draft final (IGERT poster) 04132011.pptx
The Effect of Visita/on on Prison Misconduct Sarah Tahamont Goldman School of Public Policy University of California, Berkeley
Abstract A large body of literature is devoted to
The Data: 2004 Survey of Inmates in State and
Federal Correc/ons Facili/es (SISFCF): The SISFCF is a na?onally representa?ve survey of over 14,000 state prison inmates and a small sample of federal prison inmates. In this study, we are focusing exclusively on state prison inmates. The survey includes detailed informa?on on prisoner personal characteris?cs, prior criminal history, and behavioral problems while incarcerated. The survey also includes a ques?on on whether an inmate received a visit in the past month as well as a ques?on regarding the distance between one’s current loca?on and one’s home community. For the ques?on about distance from home community the respondent was asked whether the prison was 1) less than 50 miles from home, 2) 50-‐100 miles, 3) 101-‐500 miles, 4) 501-‐1000 miles, or 5) more than 1000 miles.
Mo/va/on Empirically, the rela?onship between contact with the outside and inmate behavior is unknown. While many studies aIempt to iden?fy the rela?onship between the prison environment and inmate behavior, few consider the influence of contact with the outside world. The few studies that include measures of inmate contact with the outside world do not address how visita?on may be confounded with other factors. We hypothesize that an inmate’s contact with his outside social network affects prison misconduct. In this study, we aIempt to iden?fy and characterize the rela?onship between visita?on and prison misconduct using distance from home as an instrument for whether an inmate receives a visit.
Visita?on
Percentage of Inmates Visited By Distance From Home
Prison Misconduct 0.5
Disagreeability Higher Prison Misconduct
Fewer Visits
% Inmates Visited
iden?fying the determinants of inmate behavior. By and large, this literature neglects the considera?on of the influence of inmate interac?on with those outside the prison environment. Maintaining contact with one’s family and friends may serve as a counterweight to isola?on and nega?ve peer influences experienced while incarcerated. In this paper, we assess whether receiving in-‐prison visits impacts the degree to which state prison inmates are wriIen-‐up for behavioral infrac?ons. Using the 2004 Survey of Inmates in State and Federal Correc?ons Facili?es (SISFCF), a na?onally-‐representa?ve survey of state and federal prison inmates, we use distance between one’s ins?tu?on and home community as an instrument for whether one is visited in prison in order to iden?fy the effect of visits on a variety of behavioral outcomes including write ups for drug viola?ons and assaults. We find that, for the most part, receiving visits from friends or family reduces behavioral misconduct. The predominantly nega?ve rela?onship between visits and inmate misconduct suggests that placing inmates in closer proximity to their home community might improve prison security, poten?ally create more opportuni?es for rehabilita?ve programming and reduce the nega?ve impacts of incarcera?on on inmates and their social networks.
0.4 0.3 0.2 0.1 0
Trouble Adjus?ng to Prison Environment
Fewer than 50 to 100 101 to 500 501 to 1,000+ 50 miles miles miles 1000 miles miles
Higher Prison Misconduct
More Visits
Results: Effect of Visita/on on Any Misconduct Our preferred specifica?on , Column (4) suggests that receiving a visit reduces the likelihood of being wriIen up for any viola?on by 16%. We es?mate four models, two ordinary least squares (OLS) and two instrumental variables (two-‐stage least squares (2SLS)) models.
Outcome
(1) OLS No Covariates
(2) 2SLS No Covariates
(3) OLS Covariates
(4) 2SLS Covariates
Any Viola?on
0.0079 [0.0123]
-‐0.258** [0.0797]
-‐0.00273 [0.0101]
-‐0.157** [0.0547]
** Difference significant at 1%
*Difference significant at 5%
+ Difference significant at 10%
Control Variables: In columns (3) and (4), the other covariates include dummy variables for five year age categories, dummies for female, while, black, Hispanic, and foreign born, dummies for each educa?on category in Table 4, dummy variables for having served 1 to 3, 4 to 6, 7 to 10, 11 to 13, 14 to 17, 18, to 21, 22 to 25, or 26 plus years at the ?me of the interview, dummies for having one, two, three, four, or five or more prior prison incarcera?ons, dummies for criminal jus?ce status at ?me of prison admission (parole, proba?on), dummies for whether one commiIed a violent, property or drug crime, and indicator if immediate family have done ?me, dummies for marital status, whether one had children, dummies for year when one an?cipated being released, dummies for a disabling mental health condi?on, dummies for each offense type listed in the “Offense Codes for the Na?onal Correc?ons Repor?ng Program” (74 offense categories). We control for each offense separately and interact each offense with number of prior incarcera?ons from 1-‐5, and dummies indica?ng being in one’s cell 9 to16 hours a day or 17 to 24 hours a day.
Results: Effect of Visita/on on Specific Types of Misconduct We take our preferred specifica?on from Column (4) above and es?mate the model using specific viola?on types as the dependent variable. The findings, presented in the table at right, show that visita?on has an effect on a range of specific outcomes. A couple of the findings are par?cularly notable:
• Visita/on INCREASES the likelihood of drug viola/ons. The only significant posi?ve effect of visita?on on misconduct was for drugs, which increase by 6.5%. This is a sensible result as visitors are one of the conduits of drugs into a prison. • Visita/on REDUCES physical and verbal assault on staff and other inmates.
Es/mates of the Effect of Visita/on on Specific Types of Misconduct Physical Verbal Assault on Verbal Assault on Physical Assault on Stolen Disobeying Being Out Another Assault on Another Staff Staff Orders of Order Inmate Property Inmate Weapon
Escape or AIempted Escape Alcohol
0.1
0.0651** [0.0195]
0.05
0.00109 [0.00758]
0
-‐0.00366 -‐0.0243+ -‐0.0176+ [0.0126] [0.00945] -‐0.0437* -‐0.0394** [0.0140] [0.0182] [0.0151]
-‐0.05 -‐0.1 -‐0.15
Drugs
-‐0.121** -‐0.119** [0.0277] [0.0295]
-‐0.0887** -‐0.108** [0.0228] [0.0272]
Abstract A large body of literature is devoted to
The Data: 2004 Survey of Inmates in State and
Federal Correc/ons Facili/es (SISFCF): The SISFCF is a na?onally representa?ve survey of over 14,000 state prison inmates and a small sample of federal prison inmates. In this study, we are focusing exclusively on state prison inmates. The survey includes detailed informa?on on prisoner personal characteris?cs, prior criminal history, and behavioral problems while incarcerated. The survey also includes a ques?on on whether an inmate received a visit in the past month as well as a ques?on regarding the distance between one’s current loca?on and one’s home community. For the ques?on about distance from home community the respondent was asked whether the prison was 1) less than 50 miles from home, 2) 50-‐100 miles, 3) 101-‐500 miles, 4) 501-‐1000 miles, or 5) more than 1000 miles.
Mo/va/on Empirically, the rela?onship between contact with the outside and inmate behavior is unknown. While many studies aIempt to iden?fy the rela?onship between the prison environment and inmate behavior, few consider the influence of contact with the outside world. The few studies that include measures of inmate contact with the outside world do not address how visita?on may be confounded with other factors. We hypothesize that an inmate’s contact with his outside social network affects prison misconduct. In this study, we aIempt to iden?fy and characterize the rela?onship between visita?on and prison misconduct using distance from home as an instrument for whether an inmate receives a visit.
Visita?on
Percentage of Inmates Visited By Distance From Home
Prison Misconduct 0.5
Disagreeability Higher Prison Misconduct
Fewer Visits
% Inmates Visited
iden?fying the determinants of inmate behavior. By and large, this literature neglects the considera?on of the influence of inmate interac?on with those outside the prison environment. Maintaining contact with one’s family and friends may serve as a counterweight to isola?on and nega?ve peer influences experienced while incarcerated. In this paper, we assess whether receiving in-‐prison visits impacts the degree to which state prison inmates are wriIen-‐up for behavioral infrac?ons. Using the 2004 Survey of Inmates in State and Federal Correc?ons Facili?es (SISFCF), a na?onally-‐representa?ve survey of state and federal prison inmates, we use distance between one’s ins?tu?on and home community as an instrument for whether one is visited in prison in order to iden?fy the effect of visits on a variety of behavioral outcomes including write ups for drug viola?ons and assaults. We find that, for the most part, receiving visits from friends or family reduces behavioral misconduct. The predominantly nega?ve rela?onship between visits and inmate misconduct suggests that placing inmates in closer proximity to their home community might improve prison security, poten?ally create more opportuni?es for rehabilita?ve programming and reduce the nega?ve impacts of incarcera?on on inmates and their social networks.
0.4 0.3 0.2 0.1 0
Trouble Adjus?ng to Prison Environment
Fewer than 50 to 100 101 to 500 501 to 1,000+ 50 miles miles miles 1000 miles miles
Higher Prison Misconduct
More Visits
Results: Effect of Visita/on on Any Misconduct Our preferred specifica?on , Column (4) suggests that receiving a visit reduces the likelihood of being wriIen up for any viola?on by 16%. We es?mate four models, two ordinary least squares (OLS) and two instrumental variables (two-‐stage least squares (2SLS)) models.
Outcome
(1) OLS No Covariates
(2) 2SLS No Covariates
(3) OLS Covariates
(4) 2SLS Covariates
Any Viola?on
0.0079 [0.0123]
-‐0.258** [0.0797]
-‐0.00273 [0.0101]
-‐0.157** [0.0547]
** Difference significant at 1%
*Difference significant at 5%
+ Difference significant at 10%
Control Variables: In columns (3) and (4), the other covariates include dummy variables for five year age categories, dummies for female, while, black, Hispanic, and foreign born, dummies for each educa?on category in Table 4, dummy variables for having served 1 to 3, 4 to 6, 7 to 10, 11 to 13, 14 to 17, 18, to 21, 22 to 25, or 26 plus years at the ?me of the interview, dummies for having one, two, three, four, or five or more prior prison incarcera?ons, dummies for criminal jus?ce status at ?me of prison admission (parole, proba?on), dummies for whether one commiIed a violent, property or drug crime, and indicator if immediate family have done ?me, dummies for marital status, whether one had children, dummies for year when one an?cipated being released, dummies for a disabling mental health condi?on, dummies for each offense type listed in the “Offense Codes for the Na?onal Correc?ons Repor?ng Program” (74 offense categories). We control for each offense separately and interact each offense with number of prior incarcera?ons from 1-‐5, and dummies indica?ng being in one’s cell 9 to16 hours a day or 17 to 24 hours a day.
Results: Effect of Visita/on on Specific Types of Misconduct We take our preferred specifica?on from Column (4) above and es?mate the model using specific viola?on types as the dependent variable. The findings, presented in the table at right, show that visita?on has an effect on a range of specific outcomes. A couple of the findings are par?cularly notable:
• Visita/on INCREASES the likelihood of drug viola/ons. The only significant posi?ve effect of visita?on on misconduct was for drugs, which increase by 6.5%. This is a sensible result as visitors are one of the conduits of drugs into a prison. • Visita/on REDUCES physical and verbal assault on staff and other inmates.
Es/mates of the Effect of Visita/on on Specific Types of Misconduct Physical Verbal Assault on Verbal Assault on Physical Assault on Stolen Disobeying Being Out Another Assault on Another Staff Staff Orders of Order Inmate Property Inmate Weapon
Escape or AIempted Escape Alcohol
0.1
0.0651** [0.0195]
0.05
0.00109 [0.00758]
0
-‐0.00366 -‐0.0243+ -‐0.0176+ [0.0126] [0.00945] -‐0.0437* -‐0.0394** [0.0140] [0.0182] [0.0151]
-‐0.05 -‐0.1 -‐0.15
Drugs
-‐0.121** -‐0.119** [0.0277] [0.0295]
-‐0.0887** -‐0.108** [0.0228] [0.0272]
