Gender Differences in Predictors of PrisonViolence1

RUNNING HEAD: Gender Differences in Predictors of Prison Violence

Gender Differences in Predictors of Prison Violence
Steven Griffiths
CJA334
John Dosdall
8/13/2012












Gender Differences in Predictors of PrisonViolence2


In recent years, a large number of criminologists have yearned for practical research addressing
their uncertainties that the current prisoner risk classification system that was originally designed for men
is far less accurate in predicting female violent misconduct across the U.S. Corrections facilities across
the United States have yearned for an accurate tool that answers the question: can we accurately predict
who is going to be a violent offender once confined to a state correctional facility. The current
classification system in place is cost effective, but allows for an inmates risk for violence and escape to
slip through, allowing the assignment of an inmate to a lower than needed security rating. (Most security
ratings are from minimum to maximum, with segregation reserved for those who present a severe danger
to themselves, staff, and other offenders). The study is designed to gather information from 98 institutions
across the United States, and average the accurate predictive nature of the current classification system,
are the systems in place cost effective? And is the professional judgment of those in charge of
classification true and accurate to the offenders criminal history.
Current correctional facilities divide the inmate classification system into 2 categories: the initial
classification, and re-classification, where the staff will conduct annual reviews of inmate behaviors and
determine if they can move to a different level of security (classification). The initial classification is
based on the information received from the courts regarding their criminal history and behavior, as well
as factors for enhancing their prone to violence. In the past women were mostly classified as low (or
minimum) and granted access to working outside of the prison in place of a prison job, but problems
began to arise when the lower classified inmates were showing more signs of aggression and violence
toward other staff and inmates. According to Harer 2012 For security reasons, corrections managers are
appropriately concerned about inmate risk for escape and violent behavior in prison. Therefore, these two
concerns are most often chosen as the criteria for assessing risk classification instruments. But in any
well-run correctional facility inmates seldom escape, whereas violent or attempted violent actsalthough
not as common as the popular media might suggestoccur with greater frequency This shows that the
Gender Differences in Predictors of PrisonViolence3

past system that was in place, was only covering a few of what should have been many different areas of
consideration when classifying. The study originally showed that with the current standards, about 25%
of violent offenders were being classified lower than what they should have been, and were the cause of
an increase of offender on offender violence, as well as offender on staff violence. In 2010, prison
admissions reached about 12,000 a month, of those 12,000 admitted 40% showed violent behavior, and
30% participated in violence against staff, volunteers, inmates, and their surroundings. (Baer 2012) of
the 12,000 admitted 32% were female offenders who (under the current classification system) were
housed in minimum and medium security facilities where 25% showed violent and aggressive behavior, a
willingness to escape, and made assaults on staff and other offenders. It is safe to say that the current
system had failed to identify and properly house offenders in a setting more fitting to their behavior and
criminal history. In an earlier study, the first author of the present study conducted a factor analysis of
these violence-related categories along with several other nonviolent misconduct categories using data for
male inmates Results showed that five of these six types of violence fit into one factor. Killing or attempts
to kill did not fit any of the factors identified, but that may have been due to the very small number of
homicides in the data. The clustering of these five violence-related misconducts into one factor indicates
that a male inmate who commits any one of these types of violence is at greater risk of committing any of
the others and suggests a latent variable capturing potential for violence. Baer 2012 (para 12).
To help create a better system, a predictor table created by Baer and his research team was made
to find certain things about an inmate that would make them a better candidate for classification
placement, which included the following items:
Type of detainer measures decisions by local authorities to prosecute the offender in
addition to his or her federal conviction and, as such, measures the seriousness of the
offenders behavior and the threat the offender poses to the community.
Gender Differences in Predictors of PrisonViolence4

Severity of current offense measures whether the current offense involved violent
behavior as well as any injury inflicted on victims. Prior violence has been shown to
predict future violence.
History of escapes demonstrates a willingness to flee custody or avoid prosecution,
indicating both a failure to accept responsibility and a willingness to take actions in the
future to resist legal authority and, therefore, to violate prison rules.
History of violence, as with the severity of current offense item, has been found in past
research to predict future violence.
Precommitment status measures the courts decision whether to allow the offender to
voluntarily surrender to prison authorities to serve her or his sentence. (Baer 2012)
With this set of predictors they were accurately able to house 3,000 female offenders to a proper
security level, with and error rate of +/- 2%, which meant that there was virtually no offender on
offender violence, or offender on staff violence. Using the same chart and standards they were
able to house 7,200 male offenders with an error rate +/- 8% and although there was still some
violence, it was all offender on offender, and resulted in zero serious bodily injuries.
According to Baer 2012 Our findings imply the following two key policy implications.
First, in terms of predictive accuracy, a predictively valid male risk classification instrument is
likely to predict equally well for female prisoners. Second, because our findings also show the
seriousness of violence is much lower among women, gender-specific application of the
classification instrument is required. Separate application allows classification staff to examine
rates for more serious and less serious violent misconduct at each classification score for women
and men when grouping scores into inmate security or custody categories. Doing this is likely to
result, in as much as women admitted to federal prisons are any guide, in the vast majority of
women classified as minimal risk.

Gender Differences in Predictors of PrisonViolence5



References
Baer, A.D. (2012). Gender Differences in Predictors of Prison Violence: Assessing the Predictive Validity
of a Risk Classification System. Retrieved from
http://www.bop.gov/news/research_projects/published_reports/classification/oreprharer.pdf