Predictive Factors for Violent Misconduct in Close Custody
| Date | 01 June 2007 |
| Author | Mark D. Cunningham,Jon R. Sorensen |
| DOI | 10.1177/0032885507303752 |
| Published date | 01 June 2007 |
| Subject Matter | Articles |
Volume 87 Number 2
June 2007 241-253
© 2007 Sage Publications
Predictive Factors for
10.1177/0032885507303752
http://tpj.sagepub.com
Violent Misconduct in
hosted at
http://online.sagepub.com
Close Custody
Mark D. Cunningham
Dallas, Texas
Jon R. Sorensen
Prairie View A&M University
A retrospective review was conducted of the 2003 disciplinary records of close-
custody inmates (N = 24,514) in the Florida Department of Corrections. The
frequency of various forms of violent misconduct was inversely related to the
severity of this prison violence. A logistic regression analysis controlling for
other factors demonstrated that younger age, shorter sentence, prison gang
affiliation, prior prison violence, and prior prison term were predictive of vio-
lent institutional misconduct. Conversely, older age, longer sentence, and vio-
lent offense of conviction were associated with lower rates of violent prison
infractions. Actuarial models constructed from the logistic regression analysis
were modestly successful in predicting institutional misconduct (area under the
curve = .717 to .738, p misconduct of varying severity.
Keywords:
prison; violence; risk assessment; actuarial
The identification of factors associated with prison violence is of signifi-
cant interest to prison staff and criminal justice researchers. Such factors
have an immediate practical application at admission to prison in determina-
tions of institutional classification and facility assignment. Inmate-related
factors associated with violent prison misconduct may also inform the allo-
cation of security resources and programming within a particular custody
classification for inmates who are at midpoints in their sentences.
A number of factors associated with risk of institutional misconduct
have been reported in prior research. Some of these are expected and some
are counterintuitive. Perhaps the strongest and most consistent finding is
the inverse relationship between inmate age and the incidence of prison
misconduct, as older inmates incur fewer disciplinary infractions (e.g.,
241
242
The Prison Journal
Bench & Allen, 2003; Flanagan, 1980; Hirschi & Gottfredson, 1989). This
same inverse relationship has been demonstrated in regard to violent insti-
tutional misconduct as well (Cooper & Werner, 1990; Cunningham &
Sorensen, 2006a, b, 2007; Cunningham, Sorensen, & Reidy, 2005; Sorensen
& Pilgrim, 2000; Sorensen & Wrinkle, 1996; Wooldredge, 1991).
An inverse relationship between level of education or literacy level and
prison misconduct has also been reported. Inmates with more education
have lower annual rates of disciplinary infractions (Toch & Adams, 1986)
and lower rates of violent prison misconduct (Cunningham et al., 2005;
Cunningham & Sorensen, 2006a; Harer & Langan, 2001). Education has
been hypothesized as part of a broader preconfinement community stabil-
ity factor, also reflected in employment and marital history, reported to be
associated with lower rates of community and institutional misconduct
(Quay, 1984; Toch & Adams, 1986; Van Voorhis, 1994). A history of a prior
prison confinement has been reported to increase the risk of violence in
prison (Cunningham et al., 2005; Gendreau, Goggin, & Law, 1997;
Sorensen & Pilgrim, 2000). Paradoxically, a prior probated sentence was a
risk-reducing factor for institutional misconduct in a study of high-security
inmates (Cunningham et al., 2005).
Inmates who are affiliated with a prison gang are more frequently
involved in violent prison misconduct (Sorensen & Pilgrim, 2000). Finally,
risk of violence in prison is markedly increased by a history of assaultive
misconduct in prison (Cunningham & Reidy, 1999, 2002; Sorensen &
Pilgrim, 2000).
More counterintuitive is the relationship of offense of conviction and
sentence length to prison misconduct. For example, property offenders have
been reported to have higher rates of prison disciplinary infractions (U.S.
Department of Justice, 1999) and violent prison misconduct (Cunningham
et al., 2005) than inmates convicted of homicide. Similarly counterintuitive,
inmates serving long sentences have been observed to have lower rates of
disciplinary infractions (Flanagan, 1980) and violent prison misconduct
(Cunningham & Sorensen; 2006a; Cunningham et al., 2005). This favor-
able comparison includes inmates sentenced to life terms (Akman, 1966;
Cunningham et al., 2005; Zink, 1958).
The purpose of this study is to further investigate the predictive utility of
the factors identified above in predicting violent institutional misconduct
among high-security inmates. Furthermore, this study intends to provide
more detailed base rate information and risk analysis by utilizing a range of
operational definitions of prison violence.
Cunningham, Sorensen / Predictive Factors for Violent Misconduct
243
Method
Setting
At year end 2004, the Florida Department of Corrections (Florida DOC)
confined 76,675 inmates dispersed among five primary levels of custody clas-
sification: community, minimum, medium, close, and maximum. A total of
52 major institutions incarcerated those male inmates convicted and sen-
tenced to more than a year, representing 83.9% of the state prison inmates.
The Florida DOC was asked to provide computerized data files concerning
the disciplinary behavior of incarcerated inmates and demographic, offense,
conviction, and institutional information. The sample was limited to those
inmates (N = 51,527) who were incarcerated on January 1, 2003, and served
the entire 2003 calendar year. The current study provided a standardized “at-
risk” period of one calendar year for all inmates included in the sample.
Consistent with high-security confinement in most jurisdictions, close-
custody inmates in the Florida DOC are maintained within an armed perime-
ter or under direct armed supervision when outside of a secure perimeter. This
custody level or its equivalent is the typical initial designation of inmates fac-
ing lengthy prison sentences and those convicted of the most serious violent
offenses. Capital inmates who are spared a death sentence are placed at this
high security level in most jurisdictions. By limiting the analysis to a single
custody classification, it ensured that inmates were serving their sentences
under generally similar conditions of confinement.
Participants
The sample was limited to inmates who were classified as close custody
in the Florida DOC. To prevent potential dissimilarities in housing condi-
tions and enforcement conventions from influencing the outcome, females
were excluded. The final sample included all male inmates serving the 2003
calendar year under the restrictions of close custody (N = 24,517). These
inmates had served an average of 6 years and 3 months by January 1, 2003.
Other demographic,...
To continue reading
Request your trial