Recidivism Rates - Model Factor Tables for Inmates Released 2003-2010

Table 1: Hazard Ratios for Categorical Factors Selected by Model

Value	N	% Of Release Cohort	Recidivism Rate	Hazard Ratio
Gender
Male	216,082	89%	34%	1.598**
Female	27,780	11%	19%
Race
Black	120,845	50%	37%	1.206**
Non-Black	123,017	50%	27%
Ethnicity
Hispanic	18,141	7%	25%	0.783**
Non-Hispanic	225,721	93%	32%
Custody
Low	116,815	48%	27%	0.854**
High	34,604	14%	41%	1.091**
Most Serious Crime in Inmate History
Murder / Manslaughter	8,055	3%	28%	0.768**
Sex Offense	16,763	7%	35%	1.086**
Other Violent Offense such as assault and kidnapping	64,277	26%	32%	0.965**
Supervision to Follow
Yes	84,591	35%	44%	1.965**
No	159,271	65%	25%
Family has a Positive Influence on Inmate
Yes	179,442	84%	32%	0.964**
No	32,911	16%	34%
Friends are a Negative Influence on Inmate
Yes	17,629	8%	35%	1.06**
No	194,724	92%	32%
Friends are a Positive Influence on Inmate
Yes	118,245	56%	31%	0.958**
No	94,108	44%	34%
No Spouse
Yes	53,127	25%	34%	1.058**
No	159,226	75%	32%
Inmate has a Positive Relationship with their Attorney
Yes	100,979	48%	33%	1.045**
No	111,374	52%	32%
**: p-value ≤ 0.01; *: 0.01 < p-value ≤0.05; NS: Not Significant at α = 0.05

Table 2. Hazard Ratios for Continuous Factors

Value	Median	Mean	Hazard Ratio
Criminal History
Number of Prior Prison Commitments (0-9)	0	0.6	1.331**
Number of Burglary Offenses (0-102)	0	1.0	1.035**
Number of Theft Offenses (0-401)	0	1.9	1.011**
Number of Drug Offenses (0-86)	0	1.9	1.022**
Number of Weapons Offenses (0-27)	0	0.2	0.984**
Substance Abuse Need
Inmate-Reported Drug Screening Score (0-14)	4	5.5	1.026**
Time Served
Months served during current incarceration (1-1113)	21	36.0	1.001**
Institutional Behavior
Number of Disciplinary Reports during current incarceration (0-269)	0	2.5	1.013**
Educational Level
Most recent TABE score (grade equivalents of 1-12.9)	6.8	***	0.965**
Age
Age at release (15-89)	34	35.2	0.975**
**: p-value ≤ 0.01; *: 0.01 < p-value ≤0.05; NS: Not Significant at α = 0.05; ***: An average cannot be calculated for grade equivalent TABE scores since these are not interval scale.

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 1.598.  Since it is greater than one, it means that a male inmate is (1.598-1=0.598) 59.8% 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, a Hispanic inmate is (1-0.783=.217) 21.7% less likely to recidivate than a non-Hispanic inmate with all other factors held constant.

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 an inmate incurs while incarcerated, his likelihood of recidivating increases by (1.013-1=.013) 1.3%.  For each additional grade level tested, his likelihood of recidivating decreases by (1-0.965=0.035) 3.5%.