Making Prisons Safe:
Strategies for Reducing Violence
Donald Specter *

INTRODUCTION
Most people assume that prisons are dangerous because they
house violent convicts. In California, for example, the union
representing prison guards emphasizes the danger by calling the job
“the toughest beat in the state.” 1 Yet, in the last twenty years in only
one California prison guard has been killed by a prisoner, but
hundreds of prisoners have died from medical neglect, suicide, or
guard brutality. Prisons are dangerous, but they are far more
dangerous than they need to be.
If prison administrators provide humane conditions and require
strict adherence to commonly accepted and nationally recognized
techniques for regulating the unnecessary use of force, prisons can be
reasonably safe for both prisoners and staff. Although the threat
posed by gangs presents special problems, the traditional approach to
correctional safety—suppression and isolation—has not been
successful. The experiences of some innovative programs around the
country, as discussed below, suggests the success of a radically
* This article is based on testimony to the Commission on Safety and Abuse in
America’s Prisons, April 2005. Donald Specter, J.D. is the Director of the Prison Law Office, a
nonprofit public interest law firm located near San Francisco, California. The Prison Law
Office provides free legal services to California state prisoners related to the conditions of their
confinement. During the past thirty years the office has successfully sued California adult and
juvenile facilities over virtually all prison conditions, many of which directly relate to the safety
of prisoners, the excessive use of force, and the state’s response to prison gangs. More
information about these cases is available at the Prison Law Office website. See Prison Law
Office, http://www.prisonlaw.com (last visited July 8, 2006).
1. See California Correctional Peace Officers Association, http://www.CCPOA.org (last
visited July 8, 2006).

125

p125 Specter book pages.doc

126

Journal of Law & Policy

11/20/2006

[Vol. 22:125

different approach: closely monitored integration coupled with
incentives and tools to help prisoners leave the gangs.
THE CAUSE AND CONTROL OF PRISON VIOLENCE
The Supreme Court recently stated that “[p]risons are dangerous
places.” 2 The Court implied that prisons are dangerous because
prisoners are violent. 3 Some prisoners are violent and will be violent
no matter what the circumstances, but the degree of institutional
violence is not dependent on the prisoners. It is a direct product of
prison conditions and how the state operates its prisons.
American prisons promote violence and abuse by their design and
operation. The anti-social nature of the prisoners themselves is not
solely responsible for violent and abusive behavior. In the Stanford
Prison Experiment otherwise psychologically healthy, normal
Stanford college students changed dramatically after spending six
days as guards and prisoners in a mock prison. 4 The “prisoners”
began to perceive each other as the guards perceived them and
progressively expressed more frequent intentions to harm others:
The guards, too—who also had been carefully chosen on the
basis of their normal-average scores on a variety of personality
2. Johnson v. California, 543 U.S. 499, 515 (2005).
3. Id. Threats to the safety of prisoners also spring from other, less obvious sources. In
our experience, the single biggest threat to prisoners’ lives is the absence of adequate and
appropriate health care. Neglect and malpractice kills more prisoners than do guards or other
prisoners. Although difficult to quantify, death and serious injury due to medical neglect,
preventable suicides, and mental decompensation far exceeds the harm caused by the more
overt uses of force. For example, in 2003 only fourteen California prisoners were killed by
other prisoners or staff. CALIFORNIA DEPARTMENT OF CORRECTIONS, INMATE INCIDENTS IN
INSTITUTIONS: CALENDAR YEAR 2003, at 15 (2004), available at http://www.cdcr.ca.gov/
ReportsResearch/OffenderInfoServices/Annual/Beh1Archive.html (follow “December 2003”
hyperlink).
In a recent independent review of 193 recent deaths in California’s prisons, court-appointed
medical experts found eleven cases of terrible medical care in 2003. A review by a courtappointed special master of suicides within California’s prisons in 2003 found that out of thirtyfive suicides, twenty-five (74%) received inadequate treatment and were foreseeable or
preventable. Coleman v. Wilson, Special Master’s Report on Suicides Committed in the
California Department of Corrections in Calendar Year 2003, at 8 (2004). Therefore, although it
is beyond the scope of this presentation, any analysis of safety failures and abuse must consider
the effects of ineffective prison health care systems.
4. Craig Haney & Philip Zimbardo, The Past and Future of U.S. Prison Policy, 53 AM.
PSYCHOLOGIST 709 (1998).

p125 Specter book pages.doc

2006]

11/20/2006

Making Prisons Safe

127

measures—quickly internalized their randomly assigned role.
Many of these seemingly gentle and caring young men, some
of whom had described themselves as pacifists or Vietnam
War “doves,” soon began mistreating their peers and were
indifferent to the obvious suffering that their actions produced.
Several of them devised sadistically inventive ways to harass
and degrade the prisoners, and none of the less actively cruel
mock-guards ever intervened or complained about the abuses
they witnessed. 5
The conclusions of this experiment have profound implications
for the control of violence in our prisons:
The negative, anti-social reactions observed were not the
product of an environment created by combining a collection
of deviant personalities, but rather the result of an intrinsically
pathological situation which could distort and rechannel the
behaviour of essentially normal individuals. The abnormality
here resided in the psychological nature of the situation and not
in those who passed through it. 6
In other words, the Stanford Prison Experiment teaches that prisons,
as an institution, tend to promote aggressive and violent behavior by
correctional personnel. 7
The state is responsible for controlling that type of behavior. The
elements necessary to control the use of force in prisons are well
known. To prevent abuse, the use of force must be controlled through
(1) clear policies; (2) meaningful and constant supervision of all uses
of force; (3) timely and truthful reporting of all uses of force by the
officer involved and anyone who witnessed the incident; (4) an
accurate and unbiased investigation into allegations of excessive
force; and (5) the consistent imposition of progressive and
proportional discipline when excessive force is used or when it is not
5. Id. at 709.
6. Id. at 710 (citation omitted).
7. Craig Haney et al., Interpersonal Dynamics in a Simulated Prison, 1 INT’L J.
CRIMINOLOGY & PENOLOGY 69, 93–94 (1973).

p125 Specter book pages.doc

128

Journal of Law & Policy

11/20/2006

[Vol. 22:125

reported. 8 A breakdown in any one of these components will
inevitably lead to abuse. 9
Unfortunately, it is not uncommon for correctional supervisors to
lose control over the use of force, resulting in abuse. Abu Ghraib, 10
Pelican Bay State Prison in California, 11 the entire prison system of
Texas, 12 and Rikers Island in New York City 13 are prime examples of
prisons that became more dangerous not because of the prisoners, but
because of management breakdowns that let guards mistreat and
dehumanize their captives.
The abuse at Abu Ghraib is especially important to consider
because two factors distinguish it from other situations. Like the
Stanford Prison Experiment, and unlike maximum security prisons
where prisoners are considered extremely dangerous, in Abu Ghraib
there was no public suggestion that the prisoners were especially
threatening or that their conduct caused the guards to act abusively.
This is a clear demonstration that the situation, rather than the
prisoners themselves, was responsible for the guards’ misconduct.
The United States Army’s investigation into the abuses at Abu
Ghraib revealed another factor that can cause an institution to spiral
8. See, e.g., Madrid v. Gomez, 889 F. Supp. 1146, 1199 (N.D. Cal. 1995), rev’d and
remanded by 150 F.3d 1030 (9th Cir. 1998); U.S. ARMY SPECIAL REPORT, INVESTIGATION OF
INTELLIGENCE ACTIVITIES AT ABU GHRAIB 2–5 (2004), available at http://www4.army.mil/
ocpa/reports/ar15-6/ar15-6.pdf (citing lack of effective leadership, ambiguous policies,
inadequate resources, and ineffective supervision and discipline as reasons for the abuses).
9. This was proven true in California by the special master appointed by the federal court
to monitor the conditions at Pelican Bay who conducted an inquiry into the code of silence
within the California Department of Corrections. Madrid v. Gomez, Special Master’s Report Re
Department of Corrections “Post Powers” Investigations and Employee Discipline (Jan. 15,
2004). The inquiry arose from the decision of the director of the Department of Corrections to
terminate the investigation of perjury by several correctional officers in a federal civil rights
trial for excessive force against prisoners. The special master concluded,
A minority of rogue officers can establish a code of silence, threaten the majority,
damage cars, isolate uncooperative co-workers, and create an overall atmosphere of
deceit and corruption. And if the minority are supported by a powerful labor
organization, and the union as well as management condones the code of silence, the
consequences are severe.
Id. at 99.
10. See, e.g., U.S. ARMY SPECIAL REPORT, supra note 8.
11. See Madrid, 889 F. Supp. 1146.
12. See Ruiz v. Estelle, 503 F. Supp. 1265, 1302 (S.D. Tex. 1980).
13. See Fisher v. Koehler, 692 F. Supp. 1519 (S.D.N.Y. 1988).

p125 Specter book pages.doc

2006]

11/20/2006

Making Prisons Safe

129

out of control: the injection of an external rationale for
mistreatment. 14 In addition to finding the usual management
breakdowns in the control of force the Army investigators also
recognized that the “war on terror” and the corresponding need to
obtain intelligence were perceived by the soldiers at Abu Ghraib as a
license to exceed the bounds of sanctioned conduct. 15
A similar license was granted to guards in California that led to
even more serious abuse, often causing permanent injury and death.
At Pelican Bay, guards were led to believe that extreme force was
justified by the need to punish and control the “worst of the worst.” 16
As soon as the prison opened, officials let the guards know that the
standard rules of conduct would not apply at Pelican Bay. What
followed were not only individual instances of brutality, but a
deliberate practice of using violence and pain to control prisoners’
behavior. 17 Ruling on a constitutional challenge to the excessive use
of force at Pelican Bay, the Federal District Court for the Northern
District of California catalogued unnecessary and excessively violent
cell-extractions, hog-tying of prisoners, caging of naked prisoners
outside for long periods of time in cold and rainy weather, and staff
beatings of prisoners. 18 It concluded that violence was used by staff
“not only in good faith efforts to restore and maintain order, but also
for the very purpose of inflicting punishment and pain.” 19
This attitude pervaded other California maximum security prisons
as well. In the mid-1990s California was confronted with the
gruesome spectacle of guards in the Security Housing Unit at
Corcoran State Prison purposefully releasing rival gang members in
small exercise yards and betting on which of the “gladiators” would
14. U.S. ARMY SPECIAL REPORT, supra note 9, at 5.
15. The army acknowledged this problem in typical bureaucratic and understated fashion:
“Demands on the Human Intelligence (HUMINT) capabilities in a counterinsurgency and in the
future joint operational environment will continue to tax tactical and strategic assets.” Id. at 6.
16. Madrid, 889 F. Supp. at 1155.
17. In one instance, guards placed an African American mentally ill prisoner in a bathtub
so hot that it caused third-degree burns and made his skin peel off parts of his body and hang in
large clumps around his legs, which had turned white. Id. at 1166–67. This was in retaliation for
the man biting a guard one week earlier. Id. at 1166. A prison nurse overheard a guard remark,
“[L]ooks like we’re going to have a white boy before this is through . . . .” Id. at 1167.
18. Id. at 1162–78.
19. Id. at 1200.

p125 Specter book pages.doc

130

Journal of Law & Policy

11/20/2006

[Vol. 22:125

be victorious. 20 And in 2004 a videotape showed “counselors” at the
California Youth Authority’s maximum security prison mercilessly
beating wards as they lay passively on the day room floor. 21
These and other scandals that have plagued California’s prison
system for the last fifteen years were not hidden, nor were they
accidents. They were known and tolerated by high-level correctional
administrators who showed complete indifference to the lives and
well-being of prisoners. Their utter failure to strictly and
appropriately regulate the use of force in a manner consistent with
nationally recognized principles of correctional administration was
responsible for untold suffering.
The regulation of force is, by itself, insufficient to prevent abuse.
Prison conditions can and do breed violence. Many prisoners have
committed violent crimes, and many suffer from mental illnesses that
inhibit their ability to control their own behavior. 22 When such
inmates are placed together in overcrowded, antiquated facilities with
inadequate mental health services and nothing constructive to do
violence is inevitable. 23
The California Inspector General made clear the connection
between conditions and violence when he found that deplorable
conditions and poor management practices contributed to the murder
of a correctional officer at the California Institution for Men. 24 The
20. AMNESTY INT’L, CALIFORNIAN PRISONS: FAILURE TO PROTECT PRISONERS FROM
ABUSE (2000), available at http://origin2.amnesty.org/library/Index/ENGAMR510792000?
open&of=ENG-USA.
21. See S.F. Gate: Multimedia, http://www.sfgate.com/cgi-bin/object/article?o=0&f=/
chronicle/archive/2004/04/02/BAGLV5VDLL1.DTL (last visited Oct. 15, 2006).
22. A recent survey by the United States Department of Justice found that more than half
of the nation’s state prisoners had a mental health problem. DORIS J. JAMES & LAUREN E.
GLAZE, BUREAU OF JUSTICE STATISTICS, U.S. DEP’T OF JUSTICE, NCJ 213600, MENTAL
HEALTH PROBLEMS OF PRISON AND JAIL INMATES 1 tbl.1 (2006). An older survey found that
16% of prisoners were mentally ill. PAULA M. DITTON, BUREAU OF JUSTICE STATISTICS, U.S.
DEP’T OF JUSTICE, NCJ 174463, MENTAL HEALTH AND TREATMENT OF INMATES AND
PROBATIONERS (1999).
23. For the connection between overcrowding and violence see, for example, KATHERINE
BECKETT & THEODORE SASSON, THE POLITICS OF INJUSTICE 177 (2004); ALEXIS M. DURHAM,
CRISIS AND REFORM: CURRENT ISSUES IN AMERICAN PUNISHMENT 48–49 (1994); CRAIG
HANEY, REFORMING PUNISHMENT: PSYCHOLOGICAL LIMITS TO THE PAINS OF IMPRISONMENT
202, 205 (2005).
24. OFFICE OF THE INSPECTOR GEN., SPECIAL REVIEW INTO THE DEATH OF
CORRECTIONAL OFFICER MANUEL A. GONZALEZ, JR. ON JANUARY 10, 2005 AT THE

p125 Specter book pages.doc

2006]

11/20/2006

Making Prisons Safe

131

Inspector General found that the prison violated basic classification
procedures by permitting the prisoner, who was incarcerated for
attempting to kill a police officer and had a history of recent and
serious assaultive behavior, to remain in the general population. 25
Poor maintenance and tool control procedures permitted prisoners to
obtain and conceal weapons. 26 The victim failed to follow specific
security directives initiated after a race riot, the stabbing of a
prisoner, and the discovery of weapons in the same unit. 27
Additionally, the warden and her subordinate supervisory staff failed
to ensure compliance with those directives. 28 Finally, nobody
addressed the prisoner’s clearly identified need for immediate mental
health treatment. 29
The Inspector General’s findings express in detail what is
common sense to most correctional administrators: well-run prisons
are relatively safe, while those that are poorly managed are not. The
control of violence, therefore, depends not only on executing
accepted policies for regulating the use and supervision of force, but
also on the overall management of the facility. All of the prison’s
operations, including mental health care, must be integrated and
functioning properly if prisons are to perform their primary purpose
of incarceration and not subject their inhabitants—both prisoners and
officers—to an unacceptable risk of injury or death.
CALIFORNIA INSTITUTION FOR MEN 3–4 (2005), available at www.oig.ca.gov/reports/pdf/
Review_03-17-05.pdf. The California Institution for Men is a large, overcrowded intake center
located near Los Angeles. See California Institute for Men (CIM), http://www.cdcr.ca.gov/
Visitors/fac_prison_CIM.html (last visited July 8, 2006).
25. Id. at 16–17.
26. Id. at 37.
27. Id. at 5.
28. Id. at 19–24.
29. Id. at 79. The independent panel of experts appointed by the California State Board of
Corrections also noted the “deplorable” conditions at the prison and the fact that the reception
center was overcrowded and serving more prisoners than it could safely process. CAL. STATE
BD. OF CORRS., INDEPENDENT OPERATIONS AND INCIDENT REVIEW PANEL ON THE CALIFORNIA
INSTITUTIONS FOR MEN 12 (2005), available at http://www.bdcorr.ca.gov/special_reports/
operational_incident_review_cim/final%20report.pdf.

p125 Specter book pages.doc

132

11/20/2006

Journal of Law & Policy

[Vol. 22:125

GANG PREVENTION
The problem of gang-related violence in America’s prisons is both
well known and well documented. Both prison and street gangs are
reportedly responsible for drugs, violence, and intimidation within
prison walls. 30 The traditional prison response to gang behavior is
suppression and isolation. Gang members are forced to spend long,
indefinite terms in segregated housing units in maximum or supermaximum (supermax) security prisons and their misconduct is often
targeted for administrative or criminal prosecution.
While the success of these strategies in reducing violence is
uncertain, it is clear that they have not succeeded in eliminating
gangs or their influence. It is commonly understood that, while
locked in segregation, gang leaders continue to control the illegal
activities of their members both within the prison and in the outside
community. Perhaps the most graphic example of this is the fact that
several California prison gang leaders are now facing the death
penalty for federal criminal charges arising from their activities while
imprisoned in the Security Housing Unit (SHU) at Pelican Bay, a
supermax facility. 31
Prisoners in restricted units such as Pelican Bay’s SHU are not
provided any form of meaningful recreation, education, vocational
training, or rehabilitative services. 32 They are left in their cells every
day for up to twenty-three hours, with the remaining hour being spent
either alone or with their cellmate in a small enclosed space that
approximates a dog run. 33 In this environment normal social
relationships are impossible and these inmates are left to associate
with other gang members. It is not surprising, therefore, that they will
continue to perpetuate the gang and its activities; they have nothing
else to do.
30. See, e.g., GEORGE F. COLE & CHRISTOPHER E. SMITH, CRIMINAL JUSTICE IN AMERICA
(2005).
31. Press Release, Thom Mrozek, Pub. Affairs Officer, U.S. Attorney’s Office, Cent. Dist.
of Cal., Racketeering Indictment Targets Aryan Brotherhood (Oct. 17, 2002), available at
http://www.usdoj.gov/usao/cac/pr2002/152.html.
32. Madrid v. Gomez, 889 F. Supp. 1146, 1229 (N.D. Cal. 1995), rev’d and remanded by
150 F.3d 1030 (9th Cir. 1998).
33. Id.

p125 Specter book pages.doc

2006]

11/20/2006

Making Prisons Safe

133

Recently, some correctional systems have developed successful
programs that use the opposite approach. Instead of using extreme
forms of isolation these programs actively promote integration. Their
aim is to reintegrate the gang member into the general prison
population. In the few programs that take this approach, the strategy
has proven surprisingly successful.
These programs—utilized in Connecticut, 34 Missouri’s Division
of Youth Services, 35 and the Pelican Bay Transitional Housing
Unit 36—have several things in common. First, they create housing
units that are relatively small, consisting of between fifteen and
twenty prisoners. Second, the adult programs allow the prisoner to
choose to participate, although one program conditions that choice on
the decision to inform against his former gang. Third, they create a
set of expectations that include mandatory integration with prisoners
of other races and gangs, and a very low tolerance of misbehavior.
Fourth, prisoners are given extensive orientation about these
expectations. Fifth, prisoners are provided with counseling services to
help them control anger and violence and foster healthy relationships.
These services come in the form of formal group sessions, but also
informally through guidance provided by specially selected staff in
the units. Finally, prisoners are provided with real and substantial
incentives to complete the program. This may include contact visits
with their family, jobs and a safe environment when they return to the
general population.
The success rate in each of these three programs is reportedly very
high. Connecticut and Missouri report that the recidivism rate of gang
members is under 10%. 37 In California, prison officials report that
only 5% of prisoners fail to complete the program, the recidivism rate
34. See Connecticut Department of Correction, http://www.ct.gov/doc/ (follow
“Recidivism” hyperlink) (describing Connecticut’s Gang Management Program).
35. See Missouri Division of Youth Services, http://www.dss.mo.gov/dys/ (last visited
Oct. 15, 2006).
36. See Pelican Bay State Prison, http://www.cdcr.ca.gov/Visitors/fac_prison_PBSP.html
(last visited Oct. 15, 2006).
37. Department of Correction, supra note 34. Recidivism in these states is measured
differently. In Connecticut, the test is whether the prisoner is redesignated as a gang member.
Id. Missouri uses the more traditional measure of whether the person reoffends. Telephone
interview with Mark Steward, former Dir. of Mo. Div. of Youth Servs., Mo. Dept. of Soc.
Servs. (2004).

p125 Specter book pages.doc

134

11/20/2006

Journal of Law & Policy

[Vol. 22:125

after parole is one-third of the norm, and that less than a handful are
sent back to segregation for gang-related activities. 38
These successful programs place the prisoner or ward in a culture
where violence is not “business as usual” and provide them with the
tools they need to succeed. They prove to prisoners who have never
had a meaningful conversation with a rival gang member or a
prisoner of another race that they can live a different, less violent, and
more meaningful life. Instead of creating a culture of suppression and
isolation, they provide a transition to a more normal way of life, even
if it is limited to the confines of the prison. Given time and help, the
prisoners adapt to this culture and recognize its value.
These promising programs should be studied, evaluated, and
replicated. They offer positive alternatives to the traditional reaction
to prison gangs. It is an approach that uses the institution of a prison
to create positive change rather than to promote further violence.
CONCLUSION
It is easy to blame prisoners for prison violence. But, the lessons
of the last few decades of court intervention and academic research
have demonstrated that the amount of violence in a prison is a
function of its culture, the effectiveness of its management, and, at
times, the political reality that excuses the mistreatment of prisoners.
No prison illustrated this better than Pelican Bay in the early 1990s. It
suffered from an administration that condoned and perpetuated
violence, it opened at a time when being tough on crime meant
brutalizing prisoners and there was no effective management of the
prison. The violence in that institution has largely subsided through
better management required by intensive court intervention, and
promising inroads have been made into gang membership through the
Transitional Housing Unit program, demonstrating clearly that it is
not the prisoners but the prison as an institution that is the key to
safety in correctional facilities.
38. Telephone interview with Ted Roberts & Chris Hizer, Mgmt. Staff, Transitional Hous.
Unit, Pelican Bay State Prison (Apr. 8, 2005).