USA: +1(669)289-8683 
Sign in
In this module, we read about the current crime and policy landscape nationally, in particular, recent crime trends and the growth in all parts of the criminal justice system. The chapter by Mears emphasizes that there is little research to support current policies and, just as importantly, that little is known about the implementation and effects of many of them.
In this week’s discussion, reflect a bit on the assigned reading, and then address this key question: What are the risks of adopting unnecessary, ineffective, or inefficient criminal justice policies? How would you characterize those risks at the national and the local level?
2
Irrational Criminal Justice Policy
A CENTRAL GOAL OF THIS BOOK IS TO CONTRIBUTE TO EFFORTS TO
improve criminal justice policy and to do so by showing how the sys-
tematic use of evaluation research can lead to less bad policy and more good
policy. The ultimate aim is to help place criminal justice policy on a more
rational footing, one where it has a chance of providing the accountability
and the effectiveness that the public expects of it. At present, and as detailed
in subsequent chapters, too many criminal justice policies are ill founded,
ineffective, or inefficient, or they lack sufficient evidence to support them.
Put differently, we have too much irrational criminal justice policy. I argue
that increased reliance on the evaluation hierarchy in all parts of the crimi-
nal justice system and in the development and assessment of policy provide
s
one critical platform for correcting this situation and fulfilling the public’s
desire for effective government.
This chapter sets the stage for this argument and the subsequent chap-
ters in several ways. First, it briefly introduces the evaluation hierarchy
as a framework for critiquing current policy. The details of the hierarchy
are discussed in Chapter 3, but a discussion here provides a foothold for
understanding the context – in particular, the lack of accountability and
effective criminal justice policies – that motivates this book. Second, it pro-
vides a portrait of national crime and justice system trends, and, specif-
ically, the dramatic increase in criminal justice populations and expendi-
tures. This discussion serves to illustrate the stakes involved and to highlight
the need for creating the types of research that can allow for accountabil-
ity and evidence-based policy. Third, it describes some of the prominent
factors that have been argued to influence criminal justice policy. This
discussion highlights the fact that many barriers to research-based pol-
icy exist. In so doing, it underscores the importance of institutionalizing
7
8 AMERICAN CRIMINAL JUSTICE POLICY
evaluation research as a critical part of an effort to place the criminal jus-
tice system on a more rational – that is, a more accountable and effective –
foundation.
The Evaluation Hierarchy and the Irrationality
of Criminal Justice Policy
Imagine someone tries to sell you a used car. Before responding, you may
well ask yourself five questions. The first question is, “Do I actually need a
car?” Perhaps you do. But perhaps you do not. In some areas, having a car is
practically a requirement for getting to and from work. In others, especially
metropolitan areas, public transportation may suffice.
Assuming that you do need a car, the second question presents itself:
“Is this the kind of car I need?” For example, if you have a daily two-hour
commute, you likely would want a car with good gas mileage, so a large
pickup truck might not be the best option, all else being equal. Even so, gas
mileage may not be the only relevant consideration. If you are more than six
feet tall, some gas-efficient cars may feel uncomfortable. In addition, your
work may involve shorter trips and carting large equipment from one place
to another. A compact vehicle may not be the best option in such cases.
If you determine that the car indeed is appropriate for your purposes, you
then likely will ask a third question, namely, “What is its condition?” That is,
how has it been used and how well has it been maintained? For example,
if it is a manual-shift vehicle, did the person shift in a way that wore the
clutch down? If so, you may need to pay for a new clutch soon. Was the car
ever in an accident? If so, there may be problems not immediately evident
that could surface and lead to many costly repairs. Conversely, if the owner
supplies detailed records, including regular tune-ups, oil changes, and the
like, you may have more trust that the car will be reliable and that it will
perform in the way you expect.
Should you determine, upon inspection of the car, that it passes muster,
you might well entertain a fourth question: “What do reviews about this
type of car say about its performance?” For instance, are claims about the
car’s gas mileage supported? How often are repairs typically needed? Put
differently, how much can you trust that the car will get you where you want
to go on a regular basis with minimal maintenance?
In the event that all of the preceding questions lead you to believe that the
car is indeed worth purchasing, you typically will ask a final question: “Do
the benefits of this car outweigh the costs, especially as compared to other
cars or to other pressing needs that I may have?” Presumably, if another
IRRATIONAL CRIMINAL JUSTICE POLICY 9
car passed all of the previously mentioned criteria comparably well but cost
less, you would go with it. Similarly, if you determined that some other
pressing need was more important – say, payment for a medical procedure
not covered by insurance – you might well pass on what otherwise seems to
be a good deal.
This example is by no means accidental. Car ownership is widespread in
America, and so such questions are far from academic. Even so, the reality
is that most of us make similar sets of calculations about many decisions in
our everyday lives. I work in a university setting and students typically need
to take classes to graduate. But they do not need to take all of them. Rather,
they will want to select those courses best suited to help them complete
their major. They also will want courses that are taught well, and they will
want to learn something as a result of taking them. At a more general
level, by attending a university, they, or their parents, are proceeding on the
assumption that the benefits of additional years of schooling offset the costs
of not immediately entering the workforce.
Such calculations are the stuff of life and affect even our most mundane
decisions. Each week, my son and I go grocery shopping and must select
from what seems to be about fifty or more different types of toothpaste. We
have determined (or, really, I have) that we definitely need toothpaste. We
are not entirely clear why toothpaste helps, but we proceed on faith. We are
pretty clear that we need to put the toothpaste on the toothbrush for it to
have any chance of having an effect. However, we have not the foggiest idea
which toothpaste produces the best effects. Does brand X produce fewer
cavities? How about brand Y? Even if brand X is better at reducing cavities,
perhaps it is not as good at preventing tartar buildup, which seems like a good
thing to avoid. Perhaps it is a draw when it comes to whitening teeth. Is one
brand better at reducing tooth sensitivity? Then there is the whole question
of how well it freshens your breath. After making many assumptions about
the relative benefits of one brand versus another, we then consider the cost
of the toothpaste. Some toothpastes seem to do everything at once, but they
also tend to cost more, and there is the concern that perhaps they do some
things (e.g., cavity prevention) less well than others (e.g., tartar reduction).
In the end, I am not at all sure we end up making the best choice. But,
the steps we go through – that all of us go through in making decisions
throughout the course of every day – involve a logic and sequence highly
relevant to policy debates.
In particular, most of us employ an evaluation hierarchy in our decision
making, especially if it involves serious financial investments. Indeed, viewing
our everyday decisions as policies, and the evaluation hierarchy, as depicted
10 AMERICAN CRIMINAL JUSTICE POLICY
1. Needs evaluation:
Is there a need for
the policy?
2. Theory
evaluation: Is the
policy grounded in a
clear or well-
established theory?
Yes
3. Implementation/
process evaluation:
Is implementation
of the policy
consistent with its
theory or design?
4. Outcome and
impact evaluations:
Is the policy
associated with and
does it cause
intended outcomes?
5. Cost-efficiency
evaluation: Does the
policy achieve
outcomes cost-
efficiently?
Yes
Yes
Yes
Yes
Consider continuing,
expanding, or
terminating policy
Figure 2.1. The evaluation hierarchy. Adapted from Evaluation: A Systematic Approach, by
Peter H. Rossi, Mark W. Lipsey, and Howard E. Freeman, 7th edition, 2004, p. 80. Used with
permission from Sage Publications.
in Figure 2.1, proceeds as follows. First, we assess whether a need for a policy
exists; this is a needs evaluation. Second, we then assess whether the theory
underlying the policy is logical; coherent; and, ideally, supported by research;
this is a theory evaluation. Third, we assess how well a policy is implemented;
this is typically termed an implementation or a process evaluation. Fourth, we
assess (1) whether the policy actually is associated with intended outcomes
(this is typically termed an outcome evaluation) and (2) whether it likely
causes the outcomes (this is typically termed an impact evaluation). Fifth,
we assess whether the policy’s benefits outweigh its costs and whether the
benefits, relative to costs, are substantially greater than those of another
policy; this is a cost-efficiency evaluation. The first type of cost-efficiency
evaluation is a cost-effectiveness analysis, which compares the costs of two
or more policies aimed at achieving the same goals. The second type is a
cost-benefit analysis, which compares the costs and benefits of two or more
policies that have different goals. In a cost-benefit analysis, the impacts on
goals are monetized (i.e., we assign monetary values to them) so that we can
make apples-to-apples comparisons between two or more policies.
IRRATIONAL CRIMINAL JUSTICE POLICY 11
Each type of evaluation can be viewed as involving a particular question.
In the car example, the five questions are: (1) Do we need a car? (2) Does
the theory for selecting the car make sense? That is, are particular types of
cars best suited, by design, to meet our needs? If so, what are the relevant
considerations (e.g., appearance, comfort, gas mileage)? (3) Has the car
been implemented well in the sense that it has been well maintained? If not,
we can be reasonably sure that we may not obtain the benefits (e.g., low-cost
transportation) that we want from the car. (4) Does the car actually achieve
the impacts that we expect? Does it, for example, actually get the expected
fuel savings, and is it as comfortable as reported? (5) Is the car the most cost-
efficient way to travel from one place to the next or are there other options
that would be as effective but cheaper? Alternatively, is there another, more
pressing need (e.g., medical care) that, if addressed, would result in more
benefits than would be gained by better transportation?
The basic logic of the hierarchy is that we do not typically want to proceed
with a policy if the questions associated with each level of the hierarchy have
not been adequately answered. Consider, again, a car purchase. We typically
would proceed with buying a car only if we needed it. If we do need a car,
most of us would buy one that fit our particular need (e.g., a compact car for
commuting or a pickup truck for carting heavy equipment). Even then, we
likely would not make a purchase unless we felt confident that the car had
been well maintained and needed no expensive repairs. We would also be
likely to refrain from a purchase if we learned that the car’s performance was
or would be poor. For example, a car may be touted as getting thirty miles
per gallon, but perhaps independent tests establish that, in reality, the true
performance lies closer to twenty miles per gallon. Finally, most of us would
hold back on buying a car that performs no better than another but costs
twice as much; similarly, most of us would not buy a car when the money
could be used to pay for a life-saving surgery that we need.
Needless to say, when we employ the evaluation hierarchy in our day-to-
day lives, we may not make accurate assessments, and we may lack sufficient
information to make good judgment calls. (In all likelihood, I have been
using the wrong toothpaste for many years.) But that does not negate the
importance of the hierarchy in guiding our decision making and helping us
to arrive at better decisions.
Here is the catch – even though most of us proceed through the evaluation
hierarchy in our day-to-day decision making, criminal justice policy typically
proceeds without recourse to it. That is, quite simply, irrational. The pursuit
of crime prevention and justice constitute critical societal goals. No one
wants to be a victim of a crime. No one wants a society in which justice
12 AMERICAN CRIMINAL JUSTICE POLICY
is arbitrarily meted out. And, of course, no one wants to expend scarce
resources willy-nilly without obtaining some type of return, ideally the most
possible one, especially when the stakes are high.
The Policy Context and the Stakes Involved
A central aim of this book is to argue, through the use of the evaluation
hierarchy, that many of the nation’s most prominent criminal justice policies
lack a solid theoretical and empirical foundation and that the necessary
ingredients for holding the criminal justice system accountable and making
it effective do not yet exist. I elaborate on this argument in subsequent
chapters. Here, however, I want to turn to the national criminal justice
policy context to highlight some of the stakes involved in allowing criminal
justice policy to be irrational.
Crime Rates
To begin, let us first focus on crime rates. Many different sources of data
can be used to examine crime. For example, arrests and calls to the police
frequently serve as the basis for establishing whether crime has increased,
decreased, or remained stable. Many news accounts focus on such data. If
the number of robbery arrests increases from, for example, 100 to 110, a
news account may well report that crime is up 10 percent. That would be
incorrect. Law enforcement data reflect two factors: crime and law enforce-
ment behavior.1 Observe, for example, that a community’s true crime rate
could decrease, but if the number of police officers doubled you likely would
see a dramatic increase in arrests and possibly reported crime. So, a more
accurate news account would say that arrests have gone up 10 percent and
that the increase reflects increased crime, increased law enforcement activ-
ity, or both.
If we want to determine what the true rates of crime are, it would be far
better to conduct offender and victimization surveys.2 The first would allow
us to determine how many offenders exist and how much crime they com-
mit, while the latter would allow us to identify the total number of victims
of crime. No large-scale, nationally representative offender surveys exist in
the United States, though a number of small-scale studies exist. By contrast,
the U.S. federal government has invested a considerable amount of money
and effort into conducting a large, nationally representative victimization
survey, titled, appropriately enough, the National Crime Victimization Sur-
vey (NCVS). The first data collection for the study began in 1973 and today
IRRATIONAL CRIMINAL JUSTICE POLICY 13
0
10
20
30
40
50
60
19
73
19
74
19
75
19
76
19
77
19
78
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
20
06
20
07
20
08
V
io
le
nt
V
ic
ti
m
iz
at
io
n
R
at
e
pe
r
1,
00
0
Pe
rs
on
s
Figure 2.2. Violent victimization, 1973–2008. The increase from 2005 to 2006 is not likely a
reflection of a true increase in victimization, but rather reflects a change in the methodology used
with the National Crime Victimization Survey, the source for the victimization estimates (Rand
2008:2). Sources: Rand, Michael R. 2009. Criminal Victimization, 2008. Washington, D.C.:
Bureau of Justice Statistics. Rand, Michael R. 2008. Criminal Victimization, 2007. Washington,
D.C.: Bureau of Justice Statistics. Rand, Michael R., and Shannan Catalano. 2007. Criminal
Victimization, 2006. Washington, D.C.: Bureau of Justice Statistics.
includes more than 60,000 households and the victimization experiences of
persons ages twelve or older.3
When we examine the trends in violent crime (rape or sexual assault,
robbery, and aggravated and simple assault), which typically seem to garner
the most concern among the public and policy makers, we see that such
crime remained relatively stable throughout the 1970s and declined in the
early 1980s, as shown in Figure 2.2. Then, around 1986, it began to rise
steadily, peaking in 1994 before beginning a steady decline during the next
decade. The victimization survey entails the interviewing of individuals and
so does not capture homicides. However, for that offense, law enforcement
data tend to be relatively accurate; analyses of such data reveal that the
trend in homicides during the same time period largely mirrored the trend
for violent crime generally.4 When we turn to property crime (burglary, motor
vehicle theft, and theft) – as depicted in Figure 2.3 and as measured by the
NCVS – we see a steady decline over three decades.5 In short, except for the
14 AMERICAN CRIMINAL JUSTICE POLICY
0
100
200
300
400
500
600
19
73
19
74
19
75
19
76
19
77
19
78
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
20
06
20
07
20
08
P
ro
pe
rt
y
V
ic
ti
m
iz
at
io
n
R
at
e
pe
r
1,
00
0
H
ou
se
ho
ld
s
Figure 2.3. Property victimization, 1973–2008. Sources: Rand, Michael R. 2009. Criminal Vic-
timization, 2008. Washington, D.C.: Bureau of Justice Statistics. Rand, Michael R. 2008. Crim-
inal Victimization, 2007. Washington, D.C.: Bureau of Justice Statistics. Rand, Michael R., and
Shannan Catalano. 2007. Criminal Victimization, 2006. Washington, D.C.: Bureau of Justice
Statistics.
rise in violent crime from 1986 to 1994, crime has been stable or declining
since 1973.
Correctional System Growth
On the basis of analysis of these trends, we might reasonably hypothesize
that the criminal justice system would have increased modestly in the 1980s
to early 1990s to address the rise in violent crime but that it otherwise
would have remained stable and perhaps even decreased in size. Such a
prediction would be wrong. Juxtaposed against the overall decline in violent
and property crime has been unprecedented growth in the criminal justice
system. The growth in corrections alone is striking, as can be seen in Fig-
ure 2.4. In 1980, there were 1.8 million individuals under some form of state
or federal supervision or incarcerated in jail or prison. By 2008, that number
more than quadrupled, rising to 7.3 million.
By far, the biggest driver of that growth has been the increase in the proba-
tion population, which has risen from 1.1 million to 4.3 million. Even so, the
IRRATIONAL CRIMINAL JUSTICE POLICY 15
0
1,000,000
2,000,000
3,000,000
4,000,000
5,000,000
6,000,000
7,000,000
8,000,000
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
20
06
20
07
20
08
Total
Probation
Prison
Parole
Jail
Figure 2.4. U.S. correctional populations, 1980–2008. Source: Bureau of Justice Statistics.
2009. Correctional Populations in the United States. Washington, D.C.: U.S. Department
of Justice. Available online: http://www.ojp.usdoj.gov/bjs/glance/tables/corr2tab.htm (accessed
December 15, 2009).
jail and prison populations increased at higher rates. For example, the num-
ber of individuals in jail grew from 183,988 to 785,556, an increase of 327 per-
cent. Prison populations grew even more, increasing from 319,598 inmates
to more than 1.5 million, or 375 percent. That growth is striking given that
jails and prisons typically cost considerably more to build and operate com-
pared with the costs of probation and parole or various types of community
supervision and intermediate sanctions. They also, for all intents and pur-
poses, constitute permanent investments. For example, once a prison is built,
it generally will remain in use for decades. So, any expansion in prison capac-
ity essentially represents an indefinite commitment to increased prison costs.
Why? When states decide to expand prison capacity, they cannot easily undo
that decision if, at a later point, they determine that less capacity is needed.
Prisoner Reentry
The large-scale increase in the number of individuals incarcerated in jails
and prisons translates into a new social problem – namely, the return of
large numbers of ex-prisoners back into communities, what has been termed
16 AMERICAN CRIMINAL JUSTICE POLICY
50.4%
54.6%
68.1%
59.6%
62.5%
66.7%
62.2%
73.8%
61.7%
67.5%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Drug
Public order
Property
Violent
All released prisoners
Percentage Rearrested within 3 Years
1994
1983
Figure 2.5. Recidivism of prisoners from 15 states released in 1983 and in 1994. Source:
Langan, Patrick A., and David J. Levin. 2002. Recidivism of Prisoners Released in 1994. Wash-
ington, D.C.: Bureau of Justice Statistics.
“prisoner reentry.”6 Annually, more than 735,000 inmates leave state or fed-
eral prisons7 and undergo the process of transitioning from institutional life
to a context in which they have few opportunities for employment and fre-
quently suffer from a number of problems, including mental and physical
health problems, substance abuse, family dysfunction, histories of physical
and sexual abuse, and spotty educational and employment histories.8 Of par-
ticular concern is the high likelihood that these individuals will recidivate.
Figure 2.5 shows the rates of recidivism from one of the largest national
studies ever conducted. It reveals that, in 1994, more than two-thirds
(68 percent) of released prisoners were rearrested within three years.
Remarkably, after the large-scale increases in the correctional system and
the spate of “get tough” crime policies in the 1980s and 1990s, this level of
recidivism was higher than it was a decade earlier. (In 1983, “only” 63 per-
cent of released prisoners were rearrested within three years.) It remains
unclear why the increase occurred, although it may have stemmed in part
from a decline in educational, vocational, and treatment programming in
prisons during this time period.9 It also may have reflected more vigorous
law enforcement activity. For example, numerous efforts were taken to target
IRRATIONAL CRIMINAL JUSTICE POLICY 17
drug crimes, which would have increased drug arrests. Indirect support for
that explanation can be seen at the bottom of the figure – the percentage of
drug offenders rearrested increased from 50 percent to 67 percent between
1983 and 1994. Observe that recidivism in the study was measured using
rearrest. That means that the study included only those crimes for which
a released prisoner was arrested. If measured using self-reported offending
data, the recidivism rate assuredly would have been higher.
In short, America now faces a situation in which ever-greater numbers of
individuals are returning to communities and almost all of these individuals
continue to commit crime. The “glass half full” view of the situation is that
considerable room for improvement exists, especially given the ubiquity of
reoffending among people released from prison. The “glass half empty” view,
however, is that we may not be able to make much of a dent in recidivism
rates given the commitment to increased incarceration. Of course, it can be
argued that incarceration helps society by reducing crime through incapaci-
tation or general deterrent effects. So, even if recidivism rates remain high or
increase, perhaps that negative is offset by the positive of overall decreased
rates of crime. There is some evidence – although far from compelling – to
suggest warrant for such optimism, as will be discussed in later chapters.
Regardless, recidivism stands as a concern in its own right – few of us want
someone who may reoffend moving next door to where we live.
Criminal Justice Expenditures
Putting aside such concerns, the stakes involved in criminal justice policy
can be highlighted by turning to economic considerations. Given the growth
in the criminal justice system, it perhaps should come as no surprise that
criminal justice expenditures have dramatically increased as well, as is evi-
dent from Figure 2.6. From 1982 to 2006, the United States increased
its investment in police more than fivefold, from $19 billion to more than
$99 billion. It increased its investment in corrections almost eightfold dur-
ing the same time span, from $9 billion to $69 billion. And it increased its
investment in the judiciary – which is required to process the large influx
of new cases – from almost $8 billion to $47 billion. Adding all functions
together, from 1982 to 2006, criminal justice expenditures rose by 500 per-
cent, from $36 billion to $215 billion. Inflation accounts only for $30 billion
or so of that increase.10
The burden of these new costs has largely fallen to local jurisdictions and to
states, not the federal government. Figure 2.7 depicts the trends in criminal
justice expenditures by level of government. The most dramatic increase,
18 AMERICAN CRIMINAL JUSTICE POLICY
$0
$20
$40
$60
$80
$100
$120
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
20
06
$
(B
il
li
on
s)
Police
Corrections
Judicial
Figure 2.6. Criminal justice expenditures by function, 1982–2006. Source: Bureau of Justice
Statistics. 2009. Justice Expenditure and Employment Extracts. Washington, D.C.: U.S. Depart-
ment of Justice. Available online: http://www.ojp.usdoj.gov/bjs/glance/tables/exptyptab.htm
(accessed October 1, 2009).
in absolute amounts, clearly lies with local jurisdictions. In 1982, localities
had $21 billion in criminal justice expenditures. By 2006, their expenditures
increased to $109 billion. States’ investments in criminal justice rose almost
as dramatically during this same time period, from $11 billion to $69 billion.
And federal expenditures rose from $4 billion to $36 billion. As of 2006,
roughly 51 percent of all criminal justice expenditures were borne by local
jurisdictions, 32 percent by states, and 17 percent by the federal government.
Evidence for Current Criminal Justice Policies
Set against a backdrop of dramatic increases in criminal justice funding and
in federal funding for a wide range of crime prevention and crime control
policies is the pressing concern that too little research exists to support the
selection and continued support of many of these policies. Several criminal
justice policy reviews that have emerged in recent years suggest, in fact,
that most criminal justice policies lack a strong empirical foundation. A
National Academy of Sciences review found, for example, that “scientifically
strong impact evaluations of [crime prevention and crime control] programs,
IRRATIONAL CRIMINAL JUSTICE POLICY 19
$0
$20
$40
$60
$80
$100
$120
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
20
06
$
(B
il
li
on
s)
Local
State
Federal
Figure 2.7. Criminal justice expenditures by level of government, 1982–2006. Source: Bureau
of Justice Statistics. 2009. Justice Expenditure and Employment Extracts. Washington, D.C.:
U.S. Department of Justice. Available online: http://www.ojp.usdoj.gov/bjs/glance/expgov.htm
(accessed October 1, 2009).
while improving, are still uncommon in the context of the overall number of
programs that have received funding.”11 Similar critiques, discussed in later
chapters, have been leveled against many different parts of the criminal
justice system.
In various ways, this book will tackle the question of whether the criminal
justice policy investments of the past several decades have been wise choices.
What I will argue is that, by and large, local and state governments, and the
country as a whole, lack an institutionalized foundation for conducting the
types of research necessary to produce wise choices, much less to show that
existing choices are sensible. The result? Too little evidence exists to support
many if not most of the policies and practices that constitute the nation’s
criminal justice system.
Ultimately, the failure to use research to inform criminal justice policy con-
stitutes a profound mistake with real-world consequences. Society spends a
great deal of resources on catching and punishing as well as treating offend-
ers. It has spent even more in recent decades. Even so, a limited supply of
funds exists. We cannot, for example, build enough prisons to house every
20 AMERICAN CRIMINAL JUSTICE POLICY
person who commits a crime. As with our individual financial decisions,
mistakes about the allocation of large amounts of time and money can have
dramatic affects. They can, for example, contribute to a lack of accountabil-
ity; to a failure to identify or implement evidence-based policies; and, more
generally, to ineffective and inefficient criminal justice policies.
Influences on Criminal Justice Policy
Given the pronounced increase in calls for greater government accountabil-
ity and evidence-based practice, why does so much of the criminal justice
system and the laws, practices, programs, rules, and protocols that constitute
it remain unexamined and largely hidden in the equivalent of a “black box”?
Why, more generally, is there a seemingly large disjuncture between calls for
accountability and evidence-based policy and the realization of these calls
through the research base necessary to have accountability or to identify
evidence-based policies that are effective and cost efficient?
Many scholars have tackled these questions, including the broader one
of why any criminal justice policy is adopted.12 In the following discussion,
I describe several possible explanations. In so doing, I recognize that any
adequate account about the emergence of specific policies or policy trends
typically must reference a multitude of social and economic conditions and
their interactions with one another over time.13 Nonetheless, this discussion
serves to highlight the many and varied forces that can conspire against ratio-
nal, evidence-based policy. In turn, it underscores the need for systematic
integration of evaluation research into policy development, implementation,
and assessment to help address this situation.
Politicization of Crime
One prominent explanation for why many criminal justice policies emerge
can be summarized in one word – politicization. From this perspective, policy
makers focus on crime to advance their interests. That is, they are motivated
more by the thought of political gain than by a sincere belief that crime will
be affected.14 Of course, many policy makers sincerely believe that crime
merits attention not because of any political gain that may accrue to them
but because, in their view, it constitutes a substantial problem.15
Nonetheless, the politicization of crime has featured prominently in many
compelling accounts of crime policies. What benefits, though, does this
strategy – what some scholars characterize as “symbolic politics” – confer
upon policy makers? Among other things, it may enhance their electability
IRRATIONAL CRIMINAL JUSTICE POLICY 21
and divert attention from other, more divisive social problems.16 It also,
as scholars have argued, may enhance state power and the interests of an
elite social class at the expense of the socially disadvantaged. For example,
David Garland, who has written at great length about the crime policies of
the 1980s and 1990s in the United States and in Great Britain, has noted:
“Crime – together with associated ‘underclass’ behaviors such as drug abuse,
teenage pregnancy, single parenthood, and welfare dependency – came to
function as a rhetorical legitimation for social and economic policies that
effectively punished the poor and as a justification for the development of a
strong disciplinary state.”17
Why would crime serve as a convenient target for generating political cap-
ital among policy makers and, more generally, for increasing state power?
Garland’s work highlights that the ideological rhetoric employed in policy
discussions in the 1980s and 1990s viewed individual behavior as resulting
largely, if not exclusively, from self-discipline and moral character, not from
the social contexts and conditions in which individuals reside. Such a view,
which represents a philosophical orientation more than a scientific one,18
dovetails with the more general political ideologies of the conservative gov-
ernments that prevailed in both the United States and the United Kingdom
during these decades. Garland and others have argued that support for a
broad array of conservative policies at this time was facilitated by focus-
ing on crime and, in particular, by framing crime policy decisions using the
language of conservative political ideologies.
Crime served as a useful target for additional reasons. One is that vio-
lent crime worsened during the 1980s. Another is that little political fallout
occurs when policy makers focus on criminals. In fact, a failure to establish
a record of being tough on criminals can substantially limit a policy maker’s
career, as occurred when George H. W. Bush ran the now-famous Willy
Horton advertisement in his campaign against Massachusetts Governor
Michael Dukakis for the presidency. Horton, incarcerated in a Mas-
sachusetts prison for murder, was released on furlough during Dukakis’s time
as governor; while on furlough, he raped a woman.19 The advertisement was
widely viewed as contributing to Dukakis’s defeat. Another prominent exam-
ple arose in Texas in the early 1990s. Ann Richards, a Democrat, ran against
George W. Bush for governor in 1994 and, unlike Dukakis, attempted to
compete in part on the basis of her ability to be tougher on crime than her
opponent.20 Ultimately, she, too, lost out to a conservative candidate who
more strongly emphasized a tough-on-crime platform.
Arguments about the politicization of crime have emerged in an era in
which conservatives arguably have had a greater role in dictating the tenor
22 AMERICAN CRIMINAL JUSTICE POLICY
of criminal justice policy. Concomitantly, “get tough” approaches to crime
and punishment have predominated.21 It would be reasonable, therefore, to
assume that conservatives politicize crime and liberals do not. The assump-
tion would, however, be incorrect. Crime has been and can be politicized by
conservatives and liberals alike, as the Ann Richards example illustrates and
as research attests.22
False Dichotomies
Whether one accepts arguments about the politicization of crime, politics
may influence criminal justice policy in other ways. For example, the nature
of political debates, especially in contexts where two political parties pre-
dominate, tends to create false either-or dichotomies. In any democracy,
policy makers must strive to gauge the public will and determine which social
problems merit attention and what should be done about them. Necessarily,
then, policy makers must reduce a great deal of complexity to simplified
descriptions of the problems and the options for addressing them. Such an
approach unfortunately lends itself to creating overly simplified distillations,
and, indeed, to two-scenario options – there is X way of doing things or
Y way of doing things. That approach neatly accords with a conservative-
liberal dichotomy. Even so, it frames discussions and debates in terms that
frequently misrepresent reality.
One prominent example consists of the rehabilitation versus punishment
divide in American politics.23 Media accounts present anyone who promotes
rehabilitation as a liberal and anyone who promotes punishment as conser-
vative. The problem lies in the fact that few policy makers hew exclusively
to one or the other dimension but instead differ in the extent to which they
support both approaches to managing and sanctioning offenders. In a class
I teach on juvenile justice, students frequently express surprise that the pub-
lic strongly supports punishment of violent offenders and that the public
also strongly supports providing rehabilitative services to such offenders.24
On the face of it, the students seem justified in their surprise. How could
the public support both punishment and rehabilitation? Observe, however,
that nothing about one view precludes the other. Consider, for example,
that parents typically employ many different strategies – including a diverse
array of “carrots” and “sticks” – for managing children who break rules and,
more generally, for socializing them into the ways of the world. Few of us
would level the critique that doing so is necessarily inconsistent or odd. By
extension, there should appear nothing especially notable or contradictory
about the public supporting diverse approaches to addressing juvenile or
adult crime.
IRRATIONAL CRIMINAL JUSTICE POLICY 23
Such nuances frequently get lost in the policy-making arena, which all
too often glosses over nuance and substitutes in its place dichotomies that
not only simplify but also distort public views. To illustrate, a policy maker
who holds the view that punishment and rehabilitation should be weighted
equally may nonetheless feel compelled to emphasize one more than the
other. The tenor of a political debate may require such a packaging of one’s
views. Recent news accounts about, say, a felon who committed a vio-
lent crime while on probation, may force policy makers to articulate more
extreme versions of their viewpoints. As the Willy Horton example illus-
trates, such possibilities are far from hypothetical. During the 1980s and
1990s, it would have been difficult for many policy makers to be elected or
reelected if they argued for policies that equally balanced rehabilitation and
punishment.25
Long ago, Benjamin Franklin held up as a virtue the notion that we should
do everything in moderation. Following that dictum may not always lead to
good outcomes, but in some cases it would appear to be the better part
of wisdom. In the case of criminal justice policies, extreme policies con-
stitute the equivalent of stock-market speculation, where you sink all your
eggs into one company’s stock in the hopes that it will produce fabulous
returns. That may happen. However, it may not, and on the face of it, such
returns seem highly unlikely. Criminology offers little by way of research
that establishes whether punishment or rehabilitation produces the most or
greatest impact.26 Clearly, punishment seems like the hands-down winner if
the goal is retribution. Not everyone weights retribution in the same manner,
however. More relevant is the fact that if our goal is reduced recidivism, the
research evidence to date would suggest that either can be effective, depend-
ing on how they are implemented. That is, punishment can reduce recidi-
vism, but it also may increase it, and rehabilitation may reduce recidivism,
but it also may have no impact. Much rests on the precise type of punishment
or rehabilitation and how exactly it is implemented.27 In short, when policy
makers create or are pushed into accepting false dichotomies, the likelihood
increases that ineffective and inefficient criminal justice policies will emerge.
Swings from One Extreme to Another
The false dichotomy problem is compounded by a similar yet slightly dif-
ferent political dynamic. Specifically, the nature of many political systems,
and certainly of America’s political system, leads to dramatic swings in pol-
icy. In the United States, for example, the country’s crime policy approaches
have changed dramatically from one era to the next, most recently transition-
ing away from the rehabilitation-oriented, crime-prevention approaches that
24 AMERICAN CRIMINAL JUSTICE POLICY
prevailed in the 1960s and 1970s and toward the more punishment-oriented
approaches that have prevailed since.28 Thomas Bernard has illustrated the
problem in his account of the juvenile justice system, noting that, regardless
of juvenile crime trends, policy makers and the public become increasingly
disenchanted with the current set of policies in place and substitute in their
place policies that lie at the other end of the philosophical spectrum. As a
result, juvenile justice tends to cycle back and forth between lenient policies
and harsh, punitive policies.29 Such transitions frequently occur with little
to no assessment of the precise problems, the effectiveness of the current
set of policies, or the best mix of strategies for addressing crime and improv-
ing criminal justice operations. The end result is a costly transitioning from
one set of approaches to another and the whole-cloth adoption of many
new strategies that have been unevaluated and that, after implementation,
remain so.
As the preceding discussions have indicated, the latest swing in American
criminal justice policy has been toward “get tough” punishment-oriented
philosophies. The effect of this swing arguably has been and will be greater
than earlier ones given the dramatic growth in America’s prison population
and the attendant fixed commitment of resources for incarceration. It is rela-
tively easy to dismantle a particular law or program. Politically, however, it is
not easy to generate support for dismantling prisons, and indeed, one rarely
reads accounts where a given state’s prison capacity declines. In the past
decade, many policy makers have derided the growth in prison populations,
noting that it cannot be sustained and calling for “get smart” rather than
“get tough” options. Even so, prison populations have steadily continued to
grow.30 One benefit of the situation may be that it reduces the likelihood
of a dramatic swing toward a different set of policies. Yet it also reduces
the ability to achieve what might constitute a more balanced and ultimately
more effective portfolio of strategies for managing, reducing, and preventing
crime and for achieving justice.
Bad Cases Make for Bad Policies
Another influence on policy is a political dynamic in which cases that are
not representative of most others serve as the basis for new laws and poli-
cies. The expression “bad cases make for bad laws” captures this idea. In
any given year, atrocious examples – such as the Willy Horton case – exist
of the criminal justice system having failed. The problem arises from the
fact that virtually any policy, no matter how effective, will include failures.
Consider that, on average, and as shown in Figure 2.5, roughly two-thirds
IRRATIONAL CRIMINAL JUSTICE POLICY 25
of individuals released from prison will be rearrested within three years.31
Suppose a program in a particular state could reduce that rate of recidi-
vism to 50 percent. Such a reduction would be greater than many of the
best programs.32 Still, a large number of released inmates would still go
on to commit more crime, providing endless fodder for complaints that
somehow the criminal justice system is “broke” and requires fundamentally
new responses. Of course, if one is going to complain, it helps to focus on
the most extreme cases. The problem lies in the fact that such cases not
only always occur but also are just that: extreme and not representative of
the overwhelming majority of offenders or cases that enter the juvenile or
criminal justice systems.
This situation is complicated by the fact that policy makers attempt to
respond to the public and, at the same time, frequently must respond to
issues as they are depicted in media accounts. So, if the media, as much
research attests, is biased toward publicizing the most sensational crimes,
policy makers feel compelled to respond.33 The public’s lack of understand-
ing of many aspects of the criminal justice system, including the amount of
crime and the levels and quality of punishment and rehabilitative services,34
further compounds this problem. A dynamic thus ensues in which sensa-
tional cases, not the everyday ones, drive criminal justice policy. Indeed,
one might argue that a “perfect storm” of distortion emerges because many
public opinion polls ask only a few questions, focusing on sensational cases,
and then the media publicizes these responses. Policy makers proceed to
interpret such results as representative of public opinion about crime and
its solutions, although the findings speak only to public views about a very
particular type of crime.
In reality, public views about crime and justice are complicated, nuanced,
and highly variable depending on such factors as the issue involved and
the nature of the question wording and the response options. For example,
support for the death penalty drops roughly 20 percentage points when
respondents are asked to express their level of support in a context where
life in prison without the option of parole is included as part of the set of
options for sanctioning murderers.35
Symbolic Gestures
Paralleling and contributing to these problems is a situation in which policy
makers frequently feel pushed into doing something – anything, in fact –
that demonstrates their responsiveness to crime as a problem.36 The result
can be a penchant for responding to the latest crisis with some type of
26 AMERICAN CRIMINAL JUSTICE POLICY
new, and typically extreme, response rather than to deliberate assessment
of the problem and what can and should be done about it.37 In recent
decades, for example, many new penalties have been imposed on convicted
felons, creating a penumbra of “invisible” or hidden punishments, such as
restrictions on employment, housing, welfare, and voting, all of which go
well beyond the traditional notion of having inmates serve their time and
then reenter society as citizens with the full set of rights they had prior
to incarceration.38 Perhaps such restrictions were needed to create a more
powerful general deterrent effect to would-be offenders, and perhaps they
help to reduce the recidivism of the released prisoners. There is, however,
little theoretical or empirical research to support such a claim.
These additional invisible punishments arguably emerged not from a con-
sidered assessment of the need for them or their effectiveness, but rather
from policy makers’ desire to provide symbolic gestures of their responsive-
ness to public concern about crime. Here, again, it is important to recognize
that such arguments assume a level of political calculation that does not nec-
essarily accord with reality. For example, many policy makers clearly want to
serve the public interest and do so with a sincere commitment to pursuing
policies that they feel are needed and will be effective.
Public Opinion and Policy Makers’ Misunderstanding of It
Reviews of public opinion research consistently reveal that public views
about crime are, as discussed earlier, complicated and nuanced.39 Studies
typically show that the public supports a range of strategies, some reha-
bilitative in nature and some punishment oriented, for reducing crime and
that the level of support varies over time.40 They show that the majority of
the public views prison as a breeding ground for crime but that they also,
while supportive of rehabilitation,41 have doubts about the effectiveness of
rehabilitation as it occurs in practice. As Julian Roberts and Mike Hough
found in their review of public opinion research on views toward rehabilita-
tion in the prison system, “People around the world support the principle
that prisons should rehabilitate offenders, however, they do not believe that
in practice it succeeds in doing so.”42
The public also tends to know very little about the criminal justice system
as it operates in practice. With respect to prisons, for example, “the public
around the world underestimates the severity of life inside prison.”43 Indeed,
many view prisoners as having an “easy ride” that amounts to a vacation with
no work responsibilities and innumerable opportunities to play and watch
television in well-heated or air-conditioned housing.44 Not surprisingly, then,
IRRATIONAL CRIMINAL JUSTICE POLICY 27
studies frequently show that the public supports lengthier and tougher prison
sentences.45 The unfamiliarity with criminal justice extends to more than just
prison, however. As Francis Cullen and his colleagues have noted, for many
areas of criminal justice, “including knowledge of trends in crime rates, of the
prevalence of violent crimes, of recidivism rates, of specific criminal laws,
of legal reforms, of legal rights in the criminal justice process, and of the
extent to which the insanity plea is used successfully – the lack of knowledge
[among the public] is widespread.”46 This lack of knowledge “allows cynical
use of simplistic slogans and policies that respond to the public’s emotional
needs but do not address the substantive challenge,” as Alfred Blumstein
has noted.47
Given this lack of knowledge and the complexity of public opinion,48 it
should not be surprising that the views of the public do not readily translate
into simple either-or (e.g., rehabilitate or punish) options. As a case in point,
studies find that even when the public believes prison time is too lenient and
filled with too many amenities, they do not “necessarily want to make it
more aversive,” but rather want to do away with idleness and replace it with
work.49
Policy makers appear to operate within a political context in which the
complexity of public views sometimes must be downplayed or ignored.50 As
but one example, many of the “get tough” reforms of the 1990s emerged
from policy makers’ assumption that the public called for such reforms and
not for any other approaches. However, studies have shown that policy
makers overestimated how much the public wanted punitive sanctions and
underestimated how much they wanted such things as vocational training;
conjugal visits for inmates; counseling and therapy; and, more generally,
rehabilitation.51 One striking example of this phenomenon can be seen
in a study of Michigan policy makers, which found that “while only 12
percent of policy makers believed the public would support rehabilitation
as a criminal justice objective, in reality two-thirds of the public took this
view.”52
In a democracy, we expect that the public will should at least inform policy
discussions and debates.53 It should not necessarily dictate policy. Even so,
public views are foundational to democracy. For that reason, one of the
more striking findings from public opinion research is the fact that policy
makers frequently misestimate or distort, whether consciously or not, public
views about crime and its causes and solutions. Given the way in which
political decision making occurs, such misunderstanding helps contribute
to policies that not only reflect extreme and unrepresentative cases but also
fail to reflect public sentiment.
28 AMERICAN CRIMINAL JUSTICE POLICY
Belief in “Silver Bullet” Causes and Solutions to Crime
Another factor that contributes to ineffective policy is the seemingly
widespread belief among policy makers that there exist “silver bullet” causes
of criminal justice problems and “silver bullet” solutions to them.54 A silver
bullet approach is effective when, among other things, the following con-
ditions hold: the targeted cause is truly a cause of the outcome of interest
(e.g., criminal behavior), the cause is widespread, and the cause is easily
amenable to modification. These conditions rarely if ever hold true in crimi-
nal justice, and yet many – although certainly not all – policies take a largely
single-minded focus toward reducing crime.
To illustrate, a plethora of laws and programs have emerged that focus on
illegal drug use and selling. Many of these efforts create enhanced penalties
for such crimes, or for committing other crimes while using illegal drugs, and
are widely viewed as contributing to the dramatic growth in the correctional
system in recent decades.55 At the same time, drug courts, which specialize
in handling drug offenders and drug-using offenders, have proliferated since
the early 1990s.56
In part, the focus on drugs appears to be fueled by a belief that illegal drug
use causes crime and that it does so in a dramatic way. Notably, however, it
remains unclear how strong the relationship between individual drug use and
offending is and whether the relationship is causal.57 In asserting a causal
relationship, one might point to the fact that many people in prison have or
had drug problems or were using drugs at the time of their offense.58 If that
were the only problem that prisoners exhibited, the causal claim would be
easier to accept. Yet the reality is that the profile of the typical prisoner leads
to a host of factors – such as mental illness, homelessness, unemployment,
abuse59 – that could be the cause of their behavior, including their drug use.
Complicating matters is the fact that, while various sanctions and treatments
can reduce drug problems; resolution of these problems is not simple; relapse
is common, and many interventions are costly, especially if implemented as
intended.60 In short, two of the conditions for an effective silver bullet solu-
tion do not appear to be present when it comes to drug-related offending.
At the same time, a large body of research points to many other factors
that cause crime and to the likelihood that the most effective approaches to
crime prevention involve a focus on multiple causes of offending.61
To be clear, there seems little doubt that drugs play some role in crime
and offending. It remains unclear, however, whether a largely exclusive focus
on drug-related crime would substantially reduce overall crime or recidivism
rates.62 Certainly, a balanced approach to reducing crime might involve a
IRRATIONAL CRIMINAL JUSTICE POLICY 29
focus on illegal drugs. The silver bullet approach, however, places a pri-
mary emphasis on drugs. For example, when a criminal justice system gives
priority to drug treatment, it typically must reduce its emphasis on other
approaches to resolving crime, if only because most systems operate within
a zero-sum environment in which a limited pool of resources exist. So, if
a jurisdiction decides to create a drug court, it necessarily will have fewer
resources to devote specialized attention to other populations.
It may be argued that many jurisdictions and many correctional systems
embrace a diversified portfolio of approaches, not a silver bullet approach, to
fighting crime. In reality, however, priority is frequently given to a select few
approaches. Consider, for example, that while drug courts have proliferated
nationally, few other such specialized (e.g., mental health, drug, community)
courts have enjoyed such popularity. Consider, too, at a national level, the
dramatic increase in prison systems, which is tantamount to a belief that
increased incarceration, more than a range of other approaches, can sub-
stantially reduce crime. Here, again, the point is not that such approaches
are ineffective. Rather, it is that much criminal justice policy making is
aimed at finding silver bullet solutions rather than at creating comprehen-
sive, research-based approaches to crime reduction.
Limited Production of Policy Research
To this point, many of the factors that have been described as influencing
criminal justice policy and that serve as barriers to research-based criminal
justice policy making have centered on political factors. What about research
itself, or, perhaps more precisely, the lack of research, as a barrier to better
policies? As a general matter, most accounts – including American Society
of Criminology presidential addresses – wax pessimistic about the influence
of research on policy. They point to innumerable instances in which policies
get adopted with little to no attention to prior research and in which poorly
conceived policies continue unabated and unevaluated. Even so, a number
of scholars have pointed to evidence that, at least on occasion, research
influences policy.63 If we were to average the two views, the situation might
be aptly described as one where, by and large, research provides little by way
of a positive influence on policy but where many exceptions exist. Assuming
that this assessment is correct, the question emerges as to why research has
had relatively little influence on policy. Here, I will briefly touch on some of
the major factors scholars have identified.64
The first and perhaps most important problem is the lack of an institu-
tionalized foundation for systematically integrating evaluation research into
30 AMERICAN CRIMINAL JUSTICE POLICY
decisions about criminal justice policies and for monitoring and assessing
the criminal justice system as a whole. For example, federal funding for
crime and justice research is minimal relative to investments in other policy
areas.65 At the state, county, city, and municipal levels, few agencies allot
much funding for research, and what funding exists typically is provided for
compiling highly descriptive annual reports that say little about the need
for or the design, implementation, effectiveness, or efficiency of a range of
policies.
This situation is compounded by a second problem: specifically, policy
evaluation research – what sometimes gets referred to as “applied research”
because of the focus on applying research to policy – traditionally gets short
shrift within universities.66 Many factors have contributed to this dynamic.
By historical standards, criminology is a newcomer in the academic world. As
Joan Petersilia has noted, criminology was not offered as a major in universi-
ties until the early 1930s, and it was not until 1950 that the first criminology
program formally emerged.67 As a newcomer, considerable pressure existed
to elevate the status of the discipline to a “science,” which meant focusing
on questions about the causes of crime and not necessarily the solutions
to it.68
Not surprisingly, such circumstances can lead to a bias away from policy
research and toward so-called basic research. In turn, not only do university
scholars tend to shy away from policy research, but they also tend not to
train graduate students in the art and science of evaluation research.69 In
addition, the orientation toward basic research diminishes the likelihood
of developing strong institutional ties between university researchers and
local, state, and federal criminal justice system agencies. Notably, when
criminology programs first emerged, practitioners held more sway in the
classroom, “but since the academic has largely replaced the practitioner in
the classroom and in research, the link has grown weaker and, with it, that
kind of immediate influence.”70
Even when scholars undertake policy-related research, they typically do
not translate their findings in a way that is accessible to policy makers and
practitioners. Other researchers have the training and time to sift through
myriad statistical analyses; policy makers and practitioners typically do not.
Even so, translating sophisticated analyses so that nonresearchers can easily
digest them can be challenging, especially if, as is frequently the case, many
caveats and limitations bear emphasis. To illustrate, a study may show that
a given program reduces recidivism, but that finding may apply only to
similar programs, such as those that serve similar clients. It may be that the
IRRATIONAL CRIMINAL JUSTICE POLICY 31
effect was not particularly large. It may be that only drug recidivism was
reduced but not violent or property recidivism. It also may be that the effect
only emerged when participants fully completed all aspects of the program.
Scholars are trained to give considerable weight to such nuances and to stick
closely to the limits of what the type and quality of data and analyses allow.
Consequently, it can be a struggle to try to discuss the results in a way that
runs counter to their training. That struggle can turn into resistance if they
feel that policy makers or the media purposely or unwittingly distort such
results.
A third problem is the limited funding for criminal justice and crime policy
research. Petersilia has remarked that “the federal government is, by orders
of magnitude, the largest funder of research on criminal justice policy.”71
However, federal funding for criminal justice research has been nominal
and remains so. Consider the funding of the National Institute of Justice
(NIJ), which serves as the main federal agency focused on criminal justice
evaluation and policy researches. Writing in 1995, Blumstein and Petersilia
observed that NIJ’s budget “has been essentially flat (with slight declines
in real terms) since 1981 and has stayed in the range of about $20 million
to $30 million since then – well short of a priority.”72 By 2008, more than
a decade after that observation, funding for the agency had increased to
$37 million,73 a modest increase but still well short of constituting a priority,
especially in a context where federal and state criminal justice expenditures,
well into the billions of dollars, escalated dramatically and where baseline
levels of funding were minimal compared to federal investments in other
social policy arenas. The point was made bluntly by Petersilia, who, in 1991,
commented that “for every U.S. citizen, federal funders spend $32 on health
research, but only 13 cents on criminal justice research.”74 That situation
remains largely the same today despite the dramatic increases in violent
crime and in criminal justice system expenditures that occurred during the
decades after Petersilia made this observation.75
A fourth problem has been the relative lack of investment in high-quality
impact evaluations that rely on experimental designs. The gold standard for
impact evaluations is the experiment, precisely because, if well conducted,
the results can be interpreted in a straightforward manner as indicating that a
program “works” or does not. Most criminal justice policies go unevaluated,
and the few that are evaluated typically get examined using nonexperimental
research designs.76 The use of the latter type of designs can be problematic
because they tend to find positive impacts of programs in cases where the
impacts are not real.77 Consequently, the results of many studies rest on
32 AMERICAN CRIMINAL JUSTICE POLICY
shaky foundations and so lead to a situation in which researchers must be
highly cautious in reporting results. A typical example, created for illustrative
purposes here, would be a study in which the hypothetical conclusion reads
as follows:
The results here suggest that program X may reduce the recidivism of moderate-
risk male offenders but not necessarily affect the recidivism of high-risk or
low-risk offenders or of female offenders. In addition, the results should be
interpreted with caution given that the study sample consisted of a highly select
group, including inmates who volunteered to participate in the study, and given
that many acts of recidivism may have gone unreported to law enforcement.
Indeed, because the study could not fully address important potential selection
effects, the estimated effects of the program may be biased. Put differently, were
the selection effects better addressed, the study may have found no difference
between the treatment and comparison groups in their rates of recidivism.
A policy maker would understandably view such an account as not especially
helpful in making a decision about whether to close down, continue, or
expand the program.
The problem lies not just with a lack of experiments in criminal justice pol-
icy research but also with a lack of high-quality quasi-experimental research
(i.e., studies that attempt to approximate an experimental design).78 All too
often, the design of such research is weak, and the result tends to be a
situation in which the positive impacts of a program are overstated.79 Fre-
quently, too, experimental and quasi-experimental research designs rely on
sample sizes that are too small to allow one to detect anything other than
an extremely large impact.80 Many policies and programs that may be effec-
tive thus are reported not to be. (These and related issues are discussed in
Chapter 7.)
A fifth problem is that researchers frequently focus their attention on the
problems and policies that policy makers emphasize or that constitute the
“hot topics” of the day. In so doing, they ignore a wide range of impor-
tant policies and policy emphases. Francis Cullen and Paul Gendreau have
drawn attention, for example, to the fact that in the 1970s, a period when
rehabilitation increasingly was viewed as ineffective, “the study of correc-
tions became largely the study of social problems,” and so “criminologists
paid scant attention to ‘what works’ to change offenders.”81 Indeed, this bias
was institutionalized through publishing biases: “[scholars] were praised and
rewarded with opportunities to publish their research when they could show
that an acclaimed program did not live up to its billing.”82 Criminologists
thus tended to emphasize the negative over the positive and did so in part
because of a political climate in which rehabilitation came into disrepute.83
IRRATIONAL CRIMINAL JUSTICE POLICY 33
Other barriers related to the production of research exist as well. For
example, evaluations sometimes take years to complete, but policy makers
frequently need and want results sooner. This problem is amplified by the
lack of institutionalized linkages among universities, research organizations,
and the agencies that constitute the criminal justice system, in part because
the absence of such linkages delays access to data.
All is not bleak, however. As Petersilia has emphasized, research can and
does exert a positive effect on policy. It has “helped shape the way police
are deployed,” “demonstrated the effectiveness of career criminal programs
in prosecutors’ offices,” “improved the ability to classify offenders and to
predict recidivism,” “provided information about the relationship between
drug abuse and crime” and “participation in rehabilitation programs [does
not] necessarily [reduce] recidivism.”84
That said, substantial improvements could be made for potentially little
cost. In the book’s final chapter, I will provide a discussion of specific steps –
such as the use of systemwide, comprehensive performance measurement
and monitoring – for making such improvements. In so doing, I will spotlight
research-based efforts that hold promise for enhancing the contribution and
role of research in criminal justice policy.
Conclusion
America stands at a unique juncture in the history of its criminal justice
system. Unprecedented growth in this system and in criminal justice expen-
ditures, along with an ever-growing panoply of policies, create substantial
concerns about whether the growth, expenditures, and policies make sense.
The evaluation hierarchy highlights some of the critical concerns. Have the
growth and the investments been needed? Do the policies rest on sound
theoretical grounds? Have they been well implemented? Do they achieve
expected outcomes? And have the investments been allocated to the prob-
lems and policies where the greatest gain will accrue? By and large, and as
subsequent chapters will argue, research is silent about such questions as
they relate to the criminal justice system’s many policies, practices, proto-
cols, and rules. In those cases where research exists, it sometimes waxes
positive. However, it all too frequently suffers from critical problems or
provides pessimistic or equivocal assessments.
This situation is cause for particular concern, especially given the unprece-
dented growth in the size and costs of the American criminal justice system
and the more than 735,000 prisoners released back into society annually,
more than two-thirds of whom will recidivate. The very real possibility exists
34 AMERICAN CRIMINAL JUSTICE POLICY
that this growth and the many policies enacted in recent decades have done
little to make the public safer or as safe as might be possible with investments
in other approaches to crime control and prevention.
Juxtaposed against this situation is a context in which myriad forces –
such as the politicization of crime and the belief in silver bullet solutions
to crime – lead to continued creation of and investment in criminal justice
policies that may not be the most effective or efficient. Policies will always
result from a constellation of factors. Nonetheless, efforts to make criminal
justice more accountable or to place it on a more evidence-based foundation
will not likely succeed without more and better research. What is needed
is a systematic approach to developing, monitoring, and assessing criminal
justice policy. That approach is the focus of the next chapter.
Discussion Questions
What have been the prominent crime and criminal justice system trends
in recent decades?
What are the risks of adopting unnecessary, ineffective, or inefficient crim-
inal justice policies?
How can greater government accountability or reliance on evidence-based
criminal justice policies be achieved?
What factors influence the creation of criminal justice policies, and which
ones matter the most? Why?
What role can evaluation research play in contributing to accountability in
the criminal justice system and to more effective and efficient policies?
