punishment probationers,
90.5% of prison releases with no PRS,
and 87.9% of prison releases with PRS
had one or more prior arrests. As a
whole, the FY 2005/06 sample accounted
for a total of 182,885 prior arrests. For
offenders with prior arrests, the number
of prior arrests generally increased by
type of punishment from community
punishment to intermediate punishment
Figure 1
Offense Class for Most Serious Current Conviction
64.4
25.6
51.8
34.4
64.3
44.1
1.2 10.2 4.1
0%
20%
40%
60%
80%
100%
Probation Entries Prison Releases Entire Sample
Misdemeanor F-I Felony B1-E Felony
SOURCE: NC Sentencing and Policy Advisory Commission,
FY 2005/06 Correctional Program Evaluation Data
Note: Offenders with discrepant or unknown offense classes were
excluded from this table (n=317). Percentages may not add to 100%
due to rounding.
ii
Figure 2
Offender Risk Level by Type of Punishment
55.3
37.2
23.6 30.2
40.2
52.3
56.5
57.3
4.5 10.5
19.9 12.5
0%
20%
40%
60%
80%
100%
Community
Punishment
Intermediate
Punishment
Prison
Releases No
PRS
Prison
Releases PRS
% Low Risk % Medium Risk % High Risk
SOURCE: NC Sentencing and Policy Advisory Commission,
FY 2005/06 Correctional Program Evaluation Data
to prison – 43.5% of community punishment probationers, 25.2% of intermediate punishment
probationers, 14.9% of prison releases with no PRS, and 20.0% of prison releases with PRS had
only one prior arrest. A similar pattern was found when comparing the average number of prior
arrests, with probationers having an average of 3.3 prior arrests and prisoners having an average of
5.4 prior arrests.
For the FY 2005/06 sample, 4.1% had a most serious conviction for a Class B1 through
Class E felony, 44.1% had a most serious conviction for a Class F through Class I felony, and
51.8% had a most serious conviction for a misdemeanor offense (see Figure 1). The majority of
probationers (64.4%) had a most serious conviction for a misdemeanor offense while the majority
of prisoners (64.3%) had a most serious conviction for a Class F through Class I felony.
Differences found in the offense class composition of the sample subgroups are consistent with
Structured Sentencing, which links offense severity with type of punishment. For prisoners and
probationers with a current felony conviction (48% of the sample), the majority had convictions
for drug offenses, followed by convictions for property offenses. As anticipated, prisoners were
more likely to have a current conviction for violent offenses (22%) than probationers (12%).
Focusing on the 19,733 prison releases in the sample, for the first time in the Sentencing
Commission’s recidivism studies information was available about the reason for the prisoner’s
entry to prison – whether the prisoner entered prison for a conviction for a new crime, for a
technical revocation of probation, or for both a new crime and a revocation. Forty percent entered
prison as a result of a sentence for a new crime; 37% entered prison for a technical revocation of
their probation; and the remaining 23% entered prison for a mix of both a new crime and a
technical revocation. While the majority of offenders in all three groups were convicted for
felonies in Offense Classes F through I, the proportion of convictions for violent felonies (Class
B1 through Class E) was highest for prisoners admitted for a new crime (20.4%), and the
proportion of convictions for a misdemeanor was highest for prisoners revoked to prison (30.2%).
A risk score was computed for each
offender in the sample using a composite
measure based on individual
characteristics (e.g., social factors and
criminal record factors) identified in the
literature as increasing or decreasing an
offender’s risk of recidivating. As shown
in Figure 2, probationers sentenced to a
community punishment were much more
likely to be low risk than intermediate
punishment probationers and either
category of prison releases. Overall,
prisoners were more likely to be high risk
than probationers. Looking at prison
releases, those with no PRS were more
likely to be high risk and less likely to be
low risk than those prisoners who received
PRS. Risk levels were largely a reflection
of an offender’s criminal history and were
iii
in line with the philosophy of Structured Sentencing, assigning increasingly restrictive sanctions
for the more serious, recidivism-prone offenders.
Time at Risk
While each offender was followed for a fixed three-year period to determine whether
recidivism occurred, the same “window of opportunity” to reoffend was not necessarily available
for each offender due to periods of incarceration during follow-up. This report takes into account
each offender’s actual time at risk (i.e., their actual window of opportunity to recidivate) by
identifying their periods of incarceration in North Carolina’s prison system and subtracting the
time incarcerated from the follow-up period. The percent of the sample at risk for the entire
follow-up period decreased from 87% in the first year to 70% by the third year.
Criminal Justice Outcome Measures
Of the FY 2005/06 sample,
21.4% were rearrested during the
one-year follow-up, 32.5% were
rearrested during the two-year
follow-up, and 40.1% were
rearrested during the three-year
follow-up (see Figure 3). It should
be noted, however, that these
recidivism rates do not take into
account the fact that some offenders
were not at risk for the entire
follow-up period as a result of
incarceration.
Community punishment
probationers, intermediate
punishment probationers, and
prisoners with PRS were subject to
supervision in the community
during the three-year follow-up period. For these three groups, the timing of rearrest was
examined to determine whether the offender was rearrested while on probation or PRS for the
conviction that placed him/her in the FY 2005/06 sample. Of those rearrested during follow-up,
51.4% of community punishment probationers, 57.0% of intermediate punishment probationers,
and 35.4% of prisoners with PRS were rearrested while on probation or PRS. The higher
percentage of intermediate punishment probationers with a rearrest while on supervision compared
to the other two groups is consistent with their longer term of supervision.
Tracking the sample for three years, a clear pattern emerged: while the rates of rearrest
increased for both prisoners and probationers between the first and the third year, the highest rates
of rearrest for all groups were in the first year. In each subsequent year, rearrests increased at a
declining rate.
SOURCE: NC Sentencing and Policy Advisory Commission,
FY 2005/06 Correctional Program Evaluation Data
Figure 3
Rearrest Rate: Three-Year Follow-Up
21.4
18.8
26.9
32.5
28.2
41.3 40.1
50.0
35.4
0%
10%
20%
30%
40%
50%
60%
Probationers Prisoners Entire Sample
1-Year Follow-Up 2-Year Follow-Up 3-Year Follow-Up
iv
For those probationers and prisoners rearrested during the three years, the average time to
first rearrest was 13.2 months after entry to probation or release from prison. By the end of the
three-year follow-up, the FY 2005/06 sample accounted for 50,997 recidivist arrests, including
10,500 arrests for violent offenses.
In addition to rearrest rates, two other criminal justice outcome measures (reconviction and
reincarceration) were utilized. A summary of these three measures of recidivism for the FY
2005/06 sample is provided in Figure 4.
Overall, 9.3% of the sample had a reconviction in the first year of follow-up, 19.3% by the
second year, and 26.6% by the third year. For those with a reconviction during the three-year
follow-up, the average time to reconviction was 17.3 months. The sample accrued 23,045
recidivist convictions of which 3,682 reconvictions were for a violent offense.
Overall, 12.6% of the sample were reincarcerated by the first year, 23.6% by the second
year, and 30.1% by the third year of follow-up. The average time to first incarceration for
offenders reincarcerated during the follow-up period was 15.6 months.
Independent of the measure used or the number of years tracked, recidivism rates were in
direct correlation with the type
of punishment (see Figure 4).
However, it must be noted that
these groups were also
composed of offenders who
were very different in their
potential to reoffend, based on
a composite risk measure
developed for the study (see
Figure 2).
The lowest rearrest and
reconviction rates were for
community probationers,
followed by intermediate
probationers, with the highest
rearrest and reconviction rates
for prison releases with no
PRS. The rearrest and
conviction rates of prisoners with PRS fell between intermediate probationers and prisoners with
no PRS. Compared to the other types of punishment, probationers with an intermediate
punishment had the highest rate of reincarceration, 46.0% during the three-year follow-up period,
due in large part to their higher technical revocation rates.
Figure 4
Criminal Justice Outcomes by Type of Punishment:
Three-Year Follow-Up
32.8
41.2
50.5
45.3
20.9
27.1
35.2
29.2
19.2
46.0
36.5
29.8
0%
10%
20%
30%
40%
50%
60%
Community
Punishment
Intermediate
Punishment
Prison Releases
No PRS
Prison Releases
PRS
% Rearrest % Reconviction % Reincarceration
SOURCE: NC Sentencing and Policy Advisory Commission,
FY 2005/06 Correctional Program Evaluation
v
As shown in Figure 5,
rates for all of the criminal
justice outcome measures during
the three-year follow-up period
varied considerably by offender
risk level, with a stair-step
increase in rates from low risk to
medium risk to high risk. When
compared to low risk offenders,
high risk offenders were over
three and one-half times more
likely to be rearrested, about
four and one-half times more
likely to be reconvicted, and
about four times more likely to
be reincarcerated.
Interim Outcome Measures
In addition to the recidivism rates, information is provided on two interim outcome
measures: 1) technical revocation of probation or post-release supervision for offenders while
under supervision in the community and 2) infractions for prisoners prior to release from prison.
Revocations were limited to those that are technical in nature because revocations for new crimes
would duplicate the recidivist arrest data.
Thirteen percent of the FY 2005/06 sample, including both probationers and prisoners, had
a technical revocation during the one-year follow-up period, 22.9% had a technical revocation
during the two-year follow-up period, and 28.3% had a technical revocation during the three-year
follow-up period. Probationers with an intermediate punishment had the highest technical
revocation rates during the follow-up period, with 39.5% having a technical revocation within the
three-year follow-up. The higher technical revocation rates for intermediate probationers are
likely linked to the closer monitoring and more restrictive sanctions for these offenders while on
probation.
Forty-five percent of the FY 2005/06 prison releases had an infraction while in prison. A
higher percentage of prisoners with PRS had infractions while incarcerated compared to prisoners
with no PRS (78% and 42%, respectively), which is consistent with the greater offense seriousness
and the resulting longer time served for prisoners with PRS. The average number of infractions
based only on prisoners who had an infraction was 4.4. When examining the number of
infractions per inmate, it is important to control for time served as prisoners with longer sentences
have more time to accrue infractions. As expected, the average number of infractions increased as
time served increased.
SOURCE: NC Sentencing and Policy Advisory Commission,
FY 2005/06 Correctional Program Evaluation
21.2
48.9
74.9
12.3
32.9
54.7
15.6
36.6
57.8
0%
10%
20%
30%
40%
50%
60%
70%
80%
Low Risk Medium Risk High Risk
Figure 5
Criminal Justice Outcomes by Risk Level:
Three-Year Follow-Up
% Rearrest % Reconviction % Reincarceration
vi
Multivariate Analysis
Multivariate analyses were performed to further explore factors correlated with the
probability of recidivism. This method aimed to isolate the direction and magnitude of the impact
of an independent variable on an outcome measure, such as rearrest, while controlling for the
impact of all the other independent variables. These analyses examined two main dependent
variables as indicators of recidivism – rearrest and reincarceration – and two interim dependent
variables as indicators of offender misconduct – technical probation revocations and prison
infractions.
In the analysis of all offenders, a number of factors increased an offender’s probability of
rearrest during the three-year follow-up, including being male, black, being a youthful offender,
having a history of substance abuse, having a greater number of prior arrests, having a prior drug
arrest, and having a higher risk score. The number of prior incarcerations decreased the
probability of rearrest.
Two variables, prison infractions and technical probation revocations, were used not only
as predictors of recidivism but also as outcomes indicating prisoner or probationer misconduct.
For prisoners, being black, a youthful offender, having a mental health problem, and having a
higher number of prison incarcerations were associated with increases in the number of prison
infractions acquired. For probationers, being male, black, a youthful offender, and having a
history of substance abuse significantly increased the likelihood of a technical revocation.
Similar to rearrest, an analysis examining correlates of reincarceration for all offenders
found being male, a youthful offender, and having a felony as the current conviction increased the
probability of reincarceration. Factors associated with a decrease in the probability of
reincarceration included being black, employed, married, or having at least twelve years of
education.
Mental Health
A specific correctional population highlighted in this study is prison releases in FY
2005/06 with a mental health problem. Prisoners were identified as having a mental health
problem if they had a Pulheat score of two or greater during the current period of incarceration, or
if they received a DSM-IV mental illness diagnosis from DOP mental health services during the
period of incarceration. Using this definition, almost 28% of the 19,733 prison releases had a
mental health problem.
Compared to prisoners with no mental health problem, prisoners with a mental health
problem were less likely to be male (73.7% versus 91.3%) and less likely to be black (44.9%
versus 61.2%). Prisoners with a mental health problem were also slightly older (an average of
34.4 years of age versus 32.3 years of age) and more likely to be married (15.0% versus 12.3%).
In addition, prisoners with a mental health problem were more likely to have a substance abuse
problem than prisoners with no mental health problem (66.5% versus 54.9%).
vii
Prisoners with mental
health problems were more
likely to have a felony offense as
their most serious current
conviction (77.6%) than
prisoners with no mental health
problems (73.2%). The majority
of both groups of inmates had
convictions for Classes F-I
felonies. However, a higher
percentage of prisoners with
mental health problems had
current convictions for Classes
B1 through E felonies (18.0%)
than did prisoners with no
mental health problems (7.2%).
Consistent with the differences
in offense seriousness, prisoners
with a mental health problem
served longer sentences on
average than did prisoners with no mental health problem (20.5 months compared to 11.6 months).
In addition, prisoners with a mental health problem had a higher percentage of infractions while
incarcerated (54.4%) compared to prisoners with no mental health problem (41.7%), which can be
attributed in part to the greater offense seriousness and the resulting longer time served by the
prisoners with a mental health problem.
More offenders with a mental health problem were characterized as low risk than offenders
without a mental health problem (29.0% compared to 22.5%), and fewer offenders were
characterized as high risk than in the group of offenders without a mental health problem (15.4%
compared to 20.5%).
Three criminal justice outcomes – rearrest, reconviction, and reincarceration – were
examined in this study. As shown in Figure 6, the rates of rearrest, reconviction, and
reincarceration for prisoners with and without a mental health problem were very similar.
Prisoners with a mental health problem were slightly less likely to be rearrested and reconvicted in
the three-year follow-up period, but slightly more likely to be reincarcerated during the three-year
follow-up period. Mental health problems seem to have interacted differently with infractions and
recidivism. A multivariate analysis confirmed that having a mental health problem (as identified
by DOC) is meaningful in the prison context (through an increased likelihood of committing
infractions), but not particularly for recidivism behavior after release.
Figure 6
Criminal Justice Outcomes by Mental Health Status:
Three-Year Follow-Up
48.2 50.6 50.0
35.0 34.6
33.5 36.8 35.5 35.8
0%
10%
20%
30%
40%
50%
60%
Mental Health
Problems
No Mental Health
Problems
All Prison Releases
% Rearrest % Reconviction % Reincarceration
SOURCE: NC Sentencing and Policy Advisory Commission,
FY 2005/06 Correctional Program Evaluation
viii
The types of mental health diagnoses received by prisoners and the severity of those
diagnoses were also examined. There were differences in the proportions of female and male
prisoners with diagnosed disorders, with 41% of female prison releases and 19% of male prison
releases having a diagnosed mental
illness. The most frequently
diagnosed mental disorders were
Substance-Related Disorders,
which accounted for 36.5% of the
total diagnoses.
Finally, the co-occurrence
of substance abuse and mental
health problems was examined.
Overall, 67% of prisoners had a
substance abuse problem, a mental
health problem, or both (see Figure
7). Prisoners with only a substance
abuse problem had slightly higher
rearrest rates (52.5%) than prison
releases with both a substance
abuse problem and a mental health
problem (49.3%) or with only a
mental health problem (46.0%).
Summary and Conclusions
When information from the current report is added to the Sentencing Commission’s
previous recidivism studies, a wider array of findings and tentative conclusions emerge. These
reports, covering large samples of offenders released in North Carolina between FY 1993/94 and
FY 2005/06, provide a framework to look at trends in the state’s recidivism rates and related
factors. Generally, many of the findings that follow have remained constant over the course of the
recidivism studies and lead to the same general conclusions. Figure 8 provides a comprehensive
view of the three-year recidivism rates for FY 2005/06 sample of probationers and prisoners.
 Statewide recidivism rates have been remarkably consistent over the past twelve years,
although a slight increase in the year-to-year rates can be observed.
The six samples studied had rates ranging between 36.8% and 40.1%. While these
recidivism rates have been relatively stable, an incremental (albeit slow) growth during this
timeframe can be noted. Several reasons could account for this upward trend: 1)
improvements in technology which have led to a greater number of fingerprinted
misdemeanor arrests; 2) a decrease in the percentage of probationers and an increase in the
percentage of prisoners; and 3) a shift in the risk level distributions for the samples of
offenders to fewer low risk offenders in more recent years.
SOURCE: NC Sentencing and Policy Advisory Commission,
FY 2005/06 Correctional Program Evaluation Data
40%
18% 9%
33%
Figure 7
Prevalence of Substance Abuse and Mental
Health Problems among Prisoners
(n=19,733)
Substance Abuse Problem only Mental Health Problem only
Both Problems Neither Problem
ix
 Intermediate punishments continue to provide an effective alternative in the range of
graduated sanctions between probation and incarceration.
Findings of this and previous reports confirmed that, while the general profile of intermediate
probationers more closely mimicked that of prisoners than of community probationers, their
rearrest rates were considerably and consistently lower than those of prisoners. This finding
lends continued support to the notion of intermediate sanctions as a viable alternative to
supervise certain offenders in the community in lieu of incarceration.
 The timing and targeting of correctional resources is crucial in reducing recidivism.
When the correctional response is intensive, well-targeted for an offender’s needs, especially
during the first year of supervision, it seems to produce a correctional alternative that is less
expensive and more successful in reducing future reoffending. Focusing more supervision
and resources in the first year of an offender’s placement in the community seemed to hold
true for released prisoners as well, reaffirming the value of some type of re-entry or
supervision following release.
 The validity of offender risk scores as a predictive tool might point to its use at various points
in the criminal justice decision making process.
The use of risk scores in this and previous reports has proven to be the most comprehensive
predictive measure of recidivism. The risk score assigned to an offender, which is comprised
of preexisting personal and criminal history factors, has been consistently associated with the
disposition and program assignments imposed by the court as well as with the offender’s
probability of reoffending. Since the most expensive correctional resources (i.e., prisons) are
predominantly being used by the high risk offenders and minimal resources are required by
the low risk offenders, it may prove to be a good use of tax dollars to target medium risk
offenders for less restrictive correctional programming. This investment in offenders who
are medium risk may play an important part in reducing their possibility of recidivating and
ultimately utilizing more expensive resources. The availability of risk scores earlier in the
criminal justice process might also help inform the discretion of decision makers such as
magistrates, judges and prosecutors as they make decisions on bail, pleas, conviction and
sentencing.
The issues surrounding the planning for the future of criminal justice policies, technology
and resources are enormous and complex. While the state’s crime rate has been lower during the
past several years, there has been a steady increase in the number of offenders and, more
specifically, offenders convicted and sentenced to prison. This increase in probationers and
prisoners creates a need for correctional resources that exceed the funding that is available.
Consequently, the question for the immediate future becomes one of how the state deals with
crime in the smartest and most cost-effective way while maintaining public safety. The governor,
legislature and criminal justice agencies are currently addressing these challenges through various
initiatives that are designed to have an effect on the offender population. A few of these initiatives
underway are Justice Reinvestment, the StreetSafe Task Force, the Criminal Justice Law
Enforcement Automated Data Services (CJLEADS), the Youth Accountability Planning Task
x
Force, the DOC’s Black Mountain Substance Abuse Treatment Center for Women, and the DOC’s
Division of Community Corrections’ Offender Risk and Needs Assessment.
While this report examined the effect of personal characteristics, current offense, prior
criminal history and program participation as predictors of whether an offender will recidivate,
future research should examine how these same factors affect when an offender will recidivate.
Targeting resources to match offender needs might increase the probability of rehabilitation;
knowledge of factors that predict when offenders with certain characteristics tend to recidivate
would provide practical information to programs for developing additional treatment or
supervision protocols that could further delay, or even prevent recidivism.
In summary, expectations for correctional success in preventing future criminality should
be viewed realistically. Components of an offender's criminal history, current offense, and
experiences with the correctional system are all elements strongly correlated with continued
criminal behavior. Expectations for rehabilitative success should be articulated in this context,
and be realistic in weighing criminogenic factors brought with an offender into the system
compared to the short time and limited resources at the DOC’s disposal to reverse their impact.
xi
Figure 8
Three-Year Recidivism Rates for the FY 2005/06 Sample
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
1
CHAPTER ONE
INTRODUCTION
Introduction
With the enactment of the Structured Sentencing Act (SSA) in 1994, North Carolina
embarked on a new penal strategy. Since that time, Structured Sentencing has benefited the
criminal justice system by increasing consistency, certainty and truth in the sentencing of
offenders; setting priorities for the use of correctional resources; and balancing sentencing policies
with correctional resources. The issue of correctional resources and, specifically, their
effectiveness in increasing public safety and deterring future crime have continued to be of interest
to legislators and policy makers. It is the goal of most programs to sanction and control offenders,
to offer them opportunities that will assist in altering negative behavioral patterns, and,
consequently, to lower the risk of reoffending.
Studies which measure recidivism are a nationally accepted way to assess the effectiveness
of in-prison and community corrections programs in preventing future criminal behavior. The
North Carolina General Assembly, aware of this trend, incorporated the study of recidivism into
the Sentencing and Policy Advisory Commission’s mandate from the start. The first recidivism
study that was prepared for the Commission was completed in 1992 by Stevens Clarke and Anita
Harrison of the Institute of Government at the University of North Carolina Chapel Hill (now
named the School of Government). This recidivism study was followed by one that was
conducted in 1996 by Mark Jones and Darrell Ross of the School of Social Work at East Carolina
University. In 1997 and 1998, the Commission produced the third and fourth recidivism reports in
conjunction with the Department of Correction’s Office of Research and Planning.
During the 1998 Session, the General Assembly redrafted the Commission’s original
mandate to study recidivism and expanded its scope to include a more in-depth evaluation of
correctional programs. This legislation (Session Law 1998-212, Section 16.18) gives the
following directive:
The Judicial Department, through the North Carolina Sentencing and Policy
Advisory Commission, and the Department of Correction shall jointly conduct
ongoing evaluations of community corrections programs and in-prison treatment
programs and make a biennial report to the General Assembly. The report shall
include composite measures of program effectiveness based on recidivism rates,
other outcome measures, and costs of the programs. During the 1998-99 fiscal
year, the Sentencing and Policy Advisory Commission shall coordinate the
collection of all data necessary to create an expanded database containing
offender information on prior convictions, current conviction and sentence,
program participation and outcome measures. Each program to be evaluated
shall assist the Commission in the development of systems and collection of data
necessary to complete the evaluation process. The first evaluation report shall be
presented to the Chairs of the Senate and House Appropriations Committees and
the Chairs of the Senate and House Appropriations Subcommittees on Justice and
2
Public Safety by April 15, 2000, and future reports shall be made by April 15 of
each even-numbered year.
The first evaluation report, as required by law, was delivered to the General Assembly on
April 15, 2000. The current study is the sixth biennial Correctional Program Evaluation Report
and it contains information about offender characteristics, specific correctional programs, outcome
measures, and an expansive methodological approach to examine the relationship between
offender risk factors, correctional programs, and recidivism rates.
Defining Recidivism
The North Carolina General Assembly directed the Sentencing Commission to measure the
rates of recidivism of criminal offenders involved in state-supported correctional programs. The
legislation calling for these measurements made it clear that recidivism meant repeated criminal
behavior, and implied that measuring recidivism was to be a way of evaluating correctional
programs – that is, programs designed or used for sanctioning and, if possible, rehabilitating or
deterring convicted criminal offenders.
Correctional programs do not affect crime directly; rather, they are designed to change
offenders’ attitudes, skills, or thinking processes, in the hope that their social behavior will change
as a result. The punitive aspect of criminal sanctions might also serve as an individual deterrent
with convicted offenders. Policy makers such as legislators tend to be concerned with whether the
programs ultimately reduce criminal behavior. This concern is understandable. A program may
be successful in educating, training, or counseling offenders, but if it does not reduce their
subsequent criminal behavior, the result will simply be repeat offenders who are better educated or
have greater self-confidence.
There is no single official definition of recidivism. Researchers have used a variety of
definitions and measurements, including rearrest, reconviction, or reincarceration, depending on
their particular interests and the availability of data. Therefore, in comparing recidivism of
various groups of offenders, readers are well advised to be sure that the same definitions and
measurements are used for all groups. Official records from police, courts, and correctional
agencies are the source of most research on adult recidivism. For offenders involved in a
recidivism study, different types of records will indicate different rates of recidivism.
In its studies of recidivism, the Sentencing Commission uses rearrests as the primary
measure of recidivism, supplemented by information on reconvictions and reincarcerations, to
assess the extent of an offender’s repeat involvement in the criminal justice system. The
advantages of arrest data, compared with other criminal justice system data, outweigh the
disadvantages. Rearrests, as used in this research, take into account not only the frequency of
repeat offending but also its seriousness and the nature of the victimization (for example, crimes
against the person, crimes involving theft or property damage, or crimes involving illegal drugs).
The volume of repeat offending is handled by recording the number of arrests for crimes of
various types.
3
Structured Sentencing and Recidivism
North Carolina law prescribes the use of guidelines in sentencing its convicted felons and
misdemeanants. In theory, Structured Sentencing may affect recidivism in a variety of ways. Its
penalty framework may alter the deterrent effect of sentencing laws, with different punishments
influencing differently an individual offender's fear of the consequences of crime and thereby
changing his or her likelihood of reoffending. Guidelines might also impact recidivism by altering
the characteristics, or “mix,” of groups of offenders – for example, probationers or prisoners.
Impacting the composition of groups of offenders has been, from the start, one of the changes
contemplated by the guidelines sentencing movement, and this alteration may well affect group
recidivism rates.
Sentencing guidelines have sought to make offenders convicted of violent crimes, as well
as repeat offenders, more likely to receive active prison sentences and to serve longer prison
terms. At the same time, guidelines were intended to make first-time offenders charged with non-violent
crimes less likely to be imprisoned, and to have them serve shorter terms if imprisoned.
As a result, guidelines in North Carolina and elsewhere have tended to shift some offenders to
probation who formerly would have gone to prison, and others to prison who formerly might have
received probation. This kind of shift was expected to change recidivism rates by type of
punishment, by re-mixing not only the offense profile of various groups but, perhaps more
importantly, the profile of their criminal histories.
North Carolina’s Structured Sentencing emphasized not only the diversion of some
offenders from prison to probation, but also creating a middle option – the use of intermediate
punishments – for those diverted offenders. Intermediate punishments – i.e., enhancements to
probation such as intensive supervision, special probation (split sentences), and day reporting
centers – were meant to control the recidivism of offenders diverted from prison to probation.
Intermediate probationers, supervised more closely than community probationers but not exposed
to the detrimental effects of prisonization, tend to have recidivism rates between the rates of the
two other groups.
Some new dynamics gleaned from recidivism studies of the past ten years are worth noting
in this context:
 Risk scores, built primarily on prior behavior but much more extensive in summarizing an
offender’s risk to reoffend, were found to be of considerable value in predicting
recidivism. Risk scores, in turn, are strongly linked to the sentencing structure (through
offense class and prior record level) and to the type of disposition imposed by the judge.
 The meaning of “type of punishment” has been diluted by a high incidence of probation
revocations to prison. Revocation leads to underestimating the propensity to reoffend for
community and intermediate probationers (through a reduction in their “time-at-risk”), and
weakens the correlation between structured sentencing, punishment type, and recidivism.
The overall conclusion so far points to guidelines increasing the within-group
predictability of recidivism by changing the internal group profiles, but having little to no effect
on cohort recidivism rates, which have remained remarkably stable over time.
4
Comparison of Recidivism Rates for Previous Recidivism Studies
The Sentencing Commission’s previous recidivism reports provide a framework to look at
trends in the state’s recidivism rates. However, it should be noted that there are differences in the
recidivism studies that make comparisons difficult. For example, samples up to, but not including,
FY 1996/97 are based only on offenders convicted and sentenced under or prior to the Fair
Sentencing Act (FSA), the samples for FY 1996/97 through FY 2001/02 include a mixture of
offenders sentenced under the FSA and the SSA, and the sample for FY 2003/04 only includes
offenders sentenced under the SSA. The various studies also have different follow-up periods.
Nonetheless, some overall comparisons may be made as long as these factors are taken into
consideration.
Table 1.1 presents overall recidivism rates (measured as rearrest) from each of the
Sentencing Commission’s previous reports. The table indicates that recidivism rates for offenders
have been fairly similar over the sample years, given the differences in follow-up time and sample
composition. The 1989 study, the FY 1996/97 study, and the FY 1998/99 study had a similar
follow-up period (of approximately two years) and similar recidivism rates for all offenders,
ranging from 31% to 33%. The six other studies, with more extended follow-up periods (of
approximately three years), reported slightly higher recidivism, with rearrest rates for all offenders
between 33% and 39%.
Table 1.1 also provides a comparison of recidivism rates for probationers and prisoners for
the previous studies. The recidivism rates for FSA regular probationers can be compared across
the previous recidivism studies (sample years 1989, 1992/93, 1993/94, and 1994/95) and with the
SSA community punishment probationers in the more recent studies (sample years 1996/97,
1998/99, 2001/02, and 2003/04). Based on the studies with an approximate three-year follow-up
period, rearrest rates for FSA regular probationers ranged from 22.8% for the 1992/93 sample to
31.3% for the 1994/95 sample, while recidivism rates for SSA community punishment
probationers were around 30% for the 1998/99 and 2001/02 samples and around 34% for the
2003/04 sample.
Recidivism rates for prisoners can also be compared across the studies. The rearrest rates
provided for prisoners for sample years 1989, 1992/93, 1993/94, and 1994/95 are for prisoners
released on regular parole prior to or under FSA, while the rearrest rates provided for prisoners for
sample years 1996/97, 1998/99, and 2000/01 are for prisoners released under both FSA and SSA.
The 2003/04 sample provides the first SSA only prisoner sample. The three-year rearrest rates for
FSA prisoners ranged from 45.9% for the 1992/93 sample to 48.8% for the 1993/94 sample, while
the three-year rearrest rates for SSA and FSA prisoners were almost 50% for the 1998/99 and
2001/02 samples. Similarly, the three-year rearrest rate for SSA prisoners in the 2003/04 sample
was also around 50%. It must be noted that any comparison of FSA and SSA prisoners needs to
account for differences in the characteristics of these two groups relative to offense seriousness
and time served.
Table 1.1
Rearrest Rates for North Carolina Offenders
Sample Rearrest Rates
Year Sample Composition Sample
Size
Follow-Up
Period1
(in months)
All
Offenders Probationers2 Prisoners3
Studies with a Two-Year Follow-Up Period
1989 Offenders sentenced prior to or under FSA 37,933 26.7 31.2% 26.5% 41.3%
1996/97 Offenders sentenced under FSA and SSA 51,588 24 32.6% 26.3% 42.6%
1998/99 Offenders sentenced under FSA and SSA 58,238 24 31.2% 24.2% 41.6%
Studies with a Three-Year Follow-Up Period
1992/93 Offenders sentenced prior to or under FSA 33,111 36.7 32.6% 22.8% 45.9%
1993/94 Offenders sentenced prior to or under FSA 48,527 32.8 36.8% 30.7% 48.8%
1994/95 Offenders sentenced prior to or under FSA,
with some offenders sentenced under SSA 45,836 35.1 37.3% 31.3% 47.8%
1998/99 Offenders sentenced under FSA and SSA 58,238 36 37.8% 29.7% 49.6%
2001/02 Offenders sentenced under SSA, with some
offenders sentenced under FSA 57,973 36 38.2% 30.1% 49.8%
2003/04 Offenders sentenced under SSA 56,983 36 38.7% 33.7% 50.2%
SOURCE: NC Sentencing and Policy Advisory Commission
1 Variable follow-up periods were used for sample years 1989 through 1994/95. Fixed follow-up periods were used for sample years 1996/97, 1998/99,
2001/02, and 2003/04.
2 This category includes FSA offenders on regular probation for sample years 1989, 1992/93, 1993/94, and 1994/95 and SSA offenders on community
punishment probation for sample years 1996/97, 1998/99, 2001/02, and 2003/04.
3 This category includes FSA prisoners released on regular parole through FY 1994/95, FSA and SSA prisoners from 1996/97 – 2001/02, and SSA prisoners for
2003/04.
6
While it is too soon to draw valid comparisons between the overall recidivism of SSA
and FSA offenders, it is worth noting that recidivism rates over a two-year follow-up were
around 32% for the various samples and over a three-year follow-up were around 37-38%,
independent of composition. Structured Sentencing might have an impact on recidivism rates by
altering the deterrent effect of sentencing laws and by altering the characteristics, or “mix,” of
groups of offenders, but fluctuation in the rates will ultimately be affected by a host of social and
legal factors, in addition to the sentencing laws. Future studies will continue to examine these
issues.
Research Design and Methodology
The Sentencing Commission’s expanded legislative mandate translated to a more
comprehensive approach in capturing relevant empirical information. The theoretical model
adopted to study recidivism pointed to data collection in three time frames for each offender:
preexisting factors such as demographic characteristics and criminal history; current criminal
justice involvement including current conviction, sentence and correctional program
participation; and future measures of social reintegration such as rearrest, reconviction, and
reincarceration.4
Sample
The sample selected for study included all offenders released from prison by the North
Carolina Department of Correction (DOC) or placed on supervised probation during Fiscal Year
2005/06. The final study sample includes 60,824 offenders sentenced under Structured
Sentencing, affording a comprehensive look at the recidivism of Structured Sentencing
offenders.5
Follow-up Period
Recidivism studies utilize varying lengths of time as their follow-up period, depending on
the availability of data and other resources. With both short term and long term recidivism being
of great interest to policy makers, this report provides information on the recidivism of the FY
2005/06 sample of offenders with a fixed three-year follow-up period, with one-year, two-year,
and three-year rates provided. When not specified, recidivism will be defined based on the
three-year follow-up period.
Time at Risk
While each offender in the study sample had an equal three-year follow-up period, not all
of them were on the street and “at risk” of recidivism for the entire three years. The report takes
into account each sample offender’s actual time at risk, by identifying their periods of
4 Preexisting factors and current criminal justice involvement are also components in targeting offenders for
different correctional sanctions and treatment programs, and assessing their risk levels.
5 Pre-Fair Sentencing Act (FSA) and FSA offenders (prisoners and probationers) were excluded from the sample.
All DWI and traffic offenders were also excluded from the sample.
7
incarceration in North Carolina’s prison system within the follow-up time frame and subtracting
the time incarcerated from the follow-up period.6
Outcome and Process Measures
The outcome and process measures examined for this study include:
► Recidivism, defined broadly to cover the offender’s possible span of
reinvolvement in the North Carolina criminal justice system, to include rearrests,
reconvictions, and reincarcerations.
► For offenders supervised in the community, technical revocation of probation or
post-release supervision.
► For prisoners, prison infractions during incarceration.
Data Sources and Enhancements
Two automated data sources were utilized to collect aggregate data on the sample of
offenders:
► The North Carolina Department of Correction’s (DOC) Offender Population
Unified System (OPUS) provided demographic and prior record information,
current convicted offense and sentence,7 correctional program assignment, type of
punishment, and subsequent technical probation revocations and prison
incarcerations.
► The North Carolina Department of Justice’s (DOJ) data set was used to provide
fingerprinted arrest records for prior and recidivist arrests, as well as recidivist
convictions.
The final data set for this study consists of over 300 items of information (or variables)
for the sample of 60,824 offenders released to the community between July 1, 2005 and June 30,
2006, and followed for three years.8 A case profile was constructed for each sample offender,
comprised of personal and criminal history characteristics, the most serious current offense of
conviction, type of punishment imposed, correctional program assignments, and reinvolvement
with the criminal justice system (i.e., rearrest, reconviction, and reincarceration).
There were several enhancements made to the data provided in this report:
► Due to the growing number of prisoners released onto post-release supervision,
this group has been added as a subgroup of the prison release sample where
applicable in the report.9
6 Since each county jail maintains its own data, it was not possible to account for time served in county jails during
the follow-up period; nor was it possible to account for time incarcerated in other states during the follow-up period.
7 “Current” in the context of this study refers to the most serious conviction and sentence for which the offender
was released to the community within the sample time frame.
8 A glossary of relevant variables is included in Appendix A.
9 This group of prisoners was examined in detail in the Commission’s 2008 recidivism report.
8
► In Chapter Three, the recidivism of community punishment probationers,
intermediate punishment probationers, and prisoners with post-release supervision
were examined to determine whether any rearrests occurred while the offender was
being supervised for the conviction that placed them in the sample.
�� Recognizing the different ways that prisoners enter prison (as a result of a new
crime, a technical revocation of probation or post-release supervision, or both), an
indicator was created that identifies the type of entry for each prisoner, with
analyses that examine differences in these groups (e.g., current conviction,
recidivism) provided where applicable in the report.
► An indicator of mental health issues was added for the prisoners in the sample and
is used in the multivariate analyses in Chapter Four and examined in detail in
Chapter Five. Similar information will be analyzed for probationers as it becomes
available.
Report Outline
Chapter Two presents a descriptive statistical profile of the sample and aggregate figures
on the incidence and type of prior criminal behavior. It also describes the sample in terms of
offender risk (a composite “risk score” developed and assigned to each offender).
Chapter Three includes a descriptive analysis of the sample’s subsequent (i.e.,
recidivistic) criminal involvement, with special focus on the one-, two-, and three-year follow-up.
This analysis also allows for some comparisons between the recidivism of offenders released
from prison and those placed on some form of probation.
Chapter Four utilizes multivariate techniques to assess the relationship between
recidivism and various disposition types while controlling for other relevant preexisting factors.
Risk scores are used in the analysis to isolate the impact of correctional dispositions and
programs on the probability of recidivism while holding constant the “risk level” of the offender.
Chapter Five provides statistical information regarding mental heath issues for prisoners.
Finally, Chapter Six offers a short summary of the study’s main findings and closes with some
observations on recidivism in North Carolina following the enactment of Structured Sentencing.
9
CHAPTER TWO
STATISTICAL PROFILE OF THE FY 2005/06 SAMPLE
Type of Punishment
As described in Chapter One, the study sample comprises SSA offenders who either were
placed on probation or were released from prison during FY 2005/06.10
As shown in Figure 2.1, there were 60,824 offenders in the FY 2005/06 sample. There
were 41,091 (68%) probationers and 19,733 (32%) prisoners. These can be further subdivided
into the following four categories based on type of punishment:
Probation Entries
 probationers who received a community punishment;
 probationers who received an intermediate punishment;
Prison Releases
 prison releases with no post-release supervision (no PRS); and
 prison releases with post-release supervision (PRS).
Many of the tables in this chapter present information by probation or prison status for
individual categories of probationers and prisoners (also referred to as type of punishment) and
for the sample as a whole.
10 Offenders whose offenses were committed on or after October 1, 1994, were sentenced under the Structured
Sentencing Act (SSA).
FY 2005/06 Sample
The sample comprises all SSA offenders who were placed on supervised
probation or were released from prison during FY 2005/06, with the
following exclusions:
 offenders with a most serious current conviction for driving
while impaired (DWI); and
 offenders with a most serious current conviction for a
misdemeanor traffic offense.
10
Figure 2.1
Type of Punishment
Definitions for the Types of Punishment
Probation Entries: Offenders who were sentenced under the Structured Sentencing Act and received a probation
sentence (i.e., the active sentence was suspended).
Probation Entries with a Community Punishment: An offender who received a community
punishment. Community punishments may consist of a fine, unsupervised probation (although
unsupervised probationers were excluded from the sample), or supervised probation, alone or with one or
more of the following conditions: outpatient drug/alcohol treatment, community service, assignment to
TASC, payment of restitution, or any other conditions of probation that are not considered an intermediate
punishment. Offenders with little or no prior criminal history who commit the lowest class felonies (Class
H or I) and all misdemeanants may receive a community punishment.
Probation Entries with an Intermediate Punishment: An offender who received an intermediate
punishment. An intermediate punishment requires a period of supervised probation with at least one of the
following conditions: special probation, assignment to a residential treatment program, house arrest with
electronic monitoring, intensive probation, assignment to a day reporting center, and assignment to a drug
treatment court program. Generally, offenders who have a significant prior record and commit Class H or I
felonies and offenders who have little or no prior record and commit more serious non-violent felonies may
receive an intermediate punishment.
Prison Releases: An offender who was sentenced under the Structured Sentencing Act, served his/her maximum
sentence minus earned time and time for pre-conviction confinement, and was released back into the community,
usually without any supervision. This group included offenders who were sentenced to prison for a new crime, who
were revoked to prison for a technical violation of their probation, or both. A small number (n=2,009 or
approximately 10%) of the prison releases received post-release supervision.
Prison Releases with No Post-Release Supervision (no PRS): Under Structured Sentencing, prisoners
released with a most serious offense for Class F through Class I felonies and Class A1 through Class 3
misdemeanors are released from prison without any supervision.
Prison Releases with Post-Release Supervision (PRS): Under Structured Sentencing, prisoners released
with a most serious offense for Class B1 through Class E felonies are released on post-release supervision
for a period of nine months, with the exception of sex offenders who are supervised for five years.
See Appendix B for further descriptions of the types of punishment and for many of the programs that fall under
them.
All Probation Entries and Prison Releases
(N=60,824)
Prison Releases
32.4% (n=19,733)
Probation Entries
67.6% (n=41,091)
Community Punishment
47.1% (n=28,645)
Intermediate Punishment
20.5% (n=12,446)
Post-Release Supervision
3.3% (n=2,009)
No Post-Release Supervision
29.1% (n=17,724)
11
Personal Characteristics
Table 2.1 contains information describing the personal characteristics of the FY 2005/06
sample.11,12 Of the 60,824 offenders, 78.5% were male, 51.2% were black, 13.8% were married,
39.4% had twelve or more years of education, 39.6% were identified as having a substance abuse
problem, and their average age, at release from prison or placement on probation, was 31.3.
Probationers (and, in particular, probationers with community punishments) had a higher
percentage of females than prisoners. On average, offenders who were placed on probation were
slightly younger than offenders who were released from prison.
Criminal History
It is important to look at the number of prior arrests for the offenders in the sample
because previous research indicates that prior arrests are a strong predictor of recidivism (Clarke
and Harrison, 1992; Jones and Ross, 1996; NC Sentencing and Policy Advisory Commission,
1997; 1998; 2000; 2002; 2004; 2006; 2008). Information on prior fingerprinted arrests for the
FY 2005/06 sample is provided in Table 2.2.13 Prior arrests were defined as fingerprinted arrests
that occurred before the conviction that placed the offender in the sample. Beginning with the
current report, this measure excludes the arrest event for the conviction that placed the offender
in the sample.
Overall, 55.9% of community punishment probationers, 80.7% of intermediate
punishment probationers, 90.5% of prison releases with no PRS, and 87.9% of prison releases
with PRS had one or more prior arrests. As a whole, the FY 2005/06 sample accounted for a
total of 182,885 prior arrests. For offenders with prior arrests, the number of prior arrests
generally increased by type of punishment from community punishment to intermediate
punishment to prison. For example, 43.5% of community punishment probationers had only one
prior arrest compared to 14.9% of prison releases with no PRS. A similar pattern was found
when comparing the average number of arrests for the subgroups, with probationers having an
average of 3.3 prior arrests and prisoners having an average of 5.4 prior arrests.
With regard to arrest history, intermediate punishment probationers were more similar to
prisoners than to community punishment probationers. The average number of prior arrests for
those offenders with at least one prior arrest was 2.9 for community punishment probationers,
4.0 for intermediate punishment probationers, 5.4 for prison releases with no PRS, and 5.2 for
prison releases with PRS. These findings confirm the philosophy behind Structured Sentencing
that probationers who receive intermediate punishments are more serious offenders than those
who receive community punishments, but less serious than those who receive prison sentences.
11 See Appendix A for a description of major variables.
12 Available mental health data for prisoners in the DOC’s automated database indicated that 28% of prisoners had
some type of mental health issue. Comparable mental health data were not available for probationers. Further
examination of mental health issues for prisoners is provided in Chapter Five of this report.
13 Fingerprinted arrest data from the DOJ were used to determine prior arrests. Reflecting a methodological change
from previous reports, the measure of prior arrests excludes the arrest for the conviction that placed the offender in
the sample.
Table 2.1
Personal Characteristics by Type of Punishment
Type of Punishment N %
Male
%
Black
Mean
Age
%
Married
% With
Twelve Years
of Education
or More
% With
Substance
Abuse
Probation Entries
Community Punishment 28,645 71.1 47.0 30.1 14.6 46.4 28.0
Intermediate Punishment 12,446 83.1 52.2 31.4 13.3 38.4 37.0
Subtotal 41,091 74.7 48.6 30.5 14.2 43.9 30.7
Prison Releases
No Post-Release Supervision 17,724 85.7 55.4 32.8 13.1 30.7 58.8
Post-Release Supervision 2,009 93.6 68.3 33.3 12.8 32.6 52.3
Subtotal 19,733 86.5 56.7 32.9 13.1 30.9 58.1
TOTAL 60,824 78.5 51.2 31.3 13.8 39.4 39.6
Note: There are missing values for self-reported years of education.
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
Table 2.2
Prior Arrests by Type of Punishment
Prior Arrests for Offenders with Any Prior Arrest
(n=43,865)
Type of Punishment N Number of Prior Arrests (%)
% Any
Prior
Arrest
Total #
of Prior
Arrests
Average
Number
of Prior
Arrests 1 2 3-4 5-9 10+
Probation Entries
Community Punishment 28,645 55.9 46,161 2.9 43.5 20.9 18.6 12.7 4.2
Intermediate Punishment 12,446 80.7 40,181 4.0 25.2 19.4 24.7 23.0 7.6
Subtotal 41,091 63.4 86,342 3.3 36.5 20.3 21.0 16.7 5.5
Prison Releases
No Post-Release Supervision 17,724 90.5 87,371 5.4 14.9 15.4 24.5 30.0 15.3
Post-Release Supervision 2,009 87.9 9,172 5.2 20.0 16.4 20.6 26.6 16.3
Subtotal 19,733 90.2 96,543 5.4 15.4 15.5 24.1 29.6 15.4
TOTAL 60,824 72.1 182,885 4.2 27.9 18.4 22.2 21.9 9.6
Note: Percentages may not add to 100% due to rounding.
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
14
Most Serious Current Conviction
Overall, 48% (n=29,304) of the FY 2005/06 sample had a felony offense as the most
serious current conviction and 52% (n=31,520) had a misdemeanor offense as the most serious
current conviction.14 Figures 2.2 and 2.3 present the category of conviction (violent, property,
drug, or “other”) for probation entries and prison releases by felony/misdemeanor status.
As shown in Figure 2.2, the majority of probationers with a current felony conviction had
convictions for drug offenses (46%), followed by property offenses (35%). For prisoners with a
current felony conviction, the majority had convictions for property offenses (37%), followed by
convictions for drug offenses (32%). As anticipated, prisoners were more likely to have current
convictions for violent offenses (22%) than probationers (12%).
Figure 2.2
Most Serious Current Conviction: Felonies Only (n=29,304)
Probation Entries
(n = 14,615)
Prison Releases
(n = 14,689)
12%
7%
46% 35%
Violent Property Drug Other
37%
32%
9%
22%
Violent Property Drug Other
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
The majority of prisoners and probationers with current misdemeanor convictions were
convicted of property offenses (41% and 40%, respectively) and violent offenses (38% and 27%,
respectively), as shown in Figure 2.3. As expected, prisoners had a higher percentage of violent
convictions compared to probationers. Probationers had a higher percentage of drug convictions
(23%) compared to prisoners (16%).
The most serious current conviction by type of punishment is presented in Table 2.3.
Overall, 38.1% of the sample had a most serious current conviction for a property offense,
14 Each offender’s conviction(s) that placed him/her in the sample as a prison release or probation entry during FY
2005/06 were ranked in terms of seriousness and only the most serious conviction was used for analysis. For the
sake of brevity, the term “most serious current conviction” is often referred to as “current conviction.” See
Appendix A for information on the categorization of offenses as person, property, drug, and “other.”
Table 2.3
Most Serious Current Conviction by Type of Punishment
Type of Conviction
% Violent % Property % Drug % Other
% Total
Type of Punishment N
Fel. Misd. Fel. Misd. Fel. Misd. Fel. Misd. Fel. Misd.
Probation Entries
Community Punishment 28,645 0.5 20.4 6.6 33.7 10.4 19.4 0.7 8.4 18.1 81.9
Intermediate Punishment 12,446 13.3 10.3 25.7 7.7 30.2 4.2 6.5 2.2 75.7 24.3
Subtotal 41,091 4.4 17.4 12.3 25.8 16.4 14.8 2.5 6.5 35.6 64.4
Prison Releases
No Post-Release Supervision 17,724 9.3 10.9 29.9 11.7 26.7 4.5 5.6 1.3 71.5 28.5
Post-Release Supervision 2,009 75.5 N/A 4.2 N/A 1.5 N/A 18.7 N/A 100.0 N/A
Subtotal 19,733 16.1 9.8 27.3 10.5 24.2 4.1 6.9 1.2 74.4 25.6
TOTAL 60,824 8.2 14.9 17.2 20.9 18.9 11.3 3.9 4.8 48.2 51.8
Note: Percentages may not add to 100% due to rounding.
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
1 6
Figure 2.3
Most Serious Current Conviction: Misdemeanors Only (n=31,520)
Probation Entries
(n = 26,476)
Prison Releases
(n = 5,044)
27%
10%
23%
40%
Violent Property Drug Other
41%
16%
5%
38%
Violent Property Drug Other
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
followed by 30.2% for drug offenses, 23.1% for violent offenses, and 8.7% for “other”
offenses.15 Community punishment probationers were more likely to have a most serious current
conviction for a misdemeanor offense (81.9%) and the current conviction was most likely to be
for a misdemeanor property offense (33.7%). Seventy-six percent of intermediate punishment
probationers had a most serious current conviction for a felony offense and the current
conviction was most likely to be for a felony drug offense (30.2%) or for a felony property
offense (25.7%). Seventy-two percent of prisoners with no PRS had a most serious current
conviction for a felony offense. Prisoners with no PRS were most likely to have a most serious
conviction for a property offense (29.9%), followed by a drug offense (26.7%). Consistent with
current law, 100% of prison releases with PRS had a most serious current conviction for a felony
offense. Prisoners with PRS were most likely to have a most serious conviction for a violent
offense (75.5%), and for an offense in the “other” category (18.7%) which includes offenders
who have been convicted as habitual felons (Class C).
Table 2.4 presents information on the offense class of the most serious conviction for the
FY 2005/06 sample by type of punishment. Under Structured Sentencing, offenses are classified
based on offense seriousness, with Class A through Class E felonies considered the violent
felonies. The type of sentence imposed (community, intermediate punishment, or active
sentence) and the sentence length are based on the offense class for the most serious offense and
on the offender’s prior criminal history (see Appendix C for the felony and misdemeanor
punishment charts).16 Offenders convicted of Class B1 through Class D felonies are required to
15 Of the 14,045 offenders with a most serious current conviction for a violent offense, 8.8% (n=1,238) had a
conviction for an offense which requires registration as a sex offender under Article 27A of Chapter 14 of the NC
General Statutes. In addition, 71 offenders with a most serious conviction for an “other” offense were required to
register as sex offenders. These offenders were convicted of indecent exposure, which is categorized as a public
order offense according to DOC’s crime categorization scheme.
16 For further information about Structured Sentencing, see the NC Sentencing and Policy Advisory Commission’s
Structured Sentencing Training and Reference Manual.
1 7
receive an active sentence.17 Depending on their prior criminal history, offenders convicted of
Class E through G felonies may receive either an intermediate punishment or an active sentence,
while offenders convicted of Class H through Class I felonies or of misdemeanor offenses may
receive a community punishment, an intermediate punishment, or an active sentence.
Table 2.4
Offense Class for Most Serious Current Conviction by Type of Punishment
Offense Class for Current Conviction
Type of Punishment N %
B1-E
Felony
%
F-I
Felony
%
Misd.
Probation Entries
Community Punishment 28,390 N/A 18.1 81.9
Intermediate Punishment 12,384 4.0 71.6 24.4
Subtotal 40,774 1.2 34.4 64.4
Prison Releases
No Post-Release Supervision 17,724 N/A 71.5 28.5
Post-Release Supervision 2,009 100.0 N/A N/A
Subtotal 19,733 10.2 64.3 25.6
TOTAL 60,507 4.1 44.1 51.8
Note: Offenders with discrepant or unknown offense classes were excluded from this table (n=317). Percentages
may not add to 100% due to rounding.
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
For the FY 2005/06 sample, 4.1% had a most serious conviction for a Class B1 through E
felony,18 44.1% had a most serious conviction for a Class F through Class I felony, and 51.8%
had a most serious conviction for a misdemeanor offense. Differences found in the offense class
composition of the sample subgroups are consistent with Structured Sentencing, which links
offense severity with type of punishment. The majority of community punishment probationers
had a most serious conviction for a misdemeanor offense (81.9%), while the majority of
17 Offenders convicted of first degree murder (Class A) may receive either a death sentence or life without parole
under Structured Sentencing.
18 Structured Sentencing does not allow for the release of offenders convicted of first degree murder (Class A),
explaining the absence of Class A offenders in the FY 2005/06 sample.
1 8
intermediate punishment probationers and prisoners with no PRS had a most serious conviction
for a Class F through Class I felony (71.6% and 71.5%, respectively). Approximately 10% of
prisoners had a conviction for a Class B1 through Class E felony.
Type of Prison Entry
The FY 2005/06 sample included 19,733 offenders released from prison. For the first
time in the Sentencing Commission’s recidivism studies, information was also available about
the reason for the prisoner’s entry to prison – whether the prisoner entered prison for a
conviction for a new crime, for a technical revocation of probation, or for both a new crime and a
revocation.
Forty percent of the FY 2005/06 prison releases (n=7,847) entered prison as a result of a
sentence for a new crime and served, on average, 21.3 months. Thirty-seven percent (n=7,406)
entered prison for a technical revocation of their probation and served, on average, 6.1 months.
The remaining 23% of prison releases (n=4,480) entered prison for a mix of both a new crime
and a technical revocation and served, on average, 14.4 months. These groups are referred to as
new crime entries, technical revocation entries, and mixed entries, respectively.
Table 2.5 examines the variation in the most serious current conviction for prisoners by
their type of prison entry. While the majority of offenders in all three groups were convicted for
felonies in Offense Classes F through I, the proportion of convictions for violent felonies (Class
B1 through E) was highest for prisoners admitted for a new crime (20.4%), and the proportion of
convictions for a misdemeanor was highest for prisoners revoked to prison (30.2%).
Table 2.5
Offense Class of Most Serious Current Conviction for Prison Releases by Type of Entry
Offense Class for Current Conviction
Prison Releases by
Type of Prison Entry N %
B1-E
Felony
%
F-I
Felony
%
Misd.
New Crime Entry 7,847 20.4 58.5 21.1
Technical Revocation Entry 7,406 1.2 68.6 30.2
Mixed Entry 4,480 7.0 67.1 25.9
TOTAL 19,733 10.2 64.3 25.6
Note: Due to the length of sentences imposed for Class B1 felonies, there were no prisoners released in the FY
2005/06 sample with a most serious conviction for a Class B1 felony. Percentages may not add to 100% due to
rounding.
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
1 9
Offender Risk and Recidivism
Evaluations of correctional programs that use recidivism as the outcome measure of
“success” are fairly commonplace. However, a frequent problem encountered by both
researchers and policy makers interpreting the results of these studies is that most have no way to
control for different levels of offender risk. Offenders vary in their risk of recidivating,
independent of any intervention provided. This finding has been confirmed repeatedly in
research, and is being applied in risk assessments used for sentencing and in correctional policy
(e.g., to classify inmate custody levels and to make parole decisions).
In a perfect research setting, offenders would be randomly selected into the various
correctional programs to be evaluated. In the reality of corrections, this is not possible because
of practical, public safety, and legal considerations. Instead, this study attempts to control
statistically for types of offenders by developing a risk model that divides offenders into three
levels of risk: high, medium, and low. Using risk level as an independent control variable
allows for a comparative analysis of the recidivism rates of offenders who did and did not
participate in a particular program or intervention.
Components of Risk
Variables used to create the “risk” measure for this study are those identified in the
literature as increasing or decreasing a person’s risk of being arrested.19 For the purposes of this
study, risk is defined as the projected probability of rearrest, and is not intended to measure
seriousness of future offenses or offender dangerousness.
A composite measure, risk is made up of a number of factors that can be loosely divided
into the following three categories:
1. Personal Characteristics
► Offender’s age when placed on probation or released from prison
► Sex
► Race20
► Marital status (i.e., married or not married) at the time of entry into the sample
► Employment status at time of arrest for prisoners and at the time of probation entry
for probationers
► History of substance abuse problems as indicated by prison or probation
assessment
19 Previous recidivism studies conducted by the North Carolina Sentencing and Policy Advisory Commission have
used a measure of risk in the analysis, and found that many of the differences between programs diminished when
risk was controlled for (Clarke and Harrison, 1992; NC Sentencing and Policy Advisory Commission, 1998; 2000;
2002; 2004; 2006; 2008). See Appendix D for a more in-depth discussion of how the risk score was developed for
this study.
20 Of the FY 2005/06 sample, 51.2% were black, 43.1% were white, and the remaining 5.7% were American Indian
(1.7%), Asian (0.2%), other (3.6%), or unknown (0.2%). Based on this distribution, race was collapsed into two
categories, black and non-black. White, Asian and Indian offenders as well as offenders with an “other” or
“unknown” race were included in the non-black category.
2 0
2. Criminal History
► Age at first arrest
► Length of criminal history
► Number of prior arrests
► Number of prior drug arrests
► Most serious prior arrest
► Number of prior probationary sentences – probation, parole, post-release
supervision
► Number of prior revocations of probationary supervision
► Number of prior prison sentences
3. Current Sentence Information
► Offense class
► Maximum sentence length
A risk score was computed for all offenders in the sample based on these factors. 21 Risk
scores – not the sample of offenders – were divided into terciles. 22 The range of risk scores was
0.01 to 0.99; thus, “Low Risk” offenders had a score between 0.01 and 0.33; “Medium Risk”
offenders had a score between 0.34 and 0.66, and “High Risk” offenders had a risk score
between 0.67 and 0.99. Based on this methodology, 41.6% of the offenders were “Low Risk,”
48.0% were “Medium Risk,” and 10.4% were “High Risk.”
As shown in Figure 2.4, risk level varied by the type of punishment. Probationers
sentenced to a community punishment were much more likely to be low risk than intermediate
punishment probationers and either category of prison releases. For instance, only 23.6% of
prison releases without PRS and 30.2% of prison releases with PRS were low risk compared to
55.3% of probationers sentenced to a community punishment. Overall, prisoners were more
likely to be high risk than probationers. Looking at prison releases, those with no PRS were
more likely to be high risk (19.9%) and less likely to be low risk (23.6%) than those who
received PRS.
Profiles of offenders in each risk level by the type of punishment were examined in an
effort to understand why some “High Risk” offenders received community punishments and why
some “Low Risk” offenders received active sentences. This inspection revealed that these
offenders were sentenced to their respective punishment type based upon the Structured
Sentencing punishment charts. For example, an offender may have a low risk level, but have a
Class C or D felony as the current conviction; thus, the only sentencing option is an active
21 Risk scores are the probability that an offender will be rearrested during the follow-up period and are based on a
logistic regression model used to determine the impact of risk factors (including personal characteristics, criminal
history, and current sentence information) on recidivism.
22 Beginning with the Sentencing Commission’s 2008 Correctional Evaluation Report, a methodological change
was made in the categorization of the risk score into low, medium, and high risk groups. In reports prior to 2008,
the sample of offenders was divided into three groups of equal size according to their risk score, with the lowest
third as “Low Risk,” the middle third as “Medium Risk,” and the top third as “High Risk.” Recognizing that this
approach allowed the definition of low, medium, and high risk to shift slightly based on the distribution of risk
scores for different samples of offenders, an improvement was made in the method of grouping risk scores to
provide standardized definitions of low, medium, and high risk that do not change from sample to sample.
2 1
sentence. Conversely, an offender may have a high risk level but have a misdemeanor as the
current conviction. Although considered “High Risk,” community punishment is a sentencing
option for such offenders regardless of prior conviction level.
Figure 2.4
Offender Risk Level by Type of Punishment
55.3%
37.2%
23.6% 30.2%
40.2%
52.3%
56.5%
57.3%
4.5% 10.5%
19.9% 12.5%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Community
Punishment
Intermediate
Punishment
Prison Releases No
PRS
Prison Releases PRS
Low Risk Medium Risk High Risk
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program
Evaluation Data
Further analysis of the prison releases revealed a measurable difference between the risk
levels of these offenders by their type of prison entry (see Table 2.6). Of the three types of
prison entries, those revoked to prison were most likely to be low risk (28.0%) and least likely to
be high risk (13.5%). Conversely, those entering prison as a result of both a new crime and a
revocation (mixed entries), were least likely to be low risk (15.9%) and most likely to be high
risk (23.4%).
2 2
Table 2.6
Offender Risk Level for Prison Releases by Type of Entry
Risk Level
Prison Releases N %
Low
%
Medium
%
High
New Crime Entry 7,847 25.5 52.5 21.9
Technical Revocation Entry 7,406 28.0 58.4 13.5
Mixed Entry 4,480 15.9 60.6 23.4
TOTAL 19,733 24.3 56.6 19.1
Note: Percentages may not add to 100% due to rounding.
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
Summary
Chapter Two provided a description of the FY 2005/06 sample’s demographic
characteristics, prior criminal history, current conviction, and offender risk level. Of the 60,824
offenders placed on probation or released from prison in FY 2005/06, 79% were male, 51% were
black, 72% had at least one prior fingerprinted arrest, and 48% had a most serious current
conviction for a felony offense. Of the 41,091 probationers in the sample, 70% received a
community punishment and 30% received an intermediate punishment. Of the 19,733 prison
releases in the sample, 90% received no PRS and 10% received PRS.
As expected, fewer of the probationers had at least one prior arrest (63%) than did the
prison releases (90%). Among the prison releases group, prisoners with and without PRS had
similar percentages with at least one prior arrest. Among the probationers, significantly more of
those that received an intermediate punishment had a prior arrest (81%) than did those who
received a community punishment (56%). Looking at the most serious current conviction, most
of the probationers who received a community punishment were convicted of a misdemeanor
offense (82%), whereas most of the probationers who received an intermediate punishment were
convicted of a Class F through I felony offense (72%). Most of the prisoners released in FY
2005/06 with no PRS had a most serious current conviction for a Class F through I felony
offense (72%), whereas 100% of prisoners released with PRS had a most serious current
conviction for a Class B1 through E felony offense.
Offender risk level was found to increase by type of punishment, with community
punishment probationers having the lowest risk scores and prison releases having the highest risk
scores. However, prisoners released with no PRS had higher risk scores than prisoners released
with PRS. Chapter Three examines the sample’s subsequent criminal involvement, as measured
by rearrests, reconvictions, and reincarcerations.
2 3
CHAPTER THREE
CRIMINAL JUSTICE OUTCOME MEASURES FOR THE FY 2005/06 SAMPLE
Chapter Three examines the criminal justice outcome measures for the FY 2005/06
sample. Many of the tables in this chapter present information by probation or prison status for
individual categories of probationers and prisoners (also referred to as type of punishment) and
for the sample as a whole.23
Definition of the Follow-up Period and Time at Risk
Each offender in the FY 2005/06 sample was followed for a period of three years to
determine whether repeat criminal behavior occurred, with one-year, two-year, and three-year
recidivism rates reported.24 The three-year follow-up period was calculated on an individual
basis using the prison release date plus three years for prison releases and using the probation
entry date plus three years for probation entries. A fixed follow-up period was used in an
attempt to obtain the same “window of opportunity” for each offender to recidivate. In actuality,
the same window of opportunity was not necessarily available due to technical probation or PRS
revocations which result in incarceration or due to the commission of new crimes which result in
incarceration.25 Incarcerations resulting from technical revocations may reduce recidivist arrests
due to incapacitation since the offender no longer has the same amount of time in the community
to recidivate. As a result, offenders who were not rearrested during the follow-up may appear to
be a success but may have actually experienced another type of criminal justice failure (i.e.,
technical revocation and incarceration) during the follow-up period.
In order to take into account each offender’s window of opportunity to recidivate during
the follow-up period, each offender’s actual time at risk was calculated by identifying their
periods of incarceration in North Carolina’s prison system and by subtracting the length of time
incarcerated from the follow-up period. It is important to note that it was not possible to account
for time spent in county jails during the follow-up period since each of the State’s county jails
maintains its own data. In North Carolina, offenders who are sentenced to active terms of 90
days or less are incarcerated in county jail. Lack of automated statewide county jail data affects
the information presented in this chapter in two ways: 1) time incarcerated in county jails is not
subtracted from actual time at risk during the follow-up period and 2) incarceration in county
jails, either as a result of new sentences or technical revocations, is not included as part of the
recidivist incarceration measure.
Table 3.1 provides information on time at risk for offenders in the FY 2005/06 sample.
As expected, the percent of the sample at risk for the entire follow-up period declined across the
follow-up period. Overall, 87% of the FY 2005/06 sample were at risk for the entire one-year
23 Refer to Chapter Two, Figure 2.1 for a detailed explanation of the subgroups of probationers and prisoners.
24 Each follow-up period reported is inclusive of the previous follow-up periods, e.g., the two-year follow-up period
contains information on events that occurred during the first and second years of follow-up. As a result, the
recidivism rates reported for each follow-up period cannot be added across follow-up periods.
25 Technical revocations result from failure to comply with the conditions of probation or PRS (as opposed to a new
violation of the law), such as having positive drug tests, failing to attend court-ordered treatment, or violating
curfew.
2 4
follow-up period, 76% were at risk for the entire two-year follow-up period, and 70% were at
risk for the entire three-year follow-up period. While there was no difference between
probationers and prisoners with regards to the average time at risk for the one-year follow-up
period, differences between the two groups increased for the two- and three-year follow-up
periods, with prisoners being at risk fewer days than probationers (674 days compared to 687
days for the two-year follow-up and 990 days compared to 1,027 days for the three-year follow-up,
respectively). Of the four types of punishment, probationers with an intermediate
punishment had the lowest percentage of offenders who were at risk for the entire follow-up
period (i.e., had the entire window of opportunity to reoffend) and, correspondingly, were at risk
fewer days during follow-up.
Table 3.1
Percent at Risk and Average Time at Risk by Type of Punishment
Percent at Risk and Average Time at Risk
Type of Punishment N 1-Year
Follow-Up
(365 Days)
2-Year
Follow-Up
(730 Days)
3-Year
Follow-Up
(1,095 Days)
Probation Entries
Community Punishment 28,645 93%
358 days
85%
708 days
81%
1,058 days
Intermediate Punishment 12,446 74%
329 days
61%
638 days
54%
957 days
Subtotal 41,091 88%
349 days
78%
687 days
73%
1,027 days
Prison Releases
No Post-Release Supervision 17,724 87%
350 days
73%
674 days
64%
989 days
Post-Release Supervision 2,009 87%
347 days
78%
679 days
70%
1,002 days
Subtotal 19,733 87%
349 days
73%
674 days
64%
990 days
TOTAL 60,824 87%
349 days
76%
683 days
70%
1,015 days
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
2 5
Criminal Justice Outcome Measures
The Sentencing Commission uses rearrests as its primary measure of recidivism,
supplemented by information on reconvictions and reincarcerations, to assess the extent of an
offender’s repeat involvement in the criminal justice system. In the following sections, criminal
justice outcome measures are presented for the entire sample.
Recidivist Arrests26
Overall, 21.4% of the FY 2005/06 sample were rearrested during the one-year follow-up,
32.5% were rearrested during the two-year follow-up, and 40.1% were rearrested during the
three-year follow-up (see Table 3.2).27 Prisoners were more likely to be rearrested than
Table 3.2
Rearrest Rates by Type of Punishment
Rearrest Rates
Type of Punishment N 1-Year
Follow-Up
2-Year
Follow-Up
3-Year
Follow-Up
Probation Entries
Community Punishment 28,645 17.3 26.0 32.8
Intermediate Punishment 12,446 22.3 33.2 41.2
Subtotal 41,091 18.8 28.2 35.4
Prison Releases
No Post-Release Supervision 17,724 27.6 41.9 50.5
Post-Release Supervision 2,009 20.9 36.2 45.3
Subtotal 19,733 26.9 41.3 50.0
TOTAL 60,824 21.4 32.5 40.1
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
26 Fingerprinted arrest data from DOJ were used to determine recidivist arrests and convictions in North Carolina.
Recidivist arrests were defined as fingerprinted arrests that occurred after an offender was released from prison or
placed on probation for the conviction that placed him/her in the sample.
27 It must be noted that the rearrest rates reported in this section do not take into account the fact that some
offenders were not at risk for the entire follow-up period as a result of incarceration. It is possible to calculate
adjusted recidivism rates that estimate the rate of rearrest that would have occurred if every offender were at risk for
the entire follow-up period. For a comparison of rearrest rates with adjusted rearrest rates (i.e., rearrest rates that are
adjusted for time at risk), see the Commission’s 2004 recidivism report.
2 6
probationers, with a 50.0% rearrest rate for the three-year follow-up period. Of the four types of
punishment, probationers with a community punishment were the least likely to be rearrested
while prisoners with no PRS were the most likely to be arrested.
Community punishment probationers, intermediate punishment probationers, and
prisoners with PRS were subject to supervision in the community during the three-year follow-up
period.28 For these three groups, the timing of rearrest was examined to determine whether
the offender was rearrested while on probation or PRS for the conviction that placed him/her in
the FY 2005/06 sample. Of those rearrested during follow-up, 51.4% of community punishment
probationers, 57.0% of intermediate punishment probationers, and 35.4% of prisoners with PRS
were rearrested while on probation or PRS. The higher percentage of intermediate punishment
probationers with a rearrest while on supervision compared to the other two groups is consistent
with their longer term of supervision.
Overall, prisoners who were rearrested during follow-up had a higher number of rearrests
than probationers who were rearrested (see Figure 3.1). For example, 43% of prisoners
compared to 53% of probationers had only one rearrest, while 9% of prisoners compared to 6%
of probationers had five or more rearrests. For those who were rearrested during the three-year
follow-up period, their first rearrest occurred an average of 13.2 months after entry to probation
or release from prison. There were slight variations in the time to first rearrest among the four
groups. The average number of months to rearrest was 13.5 for community punishment
probationers, 13.0 for intermediate punishment probationers, 12.9 for prisoners with no PRS, and
14.4 for prisoners with PRS.
Figure 3.1
Number of Rearrests for Offenders with Any Rearrest
Probation Entries
(# with Any Rearrest = 14,528)
Prison Releases
(# with Any Rearrest = 9,856)
53%
6%
17%
24%
1 2 3-4 5+
26%
22%
9%
43%
1 2 3-4 5+
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
28 Under current law, although there are some exceptions, community punishment probationers receive a probation
sentence of not less than 12 and not more than 30 months, intermediate punishment probationers receive a probation
sentence of not less than 18 and not more than 36 months, and prisoners released from prison onto PRS are
supervised for 9 months. Of the probation entries in the FY 2005/06 sample, probationers receiving a community
punishment were sentenced to an average of 18 months of supervised probation while probationers receiving an
intermediate punishment were sentenced to an average of 29 months.
2 7
For the prison releases in the sample, there was also considerable variation in recidivism
rates by the type of entry to prison (see Table 3.3). Offenders who had entered prison both due
to a technical probation revocation and a new crime (referred to as a “mixed entry”) had a
rearrest rate of 56.9%, considerably higher than the rate for prison entries due to a technical
revocation of probation (45.5%) and the rate for prison entries due to a new crime (50.2%).
As noted in Chapter Two, 21.9% of new crime entries, 13.5% of technical revocation
entries, and 23.4% of the mixed entries were high risk. Table 3.3 also examines the relationship
between type of entry and rearrest while controlling for offender risk. While recidivism is
strongly correlated with risk levels, the recidivism rates of low and medium risk offenders in the
mixed group are notably higher than the low and medium risk offenders in the other two groups.
Table 3.3
Rearrest Rates by Risk Level for Prison Releases by Type of Entry
% Rearrest
During the Three-Year Follow-Up Period
By Offender Risk Level
Prison Releases by
Type of Prison Entry N
Overall
Low Medium High
New Crime Entry 7,847 50.2 21.4 52.9 77.4
Technical Revocation Entry 7,406 45.5 24.6 49.2 72.5
Mixed Entry 4,480 56.9 32.2 55.7 76.7
TOTAL 19,733 50.0 24.4 52.2 75.9
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
Table 3.4 provides information on the actual number of arrests for those who were
rearrested during the follow-up period, as well as the types of crimes for which they were
rearrested. The 24,384 offenders who were rearrested during the three-year follow-up accounted
for a total of 50,997 arrests during this period, with 10,500 arrests for violent offenses, 21,827
arrests for property offenses, 15,183 arrests for drug offenses, and 16,552 arrests for “other”
offenses.29 While probationers were less likely to be rearrested than prisoners, they accounted
for a higher volume of arrests due to the larger number of probation entries in the FY 2005/06
sample.
29 See Appendix A for information on the categorization of offenses as person, property, drug, and other.
Table 3.4
Rearrests by Type of Punishment and Crime Type
Total Number and Average Number of Arrests
During the Three-Year Follow-Up Period
Type of Punishment Overall Violent Property Drug Other
# with
Any
Rearrest
# Avg. # Avg. # Avg. # Avg. # Avg.
Probation Entries
Community Punishment 9,398 18,431 2.0 3,836 0.4 7,734 0.8 5,442 0.6 5,870 0.6
Intermediate Punishment 5,130 10,196 2.0 2,097 0.4 4,176 0.8 3,200 0.6 3,449 0.7
Subtotal 14,528 28,627 2.0 5,933 0.4 11,910 0.8 8,642 0.6 9,319 0.6
Prison Releases
No Post-Release Supervision 8,947 20,365 2.3 4,011 0.4 9,260 1.0 5,907 0.7 6,416 0.7
Post-Release Supervision 909 2,005 2.2 556 0.6 657 0.7 634 0.7 817 0.9
Subtotal 9,856 22,370 2.3 4,567 0.5 9,917 1.0 6,541 0.7 7,233 0.7
TOTAL 24,384 50,997 2.1 10,500 0.4 21,827 0.9 15,183 0.6 16,552 0.7
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
2 9
Note: Of the FY 2005/06 sample, 41.6% (n=25,271) were low risk,
48.0% (n=29,198) were medium risk, and 10.4% (n=6,355) were high
risk.
SOURCE: NC Sentencing and Policy Advisory Commission,
FY 2005/06 Correctional Program Evaluation Data
Figure 3.2
Rearrest Rates by Offender Risk Level:
Three-Year Follow-Up
74.9
48.9
21.2
0%
10%
20%
30%
40%
50%
60%
70%
80%
Low Risk Medium Risk High Risk
Table 3.4 also includes information on the average number of rearrests by offense type
for each group. The average number of arrests for those who were rearrested was 2.1 for the
three-year follow-up, with prisoners having a slightly higher average number of rearrests (2.3)
than probationers (2.0).
Based on the composite risk measure described in Chapter Two, 41.6% of the FY
2005/06 sample were low risk, 48.0% were medium risk, and 10.4% were high risk.30 As
expected, recidivism rates varied considerably by offender risk level, with a stair-step increase in
the percentage rearrested from low risk to medium risk to high risk (see Figure 3.2). High risk
offenders had a rearrest rate of 74.9% during the three-year follow-up period – over three and
one-half times higher than the rearrest rate of low risk offenders (21.2%).
The relationship between recidivism and type of punishment has been demonstrated in
Table 3.2. Figure 3.3 illustrates the relationship between type of punishment and rearrest during
the three-year follow-up period when controlling for risk level. In general, the stair-step pattern
in rearrest rates found by
offender risk level for the
entire sample was also found
when examining rearrest
rates by type of punishment
and controlling for risk level.
However, once risk level is
controlled, the differences in
rearrest rates between
offenders in the different
punishment categories are
diminished. For the three-year
follow-up period,
rearrest rates for low risk
offenders ranged from 18.8%
for prisoners with PRS to
25.2% for prison releases
with no PRS, while rearrest
rates for high risk offenders
ranged from 72.4% for
probationers with an
intermediate punishment to
76.5% for prison releases
with PRS over the three-year
follow-up period.
30 See Chapter Two for detailed information on the measure of offender risk level.
3 0
Figure 3.3
Rearrest Rates by Type of Punishment and Risk Level: Three-Year Follow-Up
74.8
46.1
19.7
72.4
48.0
22.9
75.8
52.1
25.2
76.5
52.3
18.8
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Low Risk Medium Risk High Risk
Community Punishment Intermediate Punishment Prison Releases No PRS Prison Releases PRS
Note: Of the FY 2005/06 sample, 41.6% (n=25,271) were low risk, 48.0% (n=29,198) were medium risk, and
10.4% (n=6,355) were high risk.
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
Table 3.5 provides information on rearrest rates by offender risk level and on the number
of rearrests for low, medium, and high risk offenders who were rearrested during follow-up.
While only 10% of offenders were high risk, they accounted for 26% (n=13,443) of the 50,997
recidivist arrests for the sample. As expected, the average number of arrests was also lowest for
low risk offenders and highest for high risk offenders with an average of 1.6 and 2.8 arrests,
respectively.
Offender risk level and recidivism were also examined in relation to offense class for the
most serious current conviction (see Table 3.6). In general, felons had higher risk levels than
misdemeanants. Compared to the other offense class groupings, offenders with a most serious
current conviction (hereinafter referred to as “conviction”) for a Class A1 through Class 3
misdemeanor offense had the lowest percentage of offenders who were high risk (7.0%). Among
felons, offenders with a conviction for a Class B1 through Class E felony (which are defined as
violent offenses under Structured Sentencing) had a lower percentage of high risk offenders than
those offenders with a conviction for a Class F through Class I felony.
Table 3.5
Rearrests by Risk Level
Rearrests for Offenders with Any Rearrest
(n=24,384)
Risk Level N Number of Rearrests (%)
% with
Any
Rearrest
Total
# of
Rearrests
Average
# of
Rearrests 1 2 3-4 5+
Low Risk 25,271 21.2 8,434 1.6 65.7 20.9 10.9 2.5
Medium Risk 29,198 48.9 29,120 2.0 48.9 25.5 18.9 6.7
High Risk 6,355 74.9 13,443 2.8 30.8 25.0 28.5 15.7
TOTAL 60,824 40.1 50,997 2.1 49.0 24.4 19.0 7.6
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
Table 3.6
Offender Risk Level and Rearrest Rates by Offense Class for Most Serious Current Conviction
% Rearrest
% Offender Risk Level During the Three-Year Follow-Up Period
By Offender Risk Level
Most Serious
Current Conviction N
Low Medium High
Overall
Low Medium High
Offense Class
Class B1 – E Felony 2,505 33.8 55.2 11.0 42.2 18.6 50.0 75.3
Class F – I Felony 26,685 32.4 53.1 14.5 44.4 22.0 49.5 75.5
Class A1 – 3 Misdemeanor 31,317 49.9 43.1 7.0 36.4 21.0 48.1 74.2
Specific Groups of Interest
Habitual Felons 376 8.8 66.7 24.5 56.4 30.3 56.2 66.3
Sex Offenders 1,309 57.9 38.4 3.7 25.3 15.0 36.6 70.8
TOTAL 60,507 41.5 48.0 10.5 40.2 21.3 48.9 75.0
Note: Of the 60,824 offenders in the FY 2005/06 sample, 317 were excluded from this table due to discrepant or unknown offense classes. Due to the length of
sentences imposed for Class B1 felonies, there were no prisoners released in the FY 2005/06 sample with a most serious conviction for a Class B1 felony. The
sex offenders grouped in this table are offenders who are required to register as sex offenders under Article 27A of Chapter 14 of the NC General Statutes.
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
3 3
Finally, almost 15% of offenders with a conviction for a Class F through Class I felony offense
were high risk – the highest of all offense groupings.
Overall, 42.2% of offenders with a conviction for a Class B1 through Class E felony,
44.4% of offenders with a conviction for a Class F through Class I felony, and 36.4% of
offenders with a conviction for a Class A1 through Class 3 misdemeanor were rearrested during
the three-year follow-up period. The offender risk level distribution, which is defined as risk of
rearrest, and the rearrest rate for offenders convicted of Class B1 through Class E felony indicate
that offenders convicted of a violent offense are less likely to reoffend than those convicted of
non-violent felony offense (primarily a property or drug offense).
The relationship between offender risk level and type of punishment (as shown in Figure
3.3) was also found when examining offender risk level and offense class (see Table 3.6). When
controlling for offender risk level, the differences in rearrest rates between offenders in the
different class groupings were diminished. For the three-year follow-up period, rearrest rates for
each offense class grouping ranged from 18.6% to 22.0% for low risk offenders, from 48.1% to
50.0% for medium risk offenders, and from 74.2% to 75.5% for high risk offenders. From this
analysis, it appears that offender risk level nearly negates the link between offense class and
rearrest.
Represented within Class B1 through Class E convictions is a special group of offenders
– habitual felons. An habitual felon is an offender with at least three prior felony convictions
(each conviction having occurred before he or she committed the next offense) who has currently
been convicted of a felony offense and who has been found by a jury to be an habitual felon.
(N.C.G.S '' 14-7.1 to -7.6) While habitual felons are sentenced as Class C felons, the
overwhelming majority of habitual felons have a Class F through Class I felony as their most
serious underlying conviction.31
In order to examine whether habitual felons were more similar to offenders with a
conviction for a Class B1 through Class E felony or to offenders with a conviction for a Class F
through Class I felony, their distribution by offender risk level and rearrest rates were examined.
Of the 376 habitual felons released from prison in FY 2005/06, 8.8% were low risk, 66.7% were
medium risk, and 24.5% were high risk (see Table 3.6). During the three-year follow-up period,
56.4% of habitual felons were rearrested. Based on these statistics, habitual felons were more
similar to offenders with Class F through Class I felony convictions than to offenders with Class
B1 through Class E felony convictions with respect to their distribution by risk level and rearrest
rates.
Offenders who are required to register as sex offenders under Article 27A of Chapter 14
of the NC General Statutes are also a group of special interest. Those convicted of a reportable
offense are required to register as sex offenders. A reportable offense is defined as “an offense
31 According to the NC Sentencing and Policy Advisory Commission’s annual statistical report, there were 808
habitual felon convictions in FY 2008/09 (NC Sentencing and Policy Advisory Commission, 2010). Overall, 90%
(n=728) had a conviction for a Class F through Class I felony as their most serious underlying conviction, with Class
F accounting for 7.5%, Class G for 22.9%, Class H for 44.2%, and Class I for 15.4%.
3 4
against a minor, a sexually violent offense, or an attempt to commit” such offenses.32 Of the
1,309 offenders in the sample convicted of an offense for which they are required to register as a
sex offender, 17.8% (n=233) were convicted of a Class B1 through Class E felony, 66.5%
(n=871) were convicted of a Class F through Class I felony, and the remainder were convicted of
a Class A1 misdemeanor or had a discrepant or unknown offense class (see Table 3.6). Almost
58% were low risk, 38.4% were medium risk, and 3.7% were high risk. Overall, 25.3% of the
offenders required to register as a sex offender had a recidivist arrest during the three-year
follow-up period. When compared to each offense class grouping, offenders required to register
as sex offenders had higher percentages determined to be low risk and lower percentages deemed
high risk. They also had lower overall rearrest rates. These findings are consistent with the risk
level distribution and rearrest rates found for male prison releases who participated in the Sex
Offender Accountability Responsibility (SOAR) program while in prison (see Appendix B).
Recidivist Convictions33
Overall, 9.3% of the FY 2005/06 sample had a reconviction during the one-year follow-up
period, 19.3% had a reconviction during the two-year follow-up period, and 26.6% had a
reconviction during the three-year follow-up period (see Table 3.7). Overall, prisoners had a
higher percentage of recidivist convictions than probationers. Almost 35% of prisoners had a
recidivist conviction during the three-year follow-up compared to 22.8% of probationers.
Intermediate punishment probationers had a higher percentage of recidivist convictions during
the three-year follow-up than community punishment probationers, with 27.1% of intermediate
punishment probationers having recidivist convictions compared to 20.9% of community
punishment probationers. Prisoners with no PRS had a higher percentage of recidivist
convictions during the three-year follow-up than prison releases with PRS, with 35.2% compared
to 29.2% respectively.
Table 3.8 provides information on the volume and types of recidivist convictions. The
16,197 offenders who had a recidivist conviction by the end of the three-year follow-up
accounted for 23,045 convictions during this period, with 3,682 convictions for violent offenses,
10,228 convictions for property offenses, 7,384 convictions for drug offenses, and 5,698
convictions for “other” offenses. While a lower percentage of probationers than prisoners had a
recidivist conviction, probationers accounted for a higher number of convictions than prisoners
due to the larger number of probation entries in the FY 2005/06 sample.
Table 3.8 also includes the average number of recidivist convictions for each group. The
average number of overall convictions for those with a recidivist conviction was 1.4 for the
three-year follow-up. Prisoners who were rearrested had a slightly higher average number of
recidivist convictions (1.5) than probationers (1.4). Overall, the average number of violent
convictions was 0.2 for those with a recidivist conviction during the three-year follow-up.
32 Offenses against a minor and sexually violent offenses are defined in N.C.G.S. ' 14-208.6.
33 Fingerprinted arrest data from the DOJ were used to determine recidivist arrests and convictions in North
Carolina. Recidivist convictions were defined as convictions for arrests that occurred after an offender was released
from prison or placed on probation for the conviction that placed him/her in the sample.
3 5
Table 3.7
Reconviction Rates by Type of Punishment
% Reconviction
Type of Punishment N 1-Year
Follow-Up
2-Year
Follow-Up
3-Year
Follow-Up
Probation Entries
Community Punishment 28,645 7.5 15.1 20.9
Intermediate Punishment 12,446 9.8 19.6 27.1
Subtotal 41,091 8.2 16.5 22.8
Prison Releases
No Post-Release Supervision 17,724 12.0 26.0 35.2
Post-Release Supervision 2,009 6.4 19.1 29.2
Subtotal 19,733 11.4 25.3 34.6
TOTAL 60,824 9.3 19.3 26.6
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
Recidivist conviction rates were also examined by offense class and by offender risk
level. Overall, 27.2% of offenders with a most serious current conviction for a Class B1 through
Class E felony, 30.1% of offenders with a conviction for a Class F through Class I felony, and
23.7% of offenders with a conviction for a Class A1 through Class 3 misdemeanor had a
recidivist conviction during the three-year follow-up period. As with rearrest rates, a stair-step
pattern was found in recidivist conviction rates by offender risk level, with 12.3% of low risk
offenders, 32.9% of medium risk offenders, and 54.7% of high risk offenders having a recidivist
conviction during the three-year follow-up period.
For offenders who had a recidivist conviction during the three-year follow-up period,
their first recidivist conviction occurred an average of 17.3 months after entry to probation or
release from prison. Among the four groups, prison releases with PRS had a slightly longer time
to reconviction (19.7 months). The average number of months to reconviction for community
punishment probationers, intermediate punishment probationers, and for prison releases with no
PRS was 17.2 months.
Table 3.8
Reconvictions by Type of Punishment and Crime Type
Total Number and Average Number of Convictions
During the Three-Year Follow-Up Period
Type of Punishment Overall Violent Property Drug Other
# with
Any
Conv.
# Avg. # Avg. # Avg. # Avg. # Avg.
Probation Entries
Community Punishment 5,999 8,441 1.4 1,310 0.2 3,654 0.6 2,721 0.5 2,049 0.3
Intermediate Punishment 3,372 4,558 1.4 708 0.2 1,940 0.6 1,520 0.5 1,169 0.3
Subtotal 9,371 12,999 1.4 2,018 0.2 5,594 0.6 4,241 0.5 3,218 0.3
Prison Releases
No Post-Release Supervision 6,240 9,213 1.5 1,469 0.2 4,375 0.7 2,862 0.5 2,217 0.4
Post-Release Supervision 586 833 1.4 195 0.3 259 0.4 281 0.5 263 0.4
Subtotal 6,826 10,046 1.5 1,664 0.2 4,634 0.7 3,143 0.5 2,480 0.4
TOTAL 16,197 23,045 1.4 3,682 0.2 10,228 0.6 7,384 0.5 5,698 0.4
SOURCE: NC Sentencing and Policy Advisory Commission, FY 2005/06 Correctional Program Evaluation Data
3 7
Recidivist Incarcerations34
Of the FY 2005/06 sample, 12.6% had a recidivist incarceration during the one-year
follow-up period, 23.6% had a recidivist incarceration during the two-year follow-up period, and
30.1% had a recidivist incarceration during the three-year follow-up period (as shown in Table
3.9). Recidivist incarcerations may have occurred as a result of the sentence imposed for a new
crime committed or due to a technical revocation during the follow-up period. Overall, prisoners
were more likely to have a recidivist incarceration than probationers, with a 35.8% incarceration
rate at the end of the three-year follow-up compared to 27.3% of probationers. Of the four
groups, probationers with community punishments had the lowest incarceration rate during the
follow-up period and probationers with intermediate punishments had the highest incarceration
rate during the follow-up period. The high reincarceration rate for this group is most likely
linked to their high technical revocation rate. For prison releases, those with no PRS had a
higher rate of reincarceration at 36.5% during the follow-up period compared to prisoners with
PRS at 29.8%. Of those offenders with an incarceration during the three-year follow-up period,
81.9% had one incarceration, 15.9% had two incarcerations, and 2.2% had three or more
incarcerations.
Recidivist incarceration rates were also examined by offense class and by offender risk
level. Overall, 32.1% of offenders with a most serious current conviction for a Class B1 through
Class E felony, 38.9% of offenders with a conviction for a Class F through Class I felony, and
22.5% of offenders with a conviction for a Class A1 through Class 3 misdemeanor had a
recidivist incarceration during the three-year follow-up period. It is not surprising that offenders
with Class F through Class I felony convictions had higher reincarceration rates than those with
Class B1 through Class E convictions. While offenders with Class B1 through Class E felony
convictions are more likely to be in the FY 2005/06 sample as a prison release, offenders with
Class F through I felony convictions are more likely to be in the sample as a result of a
probationary sentence. Correspondingly, their higher reincarceration rates may be a function of
technical revocations that result in incarceration, in addition to recidivist arrests that lead to
reincarceration. As with the other measures of recidivism, a stair-step pattern was found in
recidivist incarceration rates by offender risk level, with 15.6% of low risk offenders, 36.6% of
medium risk offenders, and 57.8% of high risk offenders having a recidivist incarceration during
the three-year follow-up period.
For offenders who had an incarceration during the three-year follow-up period, their first
incarceration occurred an average of 15.6 months after entry to probation or release from prison.
The average number of months to incarceration was 16.8 for community punishment
probationers, 12.6 for intermediate punishment probationers, 17.0 for prison releases with no
PRS, and 16.2 for prison releases with PRS.
34 DOC’s OPUS data were used to determine recidivist incarcerations (i.e., incarcerations that occurred during the
follow-up period). It must be noted that the data presented on recidivist incarcerations only include incarcerati