CEPP-JRI Mecklenburg
Recidivism Analysis of a Sample of NC DOC Prisoners Released to Mecklenburg County - DRAFT
Applied Research Services, Inc.
February 7, 2013
Method
Hope L. Marshall, MS, LCAS, Enterprise Management Analyst for Criminal Justice Services in the
Mecklenburg County Manager’s Office, pulled a random sample of 400 North Carolina prison inmates
who were released to Mecklenburg County, checking for the presence of a re-arrest in the two years
following their release. Ms. Marshall provided a dataset that included those former inmates (N = 184, or
46% of the total sample of prison releases) who were arrested in Mecklenburg County during the twoyear follow-up period.The data included their OTIS number (a prison identification number), their PID
(the local identification number used to identify arrestees and jail inmates in Mecklenburg County),
arrest date and arrest number. If there were arrested a second time during the same two-year follow-up
period, that informationwas also included. This data, which included data from 2012, was merged with
the four-year arrest cohort (arrests 2008 through 2011) that has been used to date in Mecklenburg’s
Justice Reinvestment Initiative efforts to date, the analyses of which are detailed in a number of
previous reports. This arrest dataset, rather than containing one record per person as is the case with
the data provided by Ms. Marshall, contains a separate record for each arrest episode. Given that each
record represents a single arrest, each record also includes all charges brought at the time of the arrest,
and is keyed off the most serious charge at the time of arrest.
Findings
As noted above, 184 (46%) of the random sample of 400 prisoners released from North Carolina prisons
to Mecklenburg County had at least one arrest in the two years following their release. A total of 119
(30%) had two arrests within the follow-up period, while 75 (19%) had three or more arrests.
Using the PID field common to the two datasets, 170 of the 184 members of the prison release cohort
were identified in the 2008 – 2011 arrest cohort. During the four years encompassed by the arrest
cohort, the average number of arrests for this sample of recidivists was 6 (range from 1 to 38), with a
median of 4 and a mode of 6. Twenty-two of the members of the sample had only one arrest in the 2008
– 2011 dataset, due to the inclusion of 2012 arrest data in Ms. Marshall’s sample. Just under nine-in-ten
arrest episodes (89.3%) involved African American arrestees, with 6.1% involving Caucasians and 4%
involving Asians. Ninety-eight percent of arrests for this group represented males.
1
CEPP-JRI Mecklenburg – Recidivism Analysis of a Sample Released from Prison
2/7/13
An examination of the crime classifications (most serious offense) involved in these arrest episodes can
be seen in Table 1, below.
Table 1. Crime Classification.
Crime Classification
Frequency
Felony
Misdemeanor
Traffic
Total
Missing
Total
472
435
58
965
2
967
Percent
48.8
45
6
99.8
0.2
100
Valid
Cumulative
Percent Percent
48.9
48.9
45.1
94
6
100
100
The types of crimes (most serious offense) involved in these arrest episodes can be seen in Table 2,
below.
Table 2. Crime Type.
Crime Type
Frequency Percent
Person
Property
Drug
Other
Total
Missing
Total
147
190
225
403
965
2
967
15.2
19.6
23.3
41.7
99.8
0.2
100
Valid
Cumulative
Percent Percent
15.2
15.2
19.7
34.9
23.3
58.2
41.8
100
100
As can be seen in the tables above, about half of arrest episodes involve a misdemeanor, and 43%
involved a drug or property offense as the most serious charge. As might be expected, a number of the
arrest episodes involved either a probation or parole violation. In fact slightly more than one-in-ten
(11.7%) did, with the violation being listed as the most serious charge in the vast majority (91%) of these
cases (see Table 3, below).
Table 3. Arrest Episode Includes a Probation/Parole Violation.
Episode Includes a
Probation/Parole Violation
No
Yes
Total
Missing
Total
2
Frequency
852
113
965
2
967
Percent
Valid
Cumulative
Percent Percent
88.1
88.3
88.3
11.7
11.7
100
99.8
100
0.2
100
CEPP-JRI Mecklenburg – Recidivism Analysis of a Sample Released from Prison
2/7/13
As previously indicated, each record in the 2008 – 2011 arrest cohort represents a single arrest episode
which can, and often does contain more than one charge brought at the time of arrest. Table 4, below,
provides the number of charges per arrest episode for this sample of released prison inmates.
Table 4. Number of Charges per Arrest Episode.
# of Charges at
this Arrest Episode
1
2
3
4
5
6 or more
Total
Frequency
397
180
128
125
40
97
967
Percent
41.1
18.6
13.2
12.9
4.1
10.0
100
Cumulative
Percent
41.1
59.7
72.9
85.8
90
100
As can be seen in the above table, 60% and 90% or arrest episodes for this sample included two and five
charges or less, respectively.
The top 25 specific charges involved in the arrest episodes for this sample can be found in Table 5,
below.
Table 5. Top 25 Charges (Charge Literal), Most Serious Charge at Arrest Episode.
Rank
Charge Literal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Probation Violation
Drug Paraphernalia - Possession Of
Trespass - Second Degree - Notified Not To Enter
Breaking and/or Entering (Felony)- With Force
Driving While License Revoked
C/S-Sch. VI - Possess Marijuana - Misdemeanor
C/S-Sch. II - Possess Cocaine
Resisting Public Officer
C/S-Sch. II - P/W/I/S/D Cocaine
Assault On A Female – Non-Agg.Phys.Force
Bond Termination
C/S-Sch. VI - P/W/I/S/D Marijuana
Assault On A Female - Agg.Phys.Force
Possession Of Firearm By Felon
Consp. Robbery Dangerous Weapon
Larceny (Misdemeanor) - Under $50
Parole Violation
Communicating Threats
Larceny (Misdemeanor) - $50-199
Break/Enter Motor Vehicle- $200 & Up
3
Frequency
Percent
90
63
61
51
39
37
30
30
26
24
24
17
14
14
13
13
13
12
12
11
CEPP-JRI Mecklenburg – Recidivism Analysis of a Sample Released from Prison
9.3
6.5
6.3
5.3
4
3.8
3.1
3.1
2.7
2.5
2.5
1.8
1.4
1.4
1.3
1.3
1.3
1.2
1.2
1.1
Cumulative
Percent
9.3
15.8
22.1
27.4
31.4
35.3
38.4
41.5
44.2
46.6
49.1
50.9
52.3
53.8
55.1
56.5
57.8
59
60.3
61.4
2/7/13
Rank
Charge Literal
21
22
Larceny (Misdemeanor) - $200 & Up
Robbery With Dangerous Weapon - Individual
(Felony)
Probation Violation - Out Of County
Assault Govt. Official/Emply – Non-Agg.Phys.
Breaking and/or Entering (Felony)-Without Force
All other charges (N = 130 other separate charges)
23
24
25
26-155
Frequency
Percent
11
11
10
9
9
967
Cumulative
Percent
1.1
62.6
1.1
63.7
1
0.9
0.9
33.4
64.7
65.7
66.6
100
As can be seen in Table 5 (above), the top 25 charges account for 66% of all arrest episodes for this
sample of prison inmates released to Mecklenburg County. Eleven charges account for fully half of all
arrest episodes for this sample.As regards the Mecklenburg County Justice Reinvestment Initiative (JRI)
efforts, four of the top six charges (representing 21.6% of this sample’s arrest episodes) are included
among those selected for consideration as part of the proposed JRI strategies. About half of the arrest
episodes involved misdemeanors, and about 12% involved either a probation or parole violation. Only
about 15% represented an offense against another person or persons.
Summary and Conclusions
The present analysis extended the findings of a separate analysis involving a random sample of 400
inmates released from North Carolina prisons to Mecklenburg County and followed through 2012,
conducted by Hope Marshall. Slightly less than half (46%) of the total random sample of 400 prisoners
released from North Carolina prisons to Mecklenburg County had at least one arrest in the two years
following their release. A total of 119 (30%) had two arrests within the follow-up period, while 75 (19%)
had three or more arrests.This sample of 184 recidivists was merged with a separate dataset containing
all arrests in Mecklenburg County from 2008 through 2011. A total of 170 of the 184 members of the
prison release cohort were identified in the 2008 – 2011 arrest dataset. During the four years
encompassed by the arrest cohort, the average number of arrests for this sample of recidivists was 6
(range from 1 to 38), with a median of 4 and a mode of 6. As such, this group of 170 represented 967
separate arrest episodes over a four-year period. Given that four of the top six charges listed as the
most serious offense involved in these arrests was for charges already identified as potential JRI
strategies, it stands to reason that effecting strategies designed to address these charges would impact
both recidivist former prison inmates as well as potential arrestees in general. Given the proportion of
these episodes which involved a probation or parole violation (12%), any reduction in arrests for these
charges would be felt most acutely in a reduction in jail bed days, given the relatively high proportion of
bed days consumed by probation/parole violators.
4
CEPP-JRI Mecklenburg – Recidivism Analysis of a Sample Released from Prison
2/7/13