2009 Florida Prison Recidivism Study
Releases From 2001 to 2008
May 2010

Florida Department of Corrections
Walter A. McNeil, Secretary
Bureau of Research and Data Analysis
dcresearch@mail.dc.state.fl.us

May 2010

2009 Florida Prison Recidivism Study

Page 1 of 18

INTRODUCTION
The 2009 Florida Prison Recidivism Study is the first report to be produced annually
that examines the issue of recidivism among Florida's released inmate population. The
use of recidivism as a performance indicator of the state's rehabilitative efforts can be
debated, but the analysis itself is of vital public importance. Basically, what is the
likelihood that an inmate who is released today will come back to prison? This question
is important for the state in terms of planning and budgeting, but more importantly to
the public and elected officials in terms of public safety. Since 88% of inmates in
Florida's prisons today will one day be released back into our communities, their success
or failure comes at a cost to public order and public safety.
The next pertinent question is, What factors influence recidivism rates? Do age, gender
and racial groups show differences in recidivism rates? The answers to these questions
are also important, as they allow one to identify groups most likely to fail when they are
released. This can be used to determine where to devote scarce correctional and
community resources.
This study finds that results for Florida are generally consistent with existing research of
the factors that influence recidivism. The Bureau of Justice Statistics report,
"Recidivism of Prisoners Released in 1994 (2002)" shows overall recidivism rates for
releases from 15 different states. That report shows a 51.8% recidivism rate (return to
prison for any reason within three years) for this group of inmates.
Unlike most states, Florida paroles very few inmates and only about a third of released
inmates have any community supervision sanction at all. Since those with supervision
after release recidivate more often than those without supervision upon release, it is
important to keep in mind that Florida's recidivism rate may be lower than another state
due to this difference in release mechanisms.

May 2010

2009 Florida Prison Recidivism Study

Page 2 of 18

TABLE OF CONTENTS

Introduction................................................................................................................... 2
Executive Summary ...................................................................................................... 4
Methodology .................................................................................................................. 5
Recidivism Rates Over Time ........................................................................................ 7
Recidivism and Gender ................................................................................................ 8
Recidivism and Violent Offenses ................................................................................. 9
Recidivism and Non-violent Offenses ....................................................................... 10
Recidivism and Age ...................................................................................................... 11
Model Factor Tables.................................................................................................... 12
Factors in Order of Predictive Ability ........................................................................ 16
References.................................................................................................................... 18

May 2010

2009 Florida Prison Recidivism Study

Page 3 of 18

EXECUTIVE SUMMARY

For this report, recidivism is defined as a return to prison, either because of a new
conviction or a violation of post prison supervision. The follow-up periods (typically
reported as three years) are calculated from prison release date to the date of
readmission to prison.
The overall three-year recidivism rate based on all released inmates from 2001 to 2008
is 33.1%. Specifically, the last three years have shown a slight decrease in the threeyear recidivism rate.
Inmates released in 2003 → 33.9% recidivism rate
Inmates released in 2004 → 33.4% recidivism rate
Inmates released in 2005 → 32.7% recidivism rate
Note that a one percentage point drop in the recidivism rate results in approximately
400 fewer inmates being admitted over a three-year period at a cost of $20,000 per year
per inmate or a cost avoidance of approximately $8,000,000.
The factors that influence an inmate's likelihood of recidivism include:
 prior prison commitments (more priors → higher recidivism);
 whether the inmate has a supervision term after release (supervised → higher
recidivism);
 their age at release (younger → higher recidivism);
 their behavior while in prison (more disciplinary reports → higher recidivism);
 their tested education level (higher grade level → lower recidivism); and
 number of theft/fraud offenses in criminal history (more offenses → higher
recidivism)
Inmates who complete education programs while in prison have lower recidivism rates
than inmates who do not complete programs. Note that this conclusion does not take
into account any other differences in these two groups and should not be given the
weight of a rigorous program evaluation.

May 2010

2009 Florida Prison Recidivism Study

Page 4 of 18

METHODOLOGY
Inmates released from Florida prisons from January 1, 2001 to December 31, 2008 were
included in the study with the following exceptions:
• Inmates who died, were executed, had their sentence vacated, or were released to
another state have been omitted from calculation of recidivism rates.
• Inmates who are missing information on the factors of interest are omitted.
These restrictions reduce the number of records analyzed by approximately 29%.
Survival analysis techniques are used to compute recidivism rates and to define the
statistical models used to determine which factors significantly influence recidivism
rates. The basic rates for tables and graphs are computed from Kaplan-Meier estimates
of the survival curve using right-censored data. The analyses of factor significance are
conducted using Cox models (proportional hazards regression) of the same data. The
analysis used a 5% level of significance and, to determine the factors in order of
importance, a stepwise selection routine for determining which factors to include.
Twenty-one (21) factors considered:
Gender is Male – Yes/No
Number of Prior Prison Commitments
Age at release
Number of Disciplinary Reports in Current Incarceration
Most Recent TABE score (education level in grade equivalents)
Time Served in Prison, Current Incarceration in Months
Worst Offense is Murder/Manslaughter – Yes/No
Worst Offense is Sex Offense – Yes/No
Worst Offense is Robbery – Yes/No
Worst Offense is Other Violent Offense like Assault or Kidnapping – Yes/No
Race is Black – Yes/No
Ethnicity is Hispanic – Yes/No
Supervision to Follow Prison – Yes/No
Low Custody (Minimum or Community Custody) – Yes/No
High Custody (Close Custody) – Yes/No
Number of Burglary offenses in criminal history
Number of Drug offenses in criminal history
Number of Theft/Fraud offenses in criminal history
Number of Weapons offenses in criminal history
Diagnosed Mental Illness – Yes/No
Substance Abuse Severity Score
For the "Worst Offense" factors the hierarchy is as follows:
Murder, Sex Offense, Robbery, Other Violent Offense
Each inmate can only be designated in at most one of the categories. If he has
committed both Sex Offenses and Robbery, he will be considered in the "Worst Offense
is Sex Offense" category, not in the "Worst Offense is Robbery" category.

May 2010

2009 Florida Prison Recidivism Study

Page 5 of 18

For the criminal history factors, only those offenses for which the inmate received either
a Florida prison sentence or Florida community supervision sentence are considered.
Arrests in Florida that resulted in other sanctions, such as fines or county jail sentences
are not considered. Arrests, supervision, or prison sentences outside of Florida are not
considered unless they are part of the inmate's Florida sentence.

May 2010

2009 Florida Prison Recidivism Study

Page 6 of 18

RECIDIVISM RATES OVER TIME
Three year recidivism rate by year of release
40.0%

38.0%

36.0%
33.9%

34.0%

Recidivism rate

32.6%

33.4%

33.2%

32.7%

32.0%

30.0%

28.0%

26.0%

24.0%

22.0%

20.0%
2001

2002

2003

2004

2005

Year of Release from Prison

The range of recidivism rates when examined one year of releases at a time, range from
32.6% to 33.9% in the five years for which three-year rates can be computed.
Recidivism rates are certainly affected by factors outside the influence of the Department of
Corrections, such as unemployment, crime rates, and local criminal justice issues such as
jail bed availability and judicial behavior. Statewide initiatives such as truth-in-sentencing,
increased use of mandatory prison terms and inconsistent funding for inmate rehabilitative
programs may also influence recidivism rates. For this reason, recidivism rates cannot be
used as the only measure of operational performance for the prison system. It is a measure
of a multitude of societal issues working for and against the released inmate, before he ever
gets sentenced to prison and after he is released.
The slight upward increase in 2002 and 2003 that is followed by decreases in 2004 and
2005 may be due to an overall increase in revocations of supervision that peaked in 2005
and subsequently declined to previous levels. This trend would affect the 2003 release
cohort more than the subsequent years of releases.
Note that a one percentage point drop in the recidivism rate results in approximately 400
fewer inmates being admitted over a three-year period. Considering that it costs taxpayers
almost $20,000 per year for each inmate incarcerated, even a relatively small decrease in
recidivism rates that persists over multiple years can result in millions of taxpayer dollars to
be used for other priorities.

May 2010

2009 Florida Prison Recidivism Study

Page 7 of 18

RECIDIVISM AND GENDER
Recidivism, gender
70.0%
All

Male

Female
60.0%

Recidivism rate

50.0%

40.0%

30.0%

20.0%

10.0%

99

102

96

93

90

87

84

81

78

75

72

69

66

63

60

57

54

51

48

45

42

39

36

33

30

27

24

21

18

15

9

12

6

3

0

0.0%

Months since release

Female inmates recidivate at a much lower rate than male inmates. In fact, it takes six
years for a female inmate to recidivate at the same rate as a male inmate does in three
years (female recidivism at six years from release is 33.2%; male recidivism rate at three
years from release is 34.7%)

May 2010

2009 Florida Prison Recidivism Study

Page 8 of 18

RECIDIVISM AND VIOLENT OFFENSES
Recidivism, current offense
70.0%
Murder/Manslaughter

Robbery

Other Violent Offense

Sex Offense

All

60.0%

Recidivism rate

50.0%

40.0%

30.0%

20.0%

10.0%

99

93
96

87
90

81
84

75
78

69
72

63
66

57
60

51
54

45
48

39
42

33
36

27
30

21
24

15
18

6

9
12

3

0

0.0%

Months since release

The graph above shows that among inmates who were in prison for violent offenses,
those in for murder or manslaughter have the lowest recidivism rates. Inmates in for
robbery offenses have higher than average recidivism rates.

May 2010

2009 Florida Prison Recidivism Study

Page 9 of 18

RECIDIVISM AND NON-VIOLENT OFFENSES
Recidivism, current offense
70.0%

60.0%
Burglary

Theft/Fraud

Weapons

All

Drugs

Recidivism rate

50.0%

40.0%

30.0%

20.0%

10.0%

99

93
96

87
90

81
84

75
78

69
72

63
66

57
60

51
54

45
48

39
42

33
36

27
30

21
24

15
18

6

9
12

3

0

0.0%

Months since release

The graph above shows that among inmates who were in prison for non-violent
offenses, those in for weapons offenses have the lowest recidivism rates. Burglars
released during this period have the highest recidivism rates. Drug offenders recidivate
at a rate similar to the overall recidivism rate for all inmates.

May 2010

2009 Florida Prison Recidivism Study

Page 10 of 18

RECIDIVISM AND AGE

Recidivism, age
70.0%
Under 25

25 to 34

35 to 49

50 to 64

65+

60.0%

Recidivism rate

50.0%

40.0%

30.0%

20.0%

10.0%

99

102

96

93

90

87

84

81

78

75

72

69

66

63

60

57

54

51

48

45

42

39

36

33

30

27

24

21

18

15

9

12

6

3

0

0.0%

Months since release

The older an inmate is at time of release, the less likely he is to return to prison. Note
that while this relationship is apparent several months after release from prison, there is
quite a bit of overlap in the first three years after release. The groups of inmates 49 and
younger recidivate at similar rates until two to three years after release, at which point
the curves shown above begin to separate.

May 2010

2009 Florida Prison Recidivism Study

Page 11 of 18

Model Factor Tables

CATEGORICAL
FACTORS
GENDER
RACE
ETHNICITY
CUSTODY
MOST
SERIOUS
CRIME IN
INMATE
HISTORY

SUPERVISION
TO FOLLOW

VALUE

N

Female
Male
Black
Non-Black
Hispanic
Non-Hispanic
Low
High

21,220
157,459
92,655
86,024
10,836
167,843
92,690
22,594

12%
88%
52%
48%
6%
94%
52%
13%

20.6%
34.7%
37.7%
28.1%
25.0%
33.6%
28.3%
42.9%

5,791
10,726
27,277
47,833

3%
6%
15%
27%

25.0%
33.8%
40.8%
33.0%

0.680**
0.939**
0.963**
0.923**

60,759
117,920

34%
66%

44.2%
27.4%

1.939**

Murder /
Manslaughter
Sex Offense
Robbery
Other Violent
Offense such
as assault
and
kidnapping
Yes
No

% OF
RECIDIVISM Hazard
RELEASE
RATE
Ratio
COHORT

DIAGNOSED
MENTAL
ILLNESS

Yes
24,177
14%
33.7%
No
154,502
86%
33.0%
**: p-value ≤ 0.01; *: 0.01 < p-value ≤0.05; NS: Not Significant at α = 0.05

May 2010

2009 Florida Prison Recidivism Study

1.506**
1.155**
0.752**
0.807**
1.070**

1.023
NS

Page 12 of 18

NUMERIC
FACTORS
CRIMINAL
HISTORY

VALUE

MEDIAN

MEAN

Number of
Prior Prison
Commitments
0
0.5
Number of
Burglary
Offenses
0
1.0
Number of
Theft Offenses
1
1.9
Number of
Drug Offenses
1
1.9
Number of
Weapons
Offenses
0
0.2
Score from
SUBSTANCE
substance
ABUSE NEED
abuse
assessment
11
21.4
Months served
TIME SERVED
during current
incarceration
21
32.8
Number
of
INSTITUTIONAL
Disciplinary
BEHAVIOR
Reports during
current
incarceration
0
2.5
Most recent
EDUCATION
TABE score
LEVEL
(grade
equivalents)
6.8
7.3
Age at release
AGE
33
34.2
**: p-value ≤ 0.01; *: 0.01 < p-value ≤0.05; NS: Not Significant at α = 0.05

Hazard
Ratio
1.425**

1.034**
1.012**
1.019**
0.965**

1.002**

1.001**

1.010**

0.963**
0.978**

Note that hazard ratios in the above tables are interpreted as the multiple of the
likelihood of failure. For example, male inmates have a hazard ratio of 1.506. Since it is
greater than one, it means that a male inmate is (1.506-1=0.506) 50.6% more likely to
fail than a female inmate with all other factors held constant (they are identical on all
factors in the model except for gender).
On the other hand, if the hazard ratio is less than one, the interpretation is a percent
reduction in likelihood to fail. For example, an Hispanic inmate is (1-0.752=.248)
24.8% less likely to recidivate than a non-Hispanic inmate with all other factors held
constant.

May 2010

2009 Florida Prison Recidivism Study

Page 13 of 18

Those measures that are expressed as numeric counts instead of dichotomous (Yes/No),
the hazard ratios show the increase or decrease PER UNIT INCREASE in the factor. For
example, for each additional disciplinary report that an inmate incurs while
incarcerated, his likelihood of recidivating increases by (1.010-1=.010) 1.0%. For each
additional grade level tested, his likelihood of recidivating decreases by (1-0.963=0.037)
3.7%.

Factors Not Included in the Model
FACTOR
RELEASE
TYPE

RELEASE
FACILITY

VALUE
Expiration of
Sentence (No
Supervision to
Follow)
Conditional Release
(mandatory
supervision for
serious offenders as
specified in F.S.
947.1405)
Expiration of
Sentence to
Probation or
Community Control
(Split Sentence)
Major Correctional
Institution
Private Correctional
Institution
Work/Forestry Camp
Work Release Center
Contracted Transition
and Work Release
Center

PROGRAM
COMPLETERS GED
Vocational Certificate

N

%

RECIDIVISM
RATE

117,868

66%

26.3%

22,468

13%

57.3%

30,594

17%

39.3%

120,233

67%

36.2%

11,780
16,622
25,565

7%
9%
14%

29.9%
34.7%
23.0%

3,242

2%

17.9%

9,227
13,141

5%
7%

29.0%
28.1%

Note that the results presented in the table above do not constitute a rigorous evaluation
of any program. For example, it is not valid to claim that if all inmates went to Work
Release Centers, one would realize a lower overall recidivism rate. Inmates who succeed
May 2010

2009 Florida Prison Recidivism Study

Page 14 of 18

at work release do not have the same characteristics as many other inmates, so the work
release impact would likely be different if expanded to include a broader group of
inmates.
Likewise, the program completion information presented in the table does not
constitute a rigorous evaluation of those programs. The results are similar to those
reported in the Washington State Institute for Public Policy study of evidence-based
practices in corrections, but are based solely on Florida Department of Corrections data.

May 2010

2009 Florida Prison Recidivism Study

Page 15 of 18

FACTORS IN ORDER OF PREDICTIVE ABILITY
Below are listed the factors in order of importance (YELLOW highlight: higher
recidivism; GREEN highlight: lower recidivism):
MALES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

PRIOR COMMITMENTS TO PRISON
SUPERVISION FOLLOWING PRISON
AGE AT RELEASE
DISCIPLINARY REPORTS WHILE IN PRISON
MOST RECENT TABE (EDUCATION LEVEL)
NUMBER OF THEFT/FRAUD OFFENSES
LOW CUSTODY
RACE IS BLACK
NUMBER OF BURGLARY OFFENSES
SUBSTANCE ABUSE SEVERITY SCORE
WORST OFFENSE IS MURDER
HISPANIC ETHNICITY
NUMBER OF DRUG OFFENSES
WORST OFFENSE IS OTHER VIOLENT CRIME
HIGH CUSTODY
NUMBER OF WEAPONS OFFENSES
TIME SERVED IN MONTHS
WORST OFFENSE IS SEX OFFENSE
WORST OFFENSE IS ROBBERY

FEMALES
1 PRIOR COMMITMENTS TO PRISON
2 SUPERVISION FOLLOWING PRISON
3 AGE AT RELEASE
4 SUBSTANCE ABUSE SEVERITY SCORE
5 MOST RECENT TABE (EDUCATION LEVEL)
6 NUMBER OF THEFT/FRAUD OFFENSES
7 LOW CUSTODY
8 NUMBER OF DRUG OFFENSES
9 DIAGNOSED MENTAL ILLNESS
10 HISPANIC ETHNICITY
11 WORST OFFENSE IS MURDER
12 WORST OFFENSE IS OTHER VIOLENT CRIME
13 NUMBER OF BURGLARY OFFENSES
14 DISCIPLINARY REPORTS WHILE IN PRISON
15 RACE IS BLACK
16 TIME SERVED IN MONTHS

May 2010

2009 Florida Prison Recidivism Study

Page 16 of 18

Factors that "change direction":
Race is Black (helps Females, hurts Males)
Months served in prison (helps Females, hurts Males)
Factors that affect Male recidivism, but not Females:
Number of Weapons Offenses
Worst Offense Committed is sex offense
Worst Offense Committed is robbery
High custody
Factors that affect Female recidivism, but not Males:
Diagnosed mental illness

May 2010

2009 Florida Prison Recidivism Study

Page 17 of 18

References
Connecticut Office of Policy & Management, Criminal Justice Policy and Planning Division; Research,
Analysis and Evaluation Unit. (2009). 2009 Connecticut Recidivism Study: 2004 Release Cohort.
Retrieved December 15, 2009 from
http://www.ct.gov/opm/lib/opm/cjppd/cjresearch/recidivismstudy/20090215_recidivismstudy.pdf
Langan, Patrick A. and Levin, David J. (2002). Recidivism of Prisoners Released in 1994. Washington,
DC: Bureau of Justice Statistics. Retrieved December 15, 2009 from
http://bjs.ojp.usdoj.gov/index.cfm?ty=pbdetail&iid=1134
Florida Department of Corrections. (2003). Recidivism Report: Inmates Released from Florida Prisons
July 1995 to June 2001. Retrieved December 15, 2009 from
http://www.dc.state.fl.us/pub/recidivism/2003/index.html
Drake, Elizabeth and Aos, Steve and Miller, Marna. (2009). Evidence-Based Public Policy Options to
Reduce Crime and Criminal Justice Costs: Implications in Washington State. Olympia, WA: Washington
State Institute for Public Policy. Retrieved December 15, 2009 from
http://www.wsipp.wa.gov/pub.asp?docid=09-00-1201

May 2010

2009 Florida Prison Recidivism Study

Page 18 of 18