Veterans Jus+ce Outreach: Suppor+ng Returning Prisoners Marshall Johnston Na+onal Center on Homelessness Among Veterans
Veterans Jus+ce Outreach: Suppor+ng Returning Prisoners
Marshall Johnston Na+onal Center on Homelessness
Among Veterans
In the last 37 years, the incarcerated popula+on has more than quadrupled In the US.
Number of sentenced prisoners admitted to and released from state or federal prison 1977-2000
Source: The Urban Ins+tute. Based on Bureau of Jus+ce Sta+s+cs na+onal prisoner sta+s+cs.
Homeless are Over-‐represented in Prison Popula+ons
• Homeless are high-‐risk for arrest/incarcera+on • 1999 – Bureau of Jus+ce Sta+s+cs study found that 12% of state & federal Inmates reported being homeless at the +me of their arrest. • 1997 – A California Department of Correc+ons study found that 10% of California parolees were homeless, and in urban areas (LA, San Fran) the number of homeless parolees ranges from 30-‐50%
Veterans overrepresented in the homeless popula+on
• Released inmates are high-‐risk for homelessness
• 33% of homeless men are veterans (Burt et al. 1999) • 1% of homeless women are veterans (Burt et al. 1999)
VA Ending Homelessness Ini+a+ve
• The recent a^en+on that these issues have drawn, prompted the VA ini+a+ve to eliminate homelessness among veterans by 2015. • VJP (Veterans Jus+ce Programs)
• VJO (Veterans Jus+ce Outreach) • HCRV (Health Care for Reentry Veterans)
What predicts homelessness in HCRV programs at program exit?
• Research Ques+on: Are VJPs (specifically HCRV) successful in reducing homelessness upon societal reentry?
Dependent Variables
• N=4868 – Housing arrangements at program exit
– Housing stability at program exit Frequency Percent
Stable (0) 4245 86.92% Unstable (1) 639 13.08%
Frequency Percent (Housed) 0 3254 66.77% (Homeless) 1 1614 33.23%
Independent Variables
• Program comple+on • Employment status • Financial benefits • Housing at the +me of arrest • Alcohol, drug, mental health, medical, social, and voca+onal issues
Descrip+ves • N=4868 – Study limited to individuals for whom housing status at program exit was known
Age Frequency Percent Missing 12 0.25 18-‐24 38 0.78 25-‐34 419 8.58 35-‐44 678 13.88 45-‐54 1576 32.27 55-‐64 1645 33.68 65+ 516 10.57
Gender Frequency Percent Missing 16 0.33 Female 88 1.8 Male 4780 97.87
Race Frequency Percent
American Indian or Alaskan 147 3.01
Asian 21 0.43
Black or African American 1428 29.24 Na+ve Hawaiian or Other
Pacific Islander 17 0.35
White 3168 64.86 Don’t Know 29 0.59
Vet declined to answer 32 0.66 Interviewer omi^ed item 42 0.86
10
Process
• Running frequencies • Recoding variables to eliminate those without enough power – Frequencies on re-‐coded variables
• Chi Square test – observing bivariate associa+ons
• Odds Ra+os – iden+fying at-‐risk popula+ons
Example Data: Ini+al Frequency ddlHCRVEQ8 (Employment) Frequency Percent Cumulative Frequency Cumulative Percent
8793 53.43 8793 53.43 Actively seeking employment 852 5.18 9645 58.6 Disabled or retired 1087 6.6 10732 65.21 Don't know 3129 19.01 13861 84.22 Full-time employment 209 1.27 14070 85.49 Other vocational training 46 0.28 14116 85.77 Part-time or temporary employment 133 0.81 14249 86.58 Student 56 0.34 14305 86.92 Unemployed 2015 12.24 16320 99.16 Unpaid volunteer 22 0.13 16342 99.3 VA's IT or CWT (VI) 116 0.7 16458 100
if ddlHCRVEQ8ID=3 then employment = 5; else if ddlHCRVEQ8ID=4 then employment = 6; else if ddlHCRVEQ8ID in (5,6,7,8) then employment = 3; else if ddlHCRVEQ8ID = 9 then employment = 4; else employment = ddlHCRVEQ8ID;
12
Example Data: Final Frequency Employment & Housing Housed Percent Housed Homeless Percent Homeless Total Total Percentage Disabled or retired 391 8.01% 560 11.47% 951 19.47% Unemployed 1662 34.03% 133 2.72% 1795 36.75% Actively seeking employment 302 6.18% 390 7.99% 692 14.17% Vocational training, unpaid volunteer, student 139 2.85% 68 1.39% 207 4.24% Donít know; 577 11.81% 255 5.22% 832 17.04% Part-time employment 137 2.81% 60 1.23% 197 4.03% Full-time employment 53 1.09% 157 3.21% 210 4.30%
Employment & stability Stable Percent Stable Unstable Percent Unstable Total Total Percentage Disabled or retired 137 2.81% 814 16.67% 951 19.47% Unemployed 207 4.24% 1588 32.51% 1795 36.75% Actively seeking employment 108 2.21% 584 11.96% 692 14.17% Vocational training, unpaid volunteer, student 24 0.49% 183 3.75% 207 4.24% Donít know; 135 2.76% 697 14.27% 832 17.04% Part-time employment 16 0.33% 181 3.71% 197 4.03% Full-time employment 12 0.25% 198 4.05% 210 4.30%
13
Example data: Chi Squared
Significance Variable Degrees of
Freedom Chi Square P-value
Employment by Housed 5 1064.8236 <.0001
14
Example Data: Odds Ra+os Housed & Full-Time Employment Point Estimate 95% Confidence Limit (minimum) Confidence Limit (maximum) employment 0 vs 6 0.542 0.366 0.801 employment 1 vs 6 0.055 0.037 0.082 employment 2 vs 6 0.436 0.299 0.634 employment 3 vs 6 0.235 0.148 0.373 employment 4 vs 6 0.257 0.174 0.378 employment 5 vs 6 0.240 0.150 0.385
Housed vs Homeless 1 ÷ 0.542 = 1.845 (disabled/re+red) 1 ÷ 0.055 = 18.18 (unemployed) 1 ÷ 0.436 = 2.294 (seeking employment) 1 ÷ 0.235 = 4.255 (voca+onal training, unpaid volunteer, student) 1 ÷ 0.257 = 3.891 (don’t know) 1 ÷ 0.24 = 4.167 (part-‐+me employment)
15
Stability Results: Odds Ra+os
• Non-‐VA benefits – Neither receiving nor planning to apply VS currently receiving and will con+nue: 0.66
• Prior Housing – homeless VS house/apartment: 2.17 – Group living or ins+tu+onalized VS house/apartment: 1.83
16
Stability Results: Odds Ra+os
• Drug – Not a problem area VS problem area with arranged treatment: 0.72
• Employment – Disabled/re+red VS full-‐+me employment: 2.35 – Unemployed VS full-‐+me employment: 2.54 – Seeking employment VS full-‐+me employment: 2.29
– Unknown VS full-‐+me employment: 3.37
17
Stability Results: Odds Ra+os
• Age – 35-‐44 VS 55-‐64: 0.7
• Voca+onal Deficits – Not a problem area VS problem area with arranged treatment: 0.73
• Social or recrea+onal deficits – Not a problem area VS problem area with arranged treatment: 1.77
18
Stability Results: No Significance
• Veteran race • Gender • Ending reason • VA benefits • Alcohol • Mental Health • Medical
19
Housing Results: Odds Ra+os
• Prior Housing – Homeless VS house/apartment: 2.0 – Trailer, mobile home, hotel, motel VS house/apartment: 1.62
– Group living or ins+tu+onalized VS house/apartment: 2.34
• Age – 18-‐24 VS 55-‐64: 0.22 – 65+ VS 55-‐64: 1.36
20
Housing Results: Odds Ra+os
• VA Benefits – Neither receiving nor planning to apply VS currently receiving: 1.45
– Planning to apply/pending applica+on VS currently receiving: 1.57
– Status unknown VS currently receiving: 2.02 • Ending reason – Released to area no served by HCRV VS successful comple+on: 0.63
– Incarcerated VS successful comple+on: 43.48
21
Housing Results: Odds Ra+os • Social
– Problem area with no arranged treatment VS Problem area with arranged treatment: 2.5
– Not a problem area VS problem area with arranged treatment: 1.73
• Alcohol – Not a problem area VS problem area with arranged treatment: 0.64
• Medical – Problem area with no arranged treatment VS Problem area with arranged treatment: 2.5
– Not a problem area VS problem area with arranged treatment: 1.73
22
Housing Results: No Significance
• Veteran race • Veteran gender • Non-‐VA benefits • Drug • Mental health • Voca+onal deficits
23
Next Steps
• More coding • Data collec+on by VJO specialists needs to be more uniform
24
My Role
• Literature review • Instruc+on on coding • Research paper • Sta+s+cal analysis
25
Lessons Learned
• Introductory sta+s+cs • SPSS • Wri+ng – Literature review – Research paper
• Importance of collec+ng quality data • Hundreds of random “fun facts” about veterans
26
Acknowledgements
• Dennis Culhane • Jonathan Rajkumar • Tom Byrne • Elizabeth Flatley • Ken Steif • Dan Treglia • Joanne Levy • Safa Browne • Hoag Levins, Megan Pellegrino • VA Na+onal Center on Homelessness, UPenn LDI