PAPER 1

 you will locate and describe a criminal justice policy of your choice, and think through the important components of that policy’s context, intentions, probable effectiveness, and apparent costs.

The paper is designed to be an exercise in critical thinking, not a “book report” that requires you to locate published research related to the policy (though you may do so, if you like). The main objective here is to get us all thinking about where CJ policy comes from, how it changes or persists over time, and whether or not we agree with Mears (2010) in describing it as “irrational.”

Mears discusses the modern landscape of “irrational” criminal justice policy, and several of the supplements that you were assigned reinforce the idea that criminal justice policies are based on biased perspectives, incomplete information, and flawed logic. The politicization of crime, along with various social, economic, structural, and other external factors all contribute to sub-optimal criminal justice policy.

Keeping in mind what Mears and the other assigned authors have written, your task in Paper #1 is to locate a real-world example of contemporary policy and to describe its purpose along with its relative strengths and weaknesses. Hopefully you will also begin to see linkages between modern policy examples and the context from which they emerged.

To begin, select a criminal justice policy of your choice, one that is intended to address crime prevention in some phase of the justice system. For example, you might search the web for examples of “Three Strikes” laws, or for specialized drug courts, or for sex offender registries. This is intended to be broad to give you maximum latitude to select an example you prefer.

In your paper, describe the following:

What is the policy you chose, and what is its context (e.g., federal, state, local)? What type of crime(s) is the policy intended to address? How is the policy intended to work?

What is your initial assessment about the effectiveness of the policy that you chose? Does it appear to reduce crime? If so, where and how? If not, why not?

Are there any available descriptions of the policy’s costs? How do you think your policy example compares to the costs of alternative criminal justice responses?

Throughout your paper, you may find it helpful to use compare and contrast statements to help support your points. For instance, if you write that the policy example you selected is probably effective at reducing crime, are you comparing that conclusion to the alternative of doing nothing, or to the alternative of other similar policies? Elaborate and provide a rationale for your position throughout your paper.

4

Needs Evaluations

F EW OF US WOULD BUY SOMETHING IF WE DID NOT NEED IT FOR SOME

purpose. We might be impulsive perhaps, but even impulsive pur-
chases typically meet some real or perceived need. Of course, our reasoning
may be flawed. But that in no way detracts from the notion that real or
perceived needs drive much of our decision making, especially when the
decisions entail large amounts of resources or sums of money.1

A similar assertion can be made about criminal justice policies. That is,
they presumably emerge from concern about a real problem, such as an
increase in crime, or at least from what policy makers perceive to be a prob-
lem. Even so, what does it mean to say that there exists a sufficiently large
problem, or that there exists a small but nonetheless important problem, as
to create a need for a particular policy response, such as more prisons? What
criteria should we use for assessing the amount of a social problem or the
need for a particular policy response?

This chapter examines these questions by describing needs evaluations
and what they entail. As will be detailed in this chapter, needs evaluations
help to identify whether a problem exists and, in turn, whether and what type
of a policy response is indicated. They provide guidance on prioritizing dif-
ferent problems. They point to research gaps that must be addressed before
it can be determined that a policy response merits implementation and, if a
response is warranted, which type. They highlight the dimensions relevant to
assessing success. And, more generally, they clarify policy discussions. The
chapter describes these benefits and provides two case studies – one focused
on mass incarceration and the other on sex crime laws – to illustrate needs
evaluations. Drawing on these case studies and other examples, the chapter
concludes with a discussion of whether the prominent crime policies that
populate today’s current criminal justice landscape are needed.

51

52 AMERICAN CRIMINAL JUSTICE POLICY

What Is a Needs Evaluation?

Needs Evaluation: Step 1 in the Hierarchy

A needs evaluation constitutes the first step in the evaluation hierarchy,
as shown in Figure 2.1 and described in Chapter 3. The reason should be
clear: without an established need – that is, without clear evidence of a
social problem – why proceed to develop and implement a policy? To use an
analogy, if we don’t have cancer, why spend money for a treatment that we
don’t need, that may draw funds away from real problems that we may have,
and that may actually cause more harm than good in the absence of the
disease? The answer? No good reason exists to do so. Treatment should be
reserved for real diseases. So, too, with social policies – we should intervene
only when a real social problem exists. To do otherwise risks misallocating
scarce resources and failing to address as forcefully as we could real or
pressing problems.

Crime clearly constitutes a social problem, so there perhaps would seem
to be little risk involved in misdiagnosing it. In reality, and as will be dis-
cussed throughout this chapter, defining need can be quite complicated.
How much crime exists, whether crime has increased, and what types of
crimes have increased are all critical considerations. In addition, needs may
arise that center around specific facets of the criminal justice system. For
example, perhaps prison officers receive insufficient training for work with
maximum-security inmates. Perhaps caseloads among defense counsel, pro-
bation officers, prosecutors, or judges greatly exceed reasonable amounts.
Many other problems exist in criminal justice, some of them directly involv-
ing crime (e.g., an insufficient response to dramatic increases in violence)
and some reflecting the specific demands and activities of components of the
criminal justice system (e.g., inmate attacks on prison officers). An effective
response requires first identifying these specific problems and their contours.

Given their importance, needs evaluations would seem to be likely can-
didates for the “most common type of evaluation” award, if such existed.
Unfortunately, no census of types of evaluation exists. However, even a cur-
sory review of the literature reveals that needs evaluations rarely occur in
criminal justice. By contrast, implementation and impact evaluations, dis-
cussed in subsequent chapters, occur more frequently. Why?

A number of factors may be relevant. Many researchers lack any formal
training in evaluation research and so hew to the two types that they most
frequently see in journals and reports. Criminal justice administrators, too,
may lack training in evaluation research and so not be sensitized to the

NEEDS EVALUATIONS 53

importance of assessing the wide range of potential problems within their
purview. In addition, policy makers have little to no training in evaluation
research. Perhaps more important, they appear to be propelled toward
action even when doing so is unnecessary or a mistake, what medical re-
searchers term action bias.2 For example, rather than slow down and evalu-
ate a problem or refrain from acting until and unless a compelling case for
a particular intervention emerges, a physician makes a diagnosis and starts
treatment. Similarly, policy makers or criminal justice administrators may
proceed with promoting a new policy or increasing an existing one (e.g.,
incarceration) even though they lack evidence about the precise nature of
the problem they wish to address, its causes, or how best to reduce it.

Other factors may come into play as well. For example, each year thou-
sands of new criminal justice policies and programs emerge. Because they
already exist, researchers might reasonably prioritize implementation or
impact studies rather than step back and examine whether a need for the
policies existed in the first place. Within the scholarly community, research
that simply describes a problem rather than explains its occurrence may
not be viewed as an important endeavor. The more prestigious criminol-
ogy and criminal justice journals tend to emphasize theoretical work, and
tenure and promotion frequently depend on publication in such journals.3

As a result, many researchers may feel pressured to avoid conducting studies
that “merely” describe a social problem. Still other factors may be relevant
as well. Regardless, needs evaluations can serve a critical – indeed, a foun-
dational – role in improving criminal justice policy.

Defining the Problem: Size, Trends, Location, Causes

Social policies serve to address problems or what alternatively can be called
needs. Crime constitutes one prominent social problem or need. Because it
is ubiquitous, we might conclude that no assessment of need is warranted.
Would that matters were so simple. Defining a social problem involves many
considerations, including reference to the (1) size, (2) trends, (3) location,
and (4) causes of it.

AN ILLUSTRATION. To illustrate this point, consider the differences in the
following two scenarios, depicted in Table 4.1, that assume cities with the
same population size and characteristics. In City A, the number of homi-
cides rose from sixty to ninety during a two-year period, an increase of thirty
homicides, or, alternatively, a 50-percent increase in homicides. Assume
that forty-five of the year-2 homicides involved some type of gang-related

54 AMERICAN CRIMINAL JUSTICE POLICY

TABLE 4.1. The crime problem in two cities

City A City B

Year Year Change Change Year Year Change Change
1 2 (No.) (%) 1 2 (No.) (%)

Homicide 60 90 +30 +50 20 30 +10 +5

0

Gang Homicide 10 45 +35 +350 1 1 None None
Burglary 1,000 700 −300 −30 1,500 2,400 +900 +60

activity, up from ten the year before. In this scenario, gang-related homi-
cides increased by 350 percent from one year to the next. Assume that
almost all of the gang-related homicides occurred in three neighborhoods.
Finally, assume that during the same time period (i.e., from year 1 to year
2), burglaries declined by 30 percent, from 1,000 to 700.

In City B, let us assume that the number of homicides rose from twenty
to thirty during the same two-year period, an increase of ten homicides and
an overall increase of 50 percent. Almost none of the past-year homicides
appear to be linked to gang-related activity, and no evidence of a trend
in gang-related homicides exists. However, three of the homicide events
accounted for ten of the thirty year-2 homicides. These were events in which
one person killed four people and the two other killers murdered three peo-
ple each. In prior years, few homicide events involved the death of more than
one person. Notably, most of the homicides seem to be relatively evenly dis-
persed throughout the city. Finally, during the same time period, burglaries
increased by 60 percent, from 1,500 to 2,400.

A number of patterns in this highly simplified example can be identified.
City A clearly suffers from a large violent crime problem relative to City B.
Whether we look at year 1 or year 2, City A reported more homicides. In
addition, in absolute terms, City A experienced a much greater change in
the number of homicides – an increase of thirty homicides from one year to
the next as compared with the increase of ten homicides during the same
time span in the other city.

In this example, observe that despite the larger increase in homicides in
City A, both cities experienced a comparable percentage change – a 50
percent increase – in the number of homicides. How could that be? City
A experienced a greater number of homicides overall as well as a greater
number of new homicides from one year to the next. However, City B had
a relatively small number of homicides (twenty) in year 1. The result in such

NEEDS EVALUATIONS 55

situations is that a city can experience a relatively small absolute increase in
homicides from one year to the next (from twenty to thirty in the case of
City B), and yet report a dramatic percentage increase in such crime.

Review of the two cities’ crime statistics shows that, in City A, gang activity
is implicated in the increase in homicides and that much of the gang-related
murder occurred in a small handful of neighborhoods. In City B, however,
the homicide increase does not appear to be readily linked to a single factor
such as gang activity nor do the homicides appear to be concentrated in one
area. Indeed, closer inspection of City B’s crime statistics reveals a different
story entirely – a large proportion of City B’s year-2 homicides, one-third of
them, resulted from just three events. That fact suggests that, in reality, no
dramatic increase in violent killers occurred in the city.

Finally, we can see that City A experienced a decrease in burglary during
the two-year period. By contrast, in City B, burglary not only increased, but it
did so more than homicide, whether measured in absolute or relative terms.
Specifically, from year 1 to year 2, there were nine 900 more burglaries, or a
60 percent increase in such crime.

This illustration serves to highlight a number of considerations related
to evaluating whether a social problem exists and, by extension, whether
there is a need for a policy response of some type. First, numbers matter.
More formally, the size of the problem matters. A big problem calls for more
attention, whereas a smaller problem calls for less. City A experienced a large
increase, in absolute and relative terms, in homicides, which would suggest
a need to respond. Ideally, of course, we could address every problem. In
reality, however, society typically must make triage decisions, focusing on
the more urgent problems first and then turning to those deemed to be less
urgent.

That challenge confronts City B, which faced an increase in both violent
and property crime. Yet how we should proceed is less clear. Yes, homicide
increased, but in absolute terms the increase of ten homicides is relatively
small – indeed, it is considerably smaller than the thirty additional homicides
that City A experienced. And it appears that most of the increase stemmed
from three offenders. In fact, if we count crime events rather than specific
victims, one might dispute whether any increase in violence occurred. At
the same time, any increase in homicide victims stands as cause for concern.
But what about the increase in burglary? City B was confronted by 900
more burglaries from year 1 to year 2, an increase of 60 percent. Burglary
does not compare with murder in terms of severity. Yet it would be odd
to allow such a dramatic increase in the frequency of property crime to go
unchecked.

56 AMERICAN CRIMINAL JUSTICE POLICY

Second, absolute and relative changes in the magnitude of a problem should
be taken into account. One or the other by itself can be misleading. For
example, if we focused purely on absolute numbers, City B has little basis
for focusing on homicide as aggressively as City A. However, if we focused
on trends over time, then both cities appear to be facing comparable prob-
lems in terms of increased violent crime. Consider the matter from a state
level. Both cities may claim that they have a comparable social problem – an
increase of 50 percent in homicides. State policy makers might be inclined
to treat the two cities equally, given this claim. However, when viewed in
absolute terms, City A unquestionably has the far more pressing homicide
problem. Which standard should local or state policy makers or law enforce-
ment agencies use to guide their decision making? If they use both, how do
they jointly consider the absolute and relative changes in homicide, relative
to burglary, in quantifying the scope of the problem? No simple solution
exists. Yet quantifying such dimensions as the absolute and relative magni-
tude of a problem – and comparing them over time across different areas
within and across cities – can ground policy discussions in a better under-
standing of the problems at hand.

Third, knowing something about the distribution of a problem in social
and geographic space can help us to delimit the contours of it and perhaps
in turn guide our thinking when developing a policy response. For example,
City A’s increased homicide rate appears to stem from gang-related activity
in a few select areas. That in turn suggests that a focus on gangs could
help reduce homicides and, more specifically, that a focus on gangs in those
areas could help reduce violent crime. Closer analysis of the problem may
prove these inferences to be false, but we at least have some reasonable
basis for creating a more targeted response. We could, of course, ignore
such information. However, we then would risk dispersing our efforts in a
diffuse manner to populations and areas that have not contributed to the
violent crime increase.

City B faces a potentially more difficult situation in that the increase
does not appear to stem from an underlying factor or to be concentrated
in one area. Thus, should the city decide to step up its efforts to combat
violent crime, it will want to develop a better understanding of what set
of factors has contributed not only to homicides in general but also to
the increase in homicides. At the same time, the fact that much of the
increase may stem from three homicide events suggests that caution should
be exercised in making too much out of the seemingly large percentage
increase in violent crime. For example, a large investment in fighting gangs
would appear unlikely to achieve much if anything by way of a reduction in
homicides.

NEEDS EVALUATIONS 57

The two-city illustration highlights the essence of needs evaluations.
Briefly, such evaluations aim to define a problem by (1) describing the
nature and extent of it in absolute and relative terms, including (2) how the
problem has changed, if at all, over time; (3) the distribution of the problem
among different populations and places; and (4) what caused it.4 Providing
insight into the possible causes of the problem is important not only because
doing so helps to describe the nature of the problem but also because it
helps to establish whether there is a need for a particular type of policy
response.

This last point bears elaboration. In the illustration, we can see that a more
refined analysis of City A’s crime statistics shows (1) gang activity to be a
significant contributor to violent crime and (2) particular areas to be host
to such gangs. Such information suggests that a more accurate depiction
of the crime problem in the city is not that homicides have increased but
rather that gang-related homicides in a select few areas have increased. That
framing of the situation describes a quite different social problem from one
where we say simply that “violent crime has increased in the city.” The former
depiction provides an arguably more accurate account of the problem and it
also provides guidance about how we might intervene to reduce homicide.
By contrast, the latter treats homicides in general as a social problem, which
seems to downplay that a more precise problem (i.e., gang-related violence)
exists. And, to make matters worse, it provides no guidance about how to
intervene.

Investigation of the causes of a problem can, of course, go much deeper.
For example, what caused the surge in gang activity? Was it a change in
drug market activity? An influx of new gangs? Increased competition in the
illegal drug market? A shift in the law enforcement department’s approach
to intervening with gangs or a shift in its emphasis on gangs to, say, one on
domestic violence or to sex crimes? Should clear answers emerge in response
to these questions, policy makers can develop more precise descriptions of
the problem and, in turn, how to respond. For example, if one gang was
the predominant instigator of gang-related violent crime, that suggests a
very different problem (and the need for a very different response) from one
where multiple gangs have been involved. In short, assessing the causes of
a problem is an essential step both in defining what the problem is and in
determining how best to respond.

THE NEED FOR ASSESSING MANY TYPES OF NEED. The previous example
focused on homicide and burglary. In reality, of course, many more crimes
exist. In addition, many problems arise in the criminal justice system that
relate not only to crime but also to providing justice and, more mundanely, to

58 AMERICAN CRIMINAL JUSTICE POLICY

executing everyday tasks (e.g., making arrests, defending and prosecuting
individuals).

The problems in criminal justice defy any simple categorization. Even so,
clarity about the types and levels of problems can help place policy on a
more rational foundation. At the broadest level, there is, of course, the fact
of crime. The two-city example underscores, however, that we typically will
want to disaggregate the problem of crime by examining specific types (e.g.,
violent, property, drug). We also will want to map crime and its change over
time as well as its distribution across different populations and places. For
example, are there “hot spots” of criminal activity? Or populations that are
“hot” in the sense that more crime occurs among them than others? At the
least, we will want to identify what, in each instance, has caused it.

The same observations hold for individual-level offending. As with eco-
logical-level crime (e.g., rates of homicide or burglary among cities), con-
siderable variation in individual-level offending exists, which has implica-
tions for describing what constitutes a “problem.” Here, consider the range
of ways in which we can measure individual-level crime problems. Partici-
pants in a program aimed at reducing recidivism could be asked whether
they committed any crime in the year after release. The outcome in this
instance is a simple binary one – a crime was committed or it was not. We
can measure offending in other ways, though. We could, for example, exam-
ine types of recidivism (e.g., violent, property, drug). We also could examine
the frequency of offending. Did participants in the program experience a
greater reduction in the amount of crime they committed compared to sim-
ilar inmates not exposed to the program? The diversity of offending might
be of interest to us. Did participants commit less violent crime compared
to similar inmates, or was the effect more diffuse, with the program reduc-
ing the occurrence of all types of recidivism? Criminologists increasingly
have become interested in examining the onset and desistance of offend-
ing, and, more broadly, trajectories of offending.5 So, a natural question is
whether programs that deal with offender populations effectively alter these
trajectories, such that participants become more likely to transition from
higher-offending trajectories to lower-offending ones.

Beyond the focus on crime, there exists, of course, the entire criminal
justice system and the varied set of policies and practices that constitute
it. These aim not only to reduce crime but also to mete out justice and,
more simply, to process large numbers of cases. Many types of problems can
be found in criminal justice. To illustrate, law enforcement agencies may
lack the resources to respond in a timely manner to 911 calls. They may
focus too much attention on some crime and not enough on another. Some

NEEDS EVALUATIONS 59

officers may abuse their authority. Others may be overworked. Domestic
violence victims may be reluctant to contact law enforcement agencies for
assistance. Communication and cooperation within and between units of
these agencies may be minimal. And so on.

Narrowing our focus somewhat, we can see that much the same can be
said of the courts – that is, a wide range of diverse problems can exist that
can undermine their effectiveness and efficiency. Prosecutors may lack the
resources to target high-impact crimes or they may expend resources on
questionable cases. They may be overly aggressive in seeking tough sanc-
tions or too lenient. Probation officers may produce low-quality case reports
for judges. They also may suffer from morale problems stemming from an
inability to execute the many and varied expectations placed upon them.
Defense counsel and public defenders may have caseloads that preclude
adequate representation of clients, and they may even refuse to take new
cases or may delay acting on them for many months. Judges within a given
jurisdiction may vary dramatically in how they sanction particular types of
cases. Victims may receive little attention or assistance from court personnel.
And so on.

Many problems also pervade, in varying degrees, correctional systems and
their component parts. Prison officers may be poorly trained for managing
violent inmates. The classification of inmates may rely on unvalidated assess-
ment instruments and there may be too few facilities for accommodating
different types of inmates. Prison graft may go largely unchecked. Unrealis-
tic policy makers’ demands may be placed on prison administrators. Parole
officer caseloads may be so high as to limit “face time” with parolees to
minutes each month. And so on.

“And so on” constitutes the operative phrase here. Each stage of the
criminal justice system entails enormous potential to do good and also to
cause harm. Many possible problems exist. In each instance, the size of the
problem may vary from the negligible to the large; it may have decreased or
increased over time (in small or large absolute or relative amounts); it may be
widely dispersed or may be concentrated among a few select groups, units,
places, and the like; and it may stem from a few specific causes or a wide
range of them. In such a context, operating without any systematic needs
evaluations of actual or proposed policies amounts to flying an airplane at
night without any equipment and with one’s eyes covered – an accident
almost inevitably will occur. More concretely, it results in many problems
going unidentified and unaddressed and it means that policy makers likely
place too little or too much emphasis on the few problems that do get
documented.

60 AMERICAN CRIMINAL JUSTICE POLICY

Determining the Type and Level of Response Needed

ASSESSING THE NEED FOR AN EXISTING OR PROPOSED POLICY

RESPONSE. To this point, the discussion of need assumes a focus on some
type of social problem. That is, we begin with determining whether a social
problem exists and whether, by extension, a need exists that should be
addressed by policy. Consider, however, a situation where we approach the
matter from the opposite direction. Specifically, we begin with an existing
or proposed policy and we determine whether a need for this specific policy
exists. This is the second major type of needs evaluation. Here, rather than
determine whether a social problem exists and then proceed to a broad-
based focus on possible policy responses, we instead determine whether a
particular policy is warranted. We end up examining similar issues. However,
conceptually the task entails a different approach because we begin with a
policy rather than a problem.

At first blush, this approach may seem odd. Presumably, legislators
implemented the policy because of a real need. As the previous discus-
sion highlights, however, and as the following case studies will reinforce,
that assumption frequently may be incorrect. Regardless, investigation of
whether existing or proposed policies are needed can provide critical infor-
mation for determining whether a shift in direction is indicated. On the basis
of the findings from a needs evaluation, we might determine, for example,
that the policy should be terminated, that it should be modified, or that a
different policy might better address the existing problem.

THREE QUESTIONS WHEN EXAMINING THE NEED FOR AN EXISTING OR

PROPOSED POLICY. A policy-focused needs evaluation entails answering
three related questions. First, are existing efforts insufficient to address
some social problem? It defies common sense to create a policy merely for
the sake of doing so, especially if other policies that address the problem
exist. Consider the typical situation in criminal justice – most states have
numerous policies in place to sanction offenders and to provide diversion
and treatment services. A lawmaker may propose new legislation to toughen
penalties for some class of offender. But any such effort assumes that the
range of sanctions currently in place somehow are insufficient and that
diversion or treatment services somehow do not work or are inadequate.
Why else go to the trouble of creating more complexity in a system already
overburdened by complexity? In short, to determine whether a particular
policy is needed requires, in part, first determining how, if at all, existing
policies fall short.

NEEDS EVALUATIONS 61

Second, are existing efforts to address some social problem not only insuf-
ficient but also not amenable to correction? Here, contemplate preparation
for an exam, because most of us have had to take a test at some time or
another. Perhaps someone comes along and tells you to adopt an entirely
new approach to preparing for your exam. That’s well and fine, but really, do
you need to undertake a whole new strategy? The person responds by observ-
ing that you flunked in all of your previous efforts and so clearly something
new is warranted. The logic seems self-evident. It is, nonetheless, specious.
It may be that the strategy you used was entirely appropriate and that you
simply failed to implement it well. That does not mean that you should
continue with it as a strategy. For example, if the strategy was to study two
hours every night, and if that strategy was unrealistic given competing obli-
gations, another strategy might well be indicated. However, it may be that
the poor implementation can be easily remedied. Perhaps, for example, you
simply need to shift the time of day when you study, or spread the study
time out over different periods of the day. In this situation, you might simply
want to revise your current approach to exam preparation, not move on to a
new one.

Third, in comparison to existing efforts to respond to some assumed or
documented social problem, is a proposed or newly implemented policy a
needed substitute or supplement? Here, assume that existing policies have
been found to be insufficient and that their inadequacy is only partially
remediable. The question remains – is a new policy in any obvious way
more responsive in addressing a social problem? Answering that question is
important because it makes little sense to invest scarce resources in policies
that do not clearly address a problem in a more direct, effective, and cost-
efficient manner than some other approach or set of approaches.

Doesn’t this last question tip us into the land of conducting theory eval-
uations of different policies? In part, yes. As will be discussed in the next
chapter, theory evaluations examine a policy’s causal logic to determine if it
rests on solid theoretical, logical, and empirical foundations. Theory evalu-
ations thus necessarily entail investigation of how well a policy may remedy
some social problem.

That observation underscores a critical point – needs evaluations and
theory evaluations are inextricably intertwined. A well-conducted evaluation
of the need for a particular policy typically will require delving into the
theoretical logic and evidence for the policy. Similarly, a well-conducted
theory evaluation typically will require delving into the evidence concern-
ing the prevalence, distribution, and nature of a social problem. Consider, for
example, that a well-designed policy may exist but the social problem it aims

62 AMERICAN CRIMINAL JUSTICE POLICY

Too Much Crime?

Increase in
Crime?

Prison
Overcrowding?

Caused by A

Caused by B

Caused by C

Caused by D

Caused by E

Caused by F

Solution A1

Solution A2

Solution B1

Solution B2

Solution C1

Solution C2

Solution D1

Solution D2

Solution E1

Solution E2

Solution F1

Solution F2

Social Problem
That Needs Addressing

Cause of
Social Problem

Solution to
Social Problem

Figure 4.1. Defining need.

to address does not. In such a situation, we would not want to implement
the policy because no problem exists.

EXAMINING WHETHER A PROBLEM EXISTS VERSUS THE NEED FOR A

SPECIFIC POLICY. To clarify the distinction between a needs evaluation
aimed at determining whether a social problem exists and one aimed at
assessing whether a particular policy is warranted, let us turn to Figure 4.1.
If we begin on the left side, we can see that our first task involves defin-
ing the social problem that we wish to address. Is it, for example, that we
feel that too much crime exists? Or is it that there appears to be a sudden
surge in crime? The two problems may be related but may not be, and so
warrant separate treatment. For example, a community with low levels of
crime may experience a doubling of crime from one year to the next and, as a

NEEDS EVALUATIONS 63

result, decide to intensively implement policies aimed at addressing this new
problem, one that is “new” not in the sense of crime existing but in the sense
of its rising dramatically. In another community with higher levels of crime,
there may have been little change in the crime rate from one year to the
next, and so perhaps no new policies are implemented. Instead, attention
may be centered on improving existing efforts to combat crime. Still another
social problem that may constitute a concern is prison overcrowding, which
may or may not be linked to levels or changes in crime. It, too, may warrant
its own evaluation.

As part of a needs evaluation, we would want to clarify which of these
problems merits the most attention. In that effort, we would want, as dis-
cussed earlier, to describe the nature and extent of the problem, how it has
changed, and its distribution among different populations and places. We
also would want to identify what caused it because doing so can help us
define the problem more precisely and lead us to a targeted policy response.
The second column depicts that focus and illustrates that multiple fac-
tors may cause a given problem. High rates of crime, for example, may
be caused by poverty, unemployment, socially disorganized communities,
weak social ties among neighbors, and so on.6 Rapid increases in crime
may be caused by increases in such factors but also may result from rapidly
expanding drug markets, changes in the demographic characteristics in a
given area, and many other such changes.7 Prison overcrowding may result
from increased crime rates and, in turn, incarceration of ever-greater num-
bers of offenders. It also may result from shifts in public opinion, concern
about specific types of crimes, changes in the prevailing political philosophy
in states or nationally, and, not least, mandating that inmates serve their
entire terms of incarceration, which creates less capacity for admitting new
offenders.8

Observe that specific causes lead to consideration of different solutions,
as indicated by the third column in Figure 4.1. For example, if we find
that high levels of poverty cause crime, then policies that alleviate poverty
would be a logical focus. Of course, many different ways exist to reduce
poverty, including various types of social welfare and efforts to bolster the
economy. Which solutions get pursued would reflect many considerations.
Perhaps the prevailing political climate rails against the notion that people
should get “handouts”; in such a context, efforts to improve the economy
presumably would garner greater political support.9 Another cause of high
crime rates may be a lack of social ties. If identified as the most salient or
tractable cause, then policies that promote stronger community ties would
make more sense. The same logic applies regardless of the type of social

64 AMERICAN CRIMINAL JUSTICE POLICY

problem under consideration, whether it is high crime rates, increased crime,
prison overcrowding, or any of myriad possibilities.

Revisiting Figure 4.1, let us now work from right to left. It frequently
seems to be the case that policy makers assume that a particular policy must
be needed. Perhaps they hear about a new policy developed in some other
state. Perhaps they believe that innovations, as a general matter, should be
adopted. Perhaps they think that nothing currently works and so something
is better than nothing. Whatever the reason, policies frequently emerge
without any clear statement or documentation of need. Instead, the need
is assumed to exist. That assumption can lead to significant mistakes in the
allocation of resources.

To illustrate, many states have implemented supermax prisons without
clear statements of need. What problems do such prisons address? System-
wide safety problems, disorder, riots, escapes, and recidivism are some pos-
sibilities.10 Most state correctional systems experience these problems. So,
the main one of interest would seem to be dramatic increases in these prob-
lems.

Here, observe that a basic challenge emerges as we work from the solu-
tion, supermax prisons, to the problem – namely, numerous problems exist
for which such prisons could be viewed as a solution. In addition, in each
instance we would want to ask what really is the scope of the problem,
how has it changed; what is its distribution; and, critically, what caused it.
Consider systemwide order and safety. Perhaps there has been a dramatic
increase in these problems in some states, but perhaps the cause stems from
prison management, not the behavior of a few “bad apples.” In this case,
supermax prisons do not seem to be a logical solution to the problem given
that they primarily focus on incapacitating a relatively small number of the
so-called bad apples, that is, those inmates thought to cause many of the
problems in prison systems.11

Other questions quickly emerge as we try to work backward from the
solution to the problem and the causes of it. Has disorder and safety actually
increased in a given system? Are the causes of it primarily or only caused by
the behavior of a few inmates? Is the problem instead increased escapes or
higher rates of recidivism or even crime rates? If so, the question again is
what has caused these problems and whether supermax prisons constitute
an obviously superior policy choice over other possibilities.

Similar questions can be asked of virtually any policy. To wit, what exactly
is the need that the policy addresses, and is it, upon closer inspection, the
most logical solution? Consider another example: drug courts. This spe-
cialized, intensive form of processing, supervising, and treating offenders

NEEDS EVALUATIONS 65

has expanded rapidly since the early 1990s. What problem does it actu-
ally address, though? The use of illegal drugs? Production and distribution
of them? Drug-related offending among individuals caught by the police?
Community-level drug crime? All of these possibilities constitute good can-
didates, but they each raise different questions. Intensive supervision and
treatment would seem to hold the promise of reducing recidivism among
drug offenders as a general matter, but would they do much to reduce it
among those offenders whose offending careers predated any involvement
in drugs? Would it do much to deter illegal drug production and distribu-
tion? More to the point, what causes each of these types of problems and do
these causes naturally lead to a focus on drug courts? For example, if illegal
drug markets prevail during downturns in the economy, it would seem that
economic interventions would have a greater chance at reducing overall lev-
els of drug production and distribution and, by extension, drug use and any
drug-induced criminal behavior.

A careful investigation of the need for a drug court would not stop there,
but also would situate the problem in relative terms. For the sake of argu-
ment, assume that a city has many thousands of people who use illegal
drugs and commit crime to support their drug habit or because of it. From
a criminal justice perspective, the fundamental concern lies with criminal
behavior. So, our primary focus then would be on documenting how much
drug-related crime contributes to overall levels or rates of crime in the city
and to changes in such crime. Doing so would enable us to determine how
much drug-related crime contributes to overall crime problems and in turn
whether it merits special focus relative to a focus on other factors that may
contribute to crime.

This example highlights an important axiom – any definition of need
necessarily implies some basis of comparison. Pointing to particular drug-
related crimes or to estimates about the large percentage of prisoners with
drug problems does little to advance discussions of need precisely because
of the lack of comparative context. However, pointing to such facts and then
highlighting that no other “competitors” exist – like mental health problems,
poor education, poverty, chronic unemployment and homelessness – would
do so. One then could argue that drug-related crime clearly predominates
over other crime and so merits special attention. However, if one finds that
crime seems more associated with other problems, such as mental illness or
poverty, then a stronger need would seem to exist for solutions focused on
them, not drugs.

It bears emphasizing that the many and sundry criminal justice policies
that exist today or are being contemplated, including supermax prisons and

66 AMERICAN CRIMINAL JUSTICE POLICY

drug courts, may well be needed. Nevertheless, asserting that fact and show-
ing it to be true are two different things. Ultimately, a science – or evidence-
based criminal justice system – would err on the side of demonstrating need,
not asserting it.

How to Conduct a Needs Evaluation

Needs evaluations – whether of social problems for which we think a policy
response might be needed or of the problems we think an existing policy
is intended to address – can be viewed as progressing through two stages.
The first is a conceptual endeavor, one aimed at describing as specifically
as possible the social problem thought to exist. The second stage involves
the development of quantitative estimates of the problem. Following this
sequence helps to ensure that we conduct the type of research relevant to
identifying and addressing a problem. For example, if we determine first
that supermax housing serves primarily to reduce disorder and violence
throughout prison systems, we know then that we need information about
the nature and levels of disorder and violence.

The sequence can be reversed, but significant risks arise in doing so. To
illustrate, consider reading a report that shows that a state experienced no
rioting after it constructed a supermax facility. We see this fact, which in turn
leads us to think that the supermax prevented future riots and, more than
that, to assume that the facility was needed. In reality, however, the relevant
facts that we want for assessing need in this case would likely be trends in the
amount and distribution of disorderly and violent acts throughout the prison
system. If we collect information necessary to produce such facts, we might
well find that disorder and violence were declining well before introduction
of the supermax, suggesting that no need for the supermax existed. We also
might find that the factors contributing to the riot had little to do with the
types of inmates housed in the supermax facility, suggesting that the housing
was not actually the type of policy response needed for responding to this
particular problem (i.e., riots).12

No single approach exists for conceptualizing a problem. The main ingre-
dients include efforts to state explicitly the nature of the problem, to com-
pare that statement with accounts in the scholarly or policy literature of
what seem to be similar problems, and to work toward a precise description
that can guide research aimed at quantifying the problem. In such an under-
taking, interviews with experts in the field and focus groups with different
stakeholder groups can be invaluable.

There also is no single approach to quantifying a problem. Many needs
evaluations begin with qualitative methodologies, such as interviews, focus

NEEDS EVALUATIONS 67

groups, and ethnographies, because these can help us to identify the types of
information we should collect.13 A well-done qualitative study can be used
to aid in designing survey questionnaires or in making tailored requests
for agency or administrative records. These in turn provide the basis for
establishing the prevalence – measured using counts, percentages, rates,
or the like – of the problem. Numerous research methods and evaluation
texts exist that describe these methodologies and how to undertake them.14

Once, again, however, the conceptualization of the problem ideally should
dictate the methods used.

The purpose of needs evaluations is to inform policy decisions. For that
reason, they should clearly identify the target population that contributes to
a problem or that can help alleviate it. For example, if we determine that
violence has increased dramatically in a prison system, we will want to aid
those charged with developing a policy response. To this end, we should
go beyond showing that violence has increased and identify who should be
the target of a response. The targets typically will be individuals or groups
who contribute to the problem, which, in this instance, likely would include
inmates who incite others to commit violence or those who actually engage
in it.

Finally, in an effort to inform the development of an effective policy res-
ponse, a needs evaluation also will want to describe the current efforts to
address the problem and the causes of it. For example, have wardens imple-
mented any initiatives to reduce violence? Have there been difficulties in
implementing them? If instigators contribute to most of the prison violence,
can they easily be identified? Would legal barriers inhibit a prison system’s
ability to use isolation as a means of controlling these inmates?

Why Are Needs Evaluations Important?

Needs evaluations serve a number of critical functions. First, and perhaps
most important, they help us to identify whether a problem exists, and,
if the evaluation is done well, the scope and nature of the problem. The
point cannot be emphasized enough – expending resources on nonexistent
problems not only is wasteful but also draws attention away from other, real
problems.

Second, a needs evaluation can provide guidance about whether a policy
should be retained. All too frequently, policies emerge in the absence of
any systematic needs evaluation. In these cases, a needs evaluation can
provide guidance on whether the problem to which the policy is addressed
actually exists. Should there be no problem, then perhaps the policy should
be terminated, especially if it entails considerable costs.

68 AMERICAN CRIMINAL JUSTICE POLICY

Third, well-done evaluations can also provide guidance about the type of
policy response that should be considered. For example, they can show not
only that homicides have increased but also that, say, gang-related homicides
have dramatically escalated in certain areas of a city. Policy makers thus have
information that can guide them, not to respond to violent crime “in general,”
but rather to focus on gang-related homicides in certain areas, and thus to
develop policies that target gangs in them.

Fourth, needs evaluations can aid in efforts to prioritize the problems
that merit attention. Typically, communities or criminal justice systems face
a wide range of problems and so they must make the difficult decision to
emphasize some over others. For example, a prosecutor’s office may face
what they perceive to be a large influx of juvenile delinquency cases and
domestic violence cases, enough so that discussions emerge about creating
specialized delinquency or domestic violence units. Ideally, such decisions
would be informed by, among other things, information about the number of
delinquency and domestic violence cases, changes in them relative to each
other and to other types of cases, and the time and money required to handle
each cases. A needs evaluation might reveal that in fact domestic violence
cases have increased much more dramatically than delinquency cases and
require far more resources. That information in turn can be used to guide
decision making about specialized units.

Fifth, needs evaluations can highlight research gaps that ideally would
be addressed before implementing a policy. A particular state may feel that
prison overcrowding exists and must be addressed by building new prisons.
However, before undertaking such a costly endeavor, it would help to know
what caused the overcrowding. Is it due, for example, to larger numbers
of admissions from the courts, from admissions in one county in the state,
or perhaps to a new law mandating that any newly admitted inmates must
serve a minimum of, say, 80 percent of their sentence in a context where
traditionally only 40 percent of inmate sentences get served? Each of these
sources suggests different policy responses. A needs evaluation could help to
identify the source of the overcrowding, and if it could not, then it could draw
attention to the fact that the new policy – building more prisons – proceeds
from an unverified premise that the problem stems from a constant source of
new pressure for beds rather than, say, a temporary spike in new admissions.
Policy makers may still feel compelled to support prison expansion, but they
would be able to make that decision in a more informed manner.

Sixth, needs evaluations can highlight the dimensions that should be used
to assess a policy’s effectiveness.15 Consider community policing, which aims
to reduce crime, an admittedly critical law enforcement priority. At least
in part, however, community policing also aims to foster greater cohesion

NEEDS EVALUATIONS 69

among community residents as well as to increase trust in and satisfaction
with the police.16 Such changes might ultimately reduce crime, but they also
constitute important goals in and of themselves.17 A needs evaluation would
likely identify this fact and thus underscore the importance of including
changes in community cohesion and trust in the police as important markers
of success.

Finally, a more general benefit of needs evaluations is that they can help
clarify policy discussions and debates. Policy makers and criminal justice
officials face a strong mandate to take action, and so they quite reasonably
tackle the problems that seem most prominent. However, policy makers and
officials come and go while criminal justice systems remain. That creates a
problematic situation. One group feels compelled to do something, yet tak-
ing action with incomplete information about the range and depth of public
safety and criminal justice system problems creates potentially more prob-
lems. In such a context, needs evaluations, especially when aimed broadly
at illuminating the spectrum of problems in criminal justice, can contribute
to more informed discussions and debates about the direction that policy
should go.

Case Studies: Mass Incarceration and

Sex Crime Laws

To illustrate several of the aforementioned points, we turn now to two case
studies, one focused on mass incarceration and the other on sex crime laws.
In each instance, the discussion centers on a description of the policy and
its goals and how need might be or is defined. Following the case studies,
we then turn to the question of whether these prominent criminal justice
policies and others like them appear to be needed.

Mass Incarceration

THE POLICY. Correctional populations in the United States have bur-
geoned, expanding almost four times in size from 1980 to 2008. If incarcer-
ation rates are the measure, the United States is the “most punitive country
in the world.”18 In 1980, there were 319,598 individuals in prison; by 2008
that number had risen to 1,518,559. The latter figure does not count the
785,556 individuals who were in jail. Combining these two figures, there
were more than 2.3 million individuals in jail or prison in 2008.19

People who hear about such increases frequently comment, “Yes, but
over that twenty-eight year period of time the U.S. population grew, so
doesn’t the growth simply reflect that change?” The answer is no. Figure
4.2 shows the changes in incarceration rates, which adjust for population

70 AMERICAN CRIMINAL JUSTICE POLICY

0

100

200

300

400

500

600

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

Figure 4.2. U.S. incarceration rates, 1980–2008. The rates refer to adult prisoners held for one
year or longer per 100,000 residents. 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/incrttab.htm (accessed December 15, 2009).

growth. In 1980, the incarceration rate was 139 per 100,000 residents. By
2008, the incarceration rate was 504 per 100,000 residents, an increase of
263 percent. In short, the use of incarceration relative to the size of the U.S.
population increased dramatically in recent decades. (Perhaps the growth
reflects increases in crime? As we will see next, that does not appear to be
the case.)

Incarceration rates increased in all but two years (2000–2001 and 2007–
2008) in that time span, but the most pronounced growth occurred from
1980 to 2000. During this period, the incarceration rate grew an average of
6.4 percent annually, with increases of 10 percent or more in some years.
Many scholars refer to the historically unprecedented growth in incarcera-
tion, especially during the 1980s and 1990s, as mass incarceration.

This growth contributed to one of the critical challenges confronting
America today – the reentry of more than 735,000 inmates annually back
into families and communities nationwide. It also led to a large body of
scholarship aimed at understanding the causes and consequences of this
sea change in incarceration policy.20 Much of this work sidesteps a basic
question, however: When exactly does society need more incarceration?

NEEDS EVALUATIONS 71

TABLE 4.2. U.S. incarceration rates, nationally, regionally, and
by state, 2008a

U.S. Total 504
Federal 60
State 445

Northeast 306
Connecticut 407
Maine 151
Massachusetts 218
New Hampshire 220
New Jersey 298
New York 307
Pennsylvania 393
Rhode Island 240
Vermont 260

South 552
Alabama 634
Arkansas 511
Delaware 463
Florida 557
Georgia 540
Kentucky 492
Louisiana 853
Maryland 403
Mississippi 735
North Carolina 368
Oklahoma 661
South Carolina 519
Tennessee 436
Texas 639

Virginia 489
West Virginia 331

Midwest 392
Illinois 351
Indiana 442
Iowa 291
Kansas 303
Michigan 488
Minnesota 179
Missouri 509
Nebraska 247
North Dakota 225
Ohio 449
South Dakota 412
Wisconsin 374

West 436
Alaska 430
Arizona 567
California 467
Colorado 467
Hawaii 332
Idaho 474
Montana 368
Nevada 486
New Mexico 316
Oregon 371
Utah 232
Washington 272
Wyoming 387

a The rates refer to adult prisoners held for one year or more per 100,000
residents.

Source: Sabol, William J., Heather C. West, and Matthew Cooper. 2009. Prison
Inmates, 2008. Washington, D.C.: Bureau of Justice Statistics, p. 30.

Before addressing that question, we should be clear that incarceration is a
policy – it represents a decision to use long-term confinement of individuals
to achieve some goal, such as retribution, rehabilitation, or public safety.21

Not only that, but it constitutes a policy that regions of the country and
states use in varying degrees. Examine, for example, Table 4.2. We can see

72 AMERICAN CRIMINAL JUSTICE POLICY

that the national incarceration rate in 2008 was 504 per 100,000 residents.
That combines the state (445) and federal (60) incarceration rates. Region-
ally, however, we can see that the use of incarceration varies dramatically.
In the Northeast, the average rate of incarceration in 2008 was 306, in the
Midwest it was 392, in the West it was 436, and in the South it was 552.
Focusing on the extremes, we can see that southern states on average incar-
cerate almost 250 more people per 100,000 residents as compared with
northeastern states. Put differently, the policy in the South, relative to the
rest of the country, is to incarcerate substantially more people.

Equally dramatic variation occurs at the state level within each of the
regions. In the Northeast, Maine has an incarceration rate of 151, whereas
Connecticut has an incarceration rate of 407. So, in that one region, the
variation reflects that across all of the regions. The range in incarceration
rates is somewhat greater in the western states, running from a low of 232
in Utah to a high of 567 in Arizona. The Midwestern states exhibit even
greater variation – the low and high incarceration rates are in Minnesota
(179) and Missouri (509), respectively. Witness the greater variability in the
South – whereas West Virginia has an incarceration rate of 331, Louisiana
has an incarceration rate of 853. In short, considerable variation exists both
across regions and within regions in the extent to which incarceration as a
policy is used.

Finally, when we turn away from regional differences and focus exclusively
on states, we can see that the variation in the use of incarceration as a policy
is even greater among states than among or within regions.22 In Figure 4.3,
for example, observe that the use of incarceration varies from a low of 151 in
Maine to a high of 853 in Louisiana. The difference in rates of incarceration
is 702. Viewed somewhat differently, Louisiana has an incarceration rate 5.7
times greater than that of Maine’s. Of course, these two states constitute
outliers – that is, the extreme ends of a range of values. Even so, we can see
that the incarceration rates among the remaining states vary greatly. Such
variation has led one scholar to remark, “Over the past twenty years, the
fifty American states have engaged in one of the great policy experiments of
modern times.”23

IDENTIFYING THE NEED FOR INCARCERATION. Returning to the question
at hand – when do we need more incarceration? Here, let us first stop to
consider the variation in the use of incarceration across states. To put such
variation in perspective, imagine an antibiotic known to cure some disease.
Then, imagine that the use of the antibiotic varies along the lines of the

NEEDS EVALUATIONS 73

0 100 200 300 400 500 600 700 800 900

Maine
Minnesota

Massachusetts
New Hampshire

North Dakota
Utah

Rhode Island
Nebraska
Vermont

Washington
Iowa

New Jersey
Kansas

New York
New Mexico

West Virginia
Hawaii
Illinois

North Carolina
Montana

Oregon
Wisconsin
Wyoming

Pennsylvania
Maryland

Connecticut
South Dakota

Alaska
Tennessee

Indiana
Ohio

Delaware
Colorado

California
Idaho

Nevada
Michigan

Virginia
Kentucky
Missouri
Arkansas

South Carolina
Georgia
Florida

Arizona
Alabama

Texas
Oklahoma

Mississippi
Louisiana

Incarceration Rates

Figure 4.3. U.S. incarceration rates by state, 2008. Source: Sabol, William J., Heather C. West,
and Matthew Cooper. 2009. Prison Inmates, 2008. Washington, D.C.: Bureau of Justice Statistics,
p. 30.

74 AMERICAN CRIMINAL JUSTICE POLICY

variation in incarceration rates. We would of course anticipate, or perhaps
hope, that there would be no variation because we know the antibiotic works
and so it would seem to be a treatment that should be used whenever and
wherever it was needed. Of course, it might be that the rate of the dis-
ease varies across places, and so that might create variation in treatment.
However, after making an adjustment for this possibility, if we still saw dra-
matic differences in the treatment of the disease, we would be concerned.24

Consider the implications of this example for decisions about incarcera-
tion. If increasing the use of prisons “works” (by some criterion and as estab-
lished by research), then there would seem to be little reason for variation in
the use of it, save for any differences due to variation in the amount of the
problem addressed by incarceration. From this perspective, the regional and
state differences in the use of incarceration stand out as cause for concern.
Why? Because, by and large, it cannot be accounted for by variation in the
problem that many of us would think incarceration addresses – crime. Stud-
ies show that increases in crime either exert no effect or a modest effect on
increases in incarceration rates. David Greenberg and Valerie West (2001)
found, for example, that increases in violent crime and property crime were
unassociated with increases in incarceration rates among states for the years
1971–1991.25 Their study merits special attention because, in contrast with
most prior research, they examined changes in crime to determine if they led
to changes in incarceration. Put differently, their study provided one of the
few direct assessments of whether crime is causally related to incarceration,
and they found that appears not to be so related.

Let us view the matter from a different angle. In the medical example,
we assumed that the medical establishment agreed that antibiotics effec-
tively treat a particular disease. If, however, we assume that the medical
establishment disagreed about the treatment’s effectiveness, we would likely
anticipate that some places would put greater faith in the treatment than
others. In turn, we would anticipate variation in the use of antibiotics to
treat the disease. Some doctors would think antibiotics work and so would
use them, and some would be skeptical and thus refrain from using them.

Seen through this lens, variation among states in the use of incarceration
perhaps should not be surprising, given that little consensus exists in the
policy-maker or scholarly community about the effectiveness of incarcera-
tion. That may seem odd. How could incarceration not be an effective way
to “treat” the problem of crime? Society takes criminals out of commission –
we incapacitate them – and sends a message to would-be offenders that
crime will get you in trouble. So clearly, in turn, crime should go down.
In reality, though, studies find that incarceration contributes, at best, to

NEEDS EVALUATIONS 75

modest reductions in crime, and the effect varies across types of crime.
As Marie Gottschalk has written, “There is some relationship between the
crime rate and the incarceration rate, but it is slight. Analysts using a variety
of methodologies have found that the deterrent and incapacitation effects
of incarceration in bringing down the crime rate are small, and that the
offenses avoided through the greater use of prisons tend to be nonviolent
rather than violent crimes.”26

One of the more generous estimates of a crime-reducing effect of incar-
ceration comes from work by William Spelman, whose analyses suggest that
increased incarceration in the 1980s and 1990s contributed to roughly one-
fourth of the drop in crime in the late 1990s.27 That effect can be seen in two
ways. On the one hand, the reduction, if real, translates into a significant
improvement in public safety. Many state and local policy makers would be
glad to take credit for a 25 percent reduction in crime that resulted from poli-
cies, like incarceration, that they may have initiated.28 On the other hand,
states must invest large amounts of resources to achieve that reduction.
In the time period Spelman used in his analyses (1971–1997), the prison
population grew fourfold, leading him to comment: “Even if imprisonment
were an incredibly inefficient means of reducing crime – and there are strong
arguments that it is exactly that – it could hardly have helped but have a
substantial effect on the crime rate, given the enormous scale of the differ-
ence.”29 Viewed somewhat differently, investment in strategies other than
prison might have produced comparable effects at less cost.

In short, if we focus on crime as a social problem that needs to be
addressed, we find some evidence to suggest that this problem leads states
to invest in prisons.30 At the same time, Gottschalk, Spelman, and others
have indicated that incarceration rates stem from many other causes. For
example, Greenberg and West’s study showed that states with more con-
servatives experienced higher rates of incarceration as well as increases in
incarceration; states with greater increases in the religious fundamentalist
population experienced greater increases in incarceration; and states with
greater levels of and increases in unemployment experienced greater incar-
ceration growth.31 Such findings suggest that incarceration may address
some other needs besides crime.

CRITERIA FOR DETERMINING THE NEED FOR INCARCERATION. The pre-
viously mentioned observations set the stage for more formally taking stock
of some of the considerations that bear on evaluating the need for incarcer-
ation. First, a needs evaluation would attempt to make clear exactly what
social problem incarceration addresses. The admittedly simple question in

76 AMERICAN CRIMINAL JUSTICE POLICY

fact defies simple analysis. For example, incarceration can be viewed as help-
ing to address two very different problems – society’s desire for retribution
when its laws are violated and society’s desire for public safety. Many policy
discussions about incarceration center on the idea of public safety, but ret-
ribution stands as a critical part of any punishment policy.32 Incarceration
also can be viewed as a response to prison overcrowding. Ultimately, an
effective policy response involves first targeting a clearly stated problem or
set of problems. Focusing on only one aspect of the problem (e.g., a need to
reduce crime) while ignoring another (e.g., societal demands for retribution)
places any subsequent policy efforts on weak footing.

Second, a needs evaluation would attempt to quantify the scope of the
problem. How much crime exists, how much has it changed, where is it occurring,
and what causes it? Few states actually quantify the true amount of crime, but
instead use proxy measures, such as arrests, even though the latter typically
do not provide an accurate basis for estimating crime or changes in crime.
For example, if states invest more heavily in law enforcement, more arrests
will occur even if true crime declines.

The problem lies deeper than this measurement issue, however. The more
fundamental question is, How much crime is bad enough to warrant a
response? More specifically, how high do crime rates have to rise before
society believes a response of some type, such as increased incarceration,
should be imposed? Similarly, how much do crime rates have to increase
before society deems that a response to be warranted? In contemplating
these questions, we confront a stunning fact – states typically provide no
explicit standards regarding the levels or changes in crime that warrant a
response. The variation among states in their use of incarceration indicates,
however, that they differ in the implicit, unarticulated criteria that they
follow.

Much the same can be said for states’ assessments of the public’s need for
retribution. Not all crime is the same, and so, by extension, demand for ret-
ribution may vary depending on the types of crimes that occur in a state. In
addition, retributive feelings may vary along social and demographic lines,33

and so an accurate assessment of the state-level need for retribution would
require taking such differences into account. Once, again, few states ever
systematically take stock of public views toward crime, much less how these
views change over time or what the specific sanctions are that the public
thinks constitutes sufficiently retributive responses to particular crimes. The
issue serves as pause for concern, if only because public opinion research
shows that policy makers consistently overestimate the punitiveness of the

NEEDS EVALUATIONS 77

public and underestimate their desire for and willingness to support rehabil-
itative measures.34

Overcrowding, too, entails consideration of similar questions. For exam-
ple, how much overcrowding must occur before more prisons should be
built? In answering that question, we would want to identify how much of a
prison sentence inmates should serve. Inmate overcrowding can be relieved,
for example, simply by having prisoners serve less time. The public tends to
be shocked when they learn that inmates rarely serve their entire sentences.
Nevertheless, it is true. For example, a Bureau of Justice Statistics study
found that “violent offenders released from State prisons in 1992 served 48
percent of the sentence that they had received”35 and that, among released
prisoners, the “average sentence for a violent crime was 8 years, and the
average time served was about 3.5 years, or just under half of their total
maximum sentence.”36 The study also found that states varied greatly in
the percent of sentence terms served by violent offenders, from a low of
21 percent (Ohio) to a high of 89 percent (Delaware).37

In the 1980s and 1990s, such differences prompted many states to enact
so-called truth-in-sentencing laws aimed at making time served more closely
mirror the length of sentences handed down at sentencing. Observe, how-
ever, that such efforts sidestep a basic question – how long should sentences
for specific offenses be? The implications of not answering that question can
be severe. To illustrate, a state might be incarcerating offenders for terms
that are twice what the public feels is appropriate and yet be releasing pris-
oners once they have completed half of their assigned terms. In that case,
the end result mirrors what the public wants in terms of the total length
of sentence actually served. Even so, policy makers might focus instead on
the fact that only half of the total assigned sentence length was served, view
this as a problem, and then enact legislation aimed at requiring prisoners to
serve their full terms. The end result? Prisoners serve lengthier prison terms
than what the public wants.

View the matter from a slightly different perspective. Almost all prison
bed-space forecasts make a highly problematic assumption – namely, that
current bed-space use is the “right” amount for the level of crime faced
in a particular state or jurisdiction.38 That is, policy makers frequently
assume that current bed-space usage constitutes the correct, or nearly cor-
rect, response for achieving some amount of crime reduction. Empirically
based models then rely on current bed space use, along with estimates of
future crime rates, to predict the need for future bed space. Observe that
“need” never actually gets defined.

78 AMERICAN CRIMINAL JUSTICE POLICY

To illustrate, imagine a state where a thousand assaults occurred. Assume
that all could be prosecuted and would likely result in conviction, but that
only three hundred result in an arrest. Of course, from a purely legalistic
perspective, virtually all of the assaults warrant punishment. Only three
hundred are arrested, however; so, from the outset, one faces the problem
that 700 of the assaulters (assuming no repeat offenders) will go unpunished.
In virtually every state, crimes such as the 700 assaults occur that never come
to the attention of law enforcement. The question, then, is when is that
amount – what might be termed the dark figure of unpunished crime – too
much? Jurisdictions do not typically answer the question, thus the need for
legal punishment remains unknown. Even for the 300 arrests, states rarely
define need, save that the most “severe,” however operationalized, typically
should be prioritized by prosecutors and the courts. Not all 300 will be given
equal attention; indeed, some will be dismissed.

If one turns from purely legalistic to crime prevention definitions of need,
questions arise about the amount of punishment needed to achieve some
specific quantity of reduced crime in society at large. How many of the 300
arrested assaulters must be punished? What is, on the one hand, the amount
of crime that “needs” to be reduced and, on the other hand, a reasonable
amount of crime reduction to expect? Can a substantial reduction be had
by punishing only a small percentage of the offenders, perhaps, for example,
the most serious assaulters? Or is it necessary to sanction all of them? If so,
by how much, and what about the 700 assaulters who went untouched? Not
least, how much of a crime prevention effect will arise through specific and
general deterrence mechanisms? That distinction is important because for
some crimes, such as homicide, one can anticipate that certain sanctions,
like the death penalty, will produce little by way of a general deterrent
effect.39 For other crimes, such as assault, it may be that a combination of
specific and general deterrence collectively gives rise to an appreciable crime
reduction.40

Against this backdrop, consider again the question of bed-space “need.”
How many prison beds are “needed” when there are 1,000 assaults? Cer-
tainly, one can anticipate that not all assaulters should be incarcerated. The
question is, Which ones should be? Prosecutors might focus on the most seri-
ous cases, those in which substantial injury was involved, but that suggests
a criterion of need – assaults in which only major injuries were sustained.
Legislators have enacted laws that endorse greater punishment for weapons-
related offenses, so perhaps “need” could be defined as assaults in which a
weapon was used. In both cases, to establish the need for bed space, one
first must identify how many assaults occurred that met the definition of

NEEDS EVALUATIONS 79

“serious.” Few jurisdictions or states in fact provide such operational defi-
nitions of need, or base their implicit or explicit definitions on the amount
of crime reduction that can be expected from using specific quantities of
incarceration for particular types of offenders.

Third, a needs evaluation would attempt to characterize the nature of
the problem, including a description of its causes. If overall crime rates
have dramatically risen but violent crime has declined, then, by extension,
there would seem to be less of the kind of problem (violent crime) typically
viewed as candidates for prison. Perhaps the public’s preferences for puni-
tiveness have changed. That would call attention to the importance of con-
sidering whether incarceration policies should be adjusted to reflect public
sentiment. A careful analysis of overcrowding might lead to similar consid-
erations. Overcrowding can stem from other factors, prison admissions in
particular. Perhaps, for example, certain counties are quicker than others to
incarcerate people for committing any kind of crime. Such counties could
disproportionately contribute to state prison populations. At the same time,
they might incarcerate the types of offenders that other counties and per-
haps citizens at large view as better punished using other sanctions. Here, a
needs evaluation, in the course of attempting to characterize the problem at
hand and its causes, might find that the problem lies not with overcrowding
but with overly aggressive use of incarceration by some counties. Of course,
the opposite finding might emerge as well – to wit, other counties may be
too lenient and not be sending appropriate cases for incarceration in state
prisons. In this case, overcrowding estimates would actually be understating
the true scope of the problem.

Fourth, a needs evaluation would attempt to ascertain not only the scope
and nature of a problem as well as its causes but also whether a need exists
for a proposed solution, such as incarceration. As the preceding example
indicates, if policy makers want to address the problem of overcrowding,
it may well be that increased incarceration constitutes the most appropri-
ate, effective, and efficient response. An analysis might show that coun-
ties throughout a state are failing to use incarceration in appropriate cases
because of concerns that the state will simply return new admissions back
as quickly as possible. In this case, increased prison capacity may be war-
ranted. However, in other cases, increased capacity may be unnecessary,
ineffective, and inefficient. For example, if a few “rogue” counties overin-
carcerate, the appropriate response might be to intercede, not build more
prison capacity.41 Otherwise, a vicious cycle ensues in which increased
capacity creates a greater demand to use incarceration in inappropriate
cases.42

80 AMERICAN CRIMINAL JUSTICE POLICY

In addition to investigating such possibilities, a needs evaluation would
want to investigate whether, among the options available to address a partic-
ular problem, incarceration constitutes the most appropriate and effective
solution or, put differently, whether it constitutes the most needed option
among those available. That is, it would assess the expected relative ben-
efits of incarceration in achieving such goals as public safety, retribution,
and reduced overcrowding relative to the expected benefits associated with
other possible approaches. When making assessments about the expected
benefits of incarceration, particular attention should be given to whether
defensible assumptions have been made about anticipated incapacitation,
specific deterrent, general deterrent, and rehabilitative effects.43

Fifth, a needs evaluation would attempt to assess the relative need for an
increase in a solution, such as incarceration, to some problem (e.g., increased
crime, public desire for retribution, overcrowding). This issue assumes con-
siderable importance when we talk about incarceration because prisons con-
stitute one of the most costly criminal justice investments society makes. Any
balanced set of criminal justice system policies would, in the ideal case, build
on a systematic assessment of needs across the entire system, not just those
of the correctional system.

Sixth, a needs evaluation would attempt to weigh and balance the poten-
tial harms that may emerge from a policy. Potentially, the need to address
some problem may outweigh the costs of some solution. Ideally, though, pol-
icy makers conduct such assessments in advance. Here, again, a discussion
of incarceration highlights the importance of this consideration given that
prisons may cause more harm than good. For example, they may divert or
obligate resources that might be expended on other potentially more effec-
tive alternatives. Incarceration may increase recidivism and worsen reentry
outcomes of released prisoners, disrupt families and communities and, by
extension, create more crime.44 Not least, according to some studies, mass
incarceration may increase poverty.45

These considerations certainly do not exhaust all of the possibilities, but
they illustrate some of those most critical to conducting a needs evaluation.
At the same time, they illustrate a key axiom of needs evaluations – namely,
such evaluations entail considerable conceptual work. Clear conceptualiza-
tion is as critical if not more so than empirical analyses, especially when the
latter occur untethered from an understanding of the complexities relating
to a seemingly straightforward “problem.” Many different and highly sophis-
ticated approaches exist for forecasting the need for prison bed space, but
none of them are a good substitute for clear conceptualization and descrip-
tion of a problem. For example, one commonly used approach to forecasting

NEEDS EVALUATIONS 81

involves the use of complicated time-series analyses, and another approach
involves equally complicated microsimulation efforts aimed at incorporating
information about prison admissions and lengths of stay for different types
of offenses and for different cohorts. However, both approaches depend
heavily on assumptions about current and past practices. They do not typ-
ically address questions about whether such practices are or were suitably
responsive, or under- or overresponsive, to the extent of actual crime or to
public preferences for specific sanctions for particular types of offenses.46

Such issues can more easily be identified by first proceeding with a concep-
tual analysis of the problem under study.

Sex Crime Laws

THE POLICY. Coinciding with the rise of mass incarceration has been a pro-
nounced effort to address sex crimes.47 One recent assessment indicated, for
example, that in 2005 alone, state legislatures enacted more than one hun-
dred sex offender laws.48 No single type of law stands out. Rather, diversity
rather than uniformity characterizes these laws, which include registries that
make public the names and addresses of sex offenders, statutes that limit
where convicted sex offenders can live, and “get tough” sentencing poli-
cies aimed at increasing incarceration of sex offenders.49 Increasingly, too,
though less common, have been state-level efforts to enact laws allowing for
the civil commitment of sex offenders in mental health facilities after their
criminal sentences have expired.50 In a few states, chemical castration has
been promoted as both a sanction and a treatment to address sex offending.
In addition, DNA data banks are beginning to emerge as a way to help law
enforcement identify and arrest suspects.51

Federal policy-making efforts spurred the development of many of the
state-level efforts. For example, under the Jacob Wetterling Crimes against
Children and Sexually Violent Offender Registration Act, enacted in 1994 –
and subsequently amended by Megan’s Law in 1996 – states are required to
“create and maintain a sex offender registration and notification program or
lose 10 percent of the Federal crime funds.”52 This federal effort was sparked
by victimizations of two young children, one of whom was abducted in 1989
when he was eleven years old (Jacob Wetterling) and the other of whom
was abducted in 1994 when she was seven years old (Megan Kanka). Other
prominent cases involving young children include the sexual assault and
murder of twelve-year-old Polly Klass in 1993 and of nine-year-old Jessica
Lunsford in 2005.53 Such cases continue to prompt new federal legislation
targeting sex crimes. For example, in 2006, Congress enacted the Adam

82 AMERICAN CRIMINAL JUSTICE POLICY

Walsh Child Protection and Safety Act, named for a boy who was abducted
and murdered in 1981. This act provided for “increased registration require-
ments for states [and creation of] a federal civil commitment program.”54

The result, according to Lisa Sample and Timothy Bray, is that all states
now have “registry, notification, and DNA laws” that “include persons con-
victed of a violent or nonviolent sex crime against any person.”55 Notably,
these laws encompass a wide range of offenses and so create the poten-
tial to greatly expand the criminal justice system’s caseloads. For example,
nonviolent sex crimes can include “crimes such as possessing, viewing, or
manufacturing child pornography; enticing a child; soliciting a minor; and
other such offenses for which offenders must register upon conviction.”56

Similarly, many states require that commission of any of a diverse array of
crimes result in a person being registered as a sex offender. These crimes
include such acts as “voyeurism, public exposure, adultery, giving obscene
material to a minor, displaying obscene material on a bumper sticker, and
bestiality.”57

IDENTIFYING THE NEED FOR SEX CRIME LAWS. From a needs evaluation
perspective, the basic question is whether a need for these laws exists.
Clearly, sex crimes occur and so constitute a social problem that should
be addressed. Such an assessment is too simplistic, however. Consider, for
example, that sex crimes have always occurred, or have so long as laws have
made such acts criminal. Thus, the relevant question is whether states sud-
denly experienced a marked increase in sex crimes in the 1980s and 1990s
sufficient to justify a sweeping set of new, “get tough” policies. The answer
appears to be no. As one recent review emphasized, “there is little convincing
evidence” that sex offending has increased.58 Indeed, if anything, the avail-
able, if limited, evidence suggests that it remained level or even declined in
recent decades. For example, from the late 1970s through 2008, rape and
sexual assault rates, as measured by the National Crime Victimization Sur-
vey, steadily declined.59 No doubt, the survey underreports the true amount
of such crime, but it likely does so consistently from one year to the next.

Unfortunately, the survey does not compile national data on the wide
range of acts considered to be sex crimes. As a result, no reliable or valid
estimates of sex crime trends exist for the country as a whole or for specific
states. One could speculate that sex crime trends as a whole run counter
to that for rape and sexual assault. However, doing so would require not
just speculation but also careful arguments that convincingly show why sex
crime trends in general would run in the opposite direction of the trend for
rape and sexual assault.

NEEDS EVALUATIONS 83

This situation is especially problematic from a needs evaluation per-
spective, given that the enactment of sex crime laws has proceeded from
an assumption that sex offending has increased. What if, however, it has
decreased? The logical implication would be that a prominent social prob-
lem had declined and so, if anything, merited less attention. Of course, some
might argue that the various laws have produced the declines in sex offend-
ing. However, the downward trend in rape and sexual assault rates began
well before many of these laws emerged and paralleled an overall decline in
violent crime in general, not just sex crime.60 An alternative view might be
that, in the past, the country failed to take adequate steps to punish and
control sex offenders. Thus, even if sex crime rates trended downward of
their own accord, new forms of punishment and new restrictions might be
seen as needed to correct for the too-lax approaches of the past. Here, how-
ever, we would want first to define the appropriate response as a benchmark
for determining whether past practice was sufficient or not.

Many sex crimes unquestionably involve horrific acts, and so for that
reason spark outrage among the public and policy makers. That alone likely
contributes to calls for tough responses to sex crimes. In addition, recidivism
statistics, as presented in media accounts, likely create additional alarm. A
widely cited and credible federal study shows, for example, that sex offenders
released from prison are four times more likely than nonsex offenders to be
rearrested for a sex crime.61 However, such studies frequently gloss over or
miss the fact that sex offenders rarely recidivate and that, as a group, they
contribute less to sex crime nationally than do nonsex offenders.

How can that be? In the federal study, which examined inmates released
from prisons in fifteen states in 1994, there were 9,691 male sex offenders.
Of this group, 517 were rearrested for a sex crime within three years of
release from prison. Put differently, 5.3 percent of the released male sex
offenders were rearrested for a sex crime within three years of release. With-
out question, some sex crimes did not result in an arrest, and the arrests that
did occur included many serious offenses. At the same time, the low base
rate of recidivism (5.3 percent) gives pause for thought.

Of course, nonsex offenders also commit sex crimes after release. The
federal study found that of the 262,420 released nonsex offenders, 3,328,
or 1.3 percent, were rearrested for a sex crime. This lower rate of sex-crime
recidivism leads to the estimate that sex offenders are four times more
likely than nonsex offenders to be rearrested for sex crimes (5.3 percent
vs. 1.3 percent). Such a fact obscures, of course, the fact that both groups
have low rates of sex crime recidivism. In addition, because the nonsex
offender group is much larger (262,420 released inmates), it results in a

84 AMERICAN CRIMINAL JUSTICE POLICY

far greater number of sex crime recidivists. In absolute terms, there were,
in the fifteen-state study, 517 sex offenders who were rearrested for a sex
crime whereas 3,328 nonsex offenders were rearrested for a sex crime. Put
differently, nonsex offender recidivists comprised the overwhelming bulk –
87 percent – of all released prisoners who were rearrested for a sex crime,
whereas sex offenders comprised only 13 percent.

The results of this study and those like it point to two broad sets of
findings. On the one hand, they suggest that sex offenders are more likely to
be rearrested for sex crimes. On the other hand, they indicate that sex crime
recidivism occurs rarely and that the bulk of such recidivism stems from
crimes committed by nonsex offenders. Moreover, when attention turns to
recidivism in general, not recidivism only for sex crimes, studies show that sex
offenders have lower recidivism rates compared to other types of offenders.
For example, in the fifteen-state federal study, 43 percent of sex offenders
(4,163 of 9,691) were rearrested for any crime within three years of release,
compared to a rearrest rate of 68 percent among nonsex offenders (179,391
of 262,420).62 Perhaps, then, sex offenders merit closer attention if we wish
to address sex crime. However, in so doing, we might fail to address most
sex crime as well as other types of crimes if we focused more on them rather
than other offenders.

CRITERIA FOR DETERMINING THE NEED FOR SEX CRIME LAWS. In short,
if need is defined as increased rates of sex crime, there would appear to
be less, not more, need for ever-increasing numbers of laws aimed at sex
offenders. Of course, need could be defined differently. Perhaps certain
definitions, when coupled with empirical research, might indicate that such
laws should exist and perhaps be expanded. For example, many different
types of sex offenders exist, as reflected in part by the range of crimes
classified as sex offenses (indecent exposure, accessing child pornography,
rape), and maybe the rates of serious sex crimes have dramatically increased
in some areas. If so, a need for a response would be indicated.

Efforts to describe the scope of and trends in different types of sex crimes
would want to take into account that considerable disagreement among
scholars exists concerning the most appropriate ways to classify sex offend-
ers. The problem in arriving at consensus stems in part from the fact that
relatively few studies have empirically tested different theories of sex crime
or the relative recidivism rates of different types of sex offenders. Why? The
low base rates of sex crime and recidivism require that studies have large
study samples, as with the fifteen-state federal study discussed earlier. That
raises study costs considerably, especially if new data, such as information

NEEDS EVALUATIONS 85

about psychological characteristics of offenders, must be collected. The sum
result? At present, we know little about the best way of classifying sex offend-
ers or how to predict which offenders will go on to commit new sex crimes.63

We know even less about the qualitative aspects of sex crimes. As one recent
study emphasized, current actuarial risk assessment instruments “conceptu-
alize violence risk solely in terms of probability of future violence, ignoring
other facets of risk, such as the possible nature, severity, imminence, dura-
tion, or frequency of future violence.”64

A needs evaluation also would consider the causes of any identified prob-
lem because such information would help in developing a policy response
tailored to the specific causes of the problem. At present, there remains
little empirical basis for claims that rates of sex crimes have increased or
that sex crimes today are more serious than in the past. At the same time,
many studies suggest that the primary drivers of today’s sex offender policies
include public concern about crime, which has resulted in part from highly
publicized and sensationalized accounts of select tragic cases, and policy
makers’ efforts to demonstrate a clear record of being tough on crime.65

The policies seem to have emerged from incorrect assumptions not only
about sex crime trends but also other facets of sex crime. For example,
sex offenders can be treated in some instances, but many policy makers
and members of the public believe otherwise.66 In addition, accounts of sex
crimes frequently treat all such crime as the same. Clearly, however, some
crimes, such as indecent exposure, differ greatly in kind and severity from
aggravated sexual assaults.

Assuming that a needs evaluation identified that increases in some types
of sex crimes constituted a social problem meriting a new response, the
question then would be what type of response to pursue. One option simply
would be to expand current efforts. Law enforcement could be directed, for
example, to expend more resources on any reports related to sex offending.
Another option, the one pursued by most states in recent decades, is to
implement many new types of laws. Ideally, however, a given law would be
pursued only after establishing that a need existed specifically for it.

To illustrate, sex offender and community notification registries make
some sense if they are likely to be accessed by residents in areas where
sex offenders reside. They also make sense if most sex crimes involve situa-
tions where the victim does not know the offender. Neither claim, however,
comports with research. For example, stranger victimizations tend to be
rare. Instead, most sex crimes are committed by family members, friends, or
acquaintances of victims. Notably, many sex crime laws have been prompted
by horrific accounts of strangers who victimized children, yet “93 percent

86 AMERICAN CRIMINAL JUSTICE POLICY

of offenses against children are committed by family members and acquain-
tances.”67 Such statistics arguably lead to the inference that sex offender
registries may not be needed because most victims know who victimized
them.

Another example involves residency restrictions, which limit where regis-
tered sex offenders can live. For example, such offenders may not be allowed
to live within 1,000 to 2,000 feet of schools, bus stops, day-care centers, or
other places where children congregate.68 This type of a policy assumes,
however, that most sex crimes involve children and that the crimes involving
children occur in or near these places. That assumption may hold true in
some places but in other places would not.69 The more important point is
that a needs evaluation would highlight if a policy aimed at limiting where
sex offenders live should be pursued or whether perhaps a different type of
policy should be created.

In assessing the need for specific policies, we ideally would have informa-
tion about the relative need for them. Expanding the range of tough sen-
tences for sex crimes may be needed in some states, but perhaps a greater
need exists for a registry, residency restrictions, or treatment. Ideally, of
course, states could implement every possible strategy. They operate, how-
ever, with limited resources and must make decisions about how to allocate
them. To this end, their efforts can be greatly facilitated by needs evaluations
that identify and measure a problem, that identify the most salient dimen-
sions of the problem or those dimensions most amenable to policy influence,
and that describe the relative need for different policies. In the case of sex
crime, one critical type of need to consider is whether the public adequately
understands important nuances concerning sex crimes. Some studies sug-
gest, for example, that the public dramatically overstates the recidivism rates
of sex offenders.70

Not least, needs evaluations of sex crime policies would take stock of
the potential or likely harms associated with specific policies. Sex offender
residency restrictions, for example, may inadvertently increase crime by forc-
ing offenders to live with family members or to be homeless and they may
also preclude offenders from obtaining employment.71 The potential is far
from academic. Jill Levenson and David D’Amora have noted that a vari-
ety of obstacles “leave many offenders no choice but to reside with family
members, but when family members are located within restricted zones,
sex offenders are left with literally nowhere to go.”72 They also have noted
that the recent trend nationally “has been for cities and towns to expand
residency restrictions to 2,500 feet [from schools, parks, playgrounds, etc.],
essentially banning sex offenders from metropolitan areas.”73 Most jobs lie

NEEDS EVALUATIONS 87

in metropolitan areas, and so residency restrictions can preclude sex offend-
ers from finding work in those areas with the best prospects for gainful
employment. The restrictions can also contribute to homelessness, which
makes it more difficult to track and monitor sex offenders.74

Many sex offender policies are quite broad in their focus. Some benefits
may accrue to the broad-based approach, but ideally a needs evaluation
would provide a systematic assessment of different approaches and discus-
sion of which seem most needed. Such an approach might reveal that a
better investment would be in risk-based classification systems, which hold
the potential for “resources to be used more efficiently to intensively moni-
tor, treat, and restrict dangerous offenders while not disrupting the stability
of lower risk offenders and their families.”75

Here, consider that many policies proceed from incorrect assumptions
about the relative risks of recidivism of different groups, assumptions that
are easier to sustain in the absence of valid classification systems. Levenson
and D’Amora have noted, for example, that “some recent policies target
child abusers and exclude rapists, who, as a group, have higher recidivism
rates, and are more likely than child molesters to target strangers and to
cause severe physical injury to their victims.”76 In short, a greater need may
exist for investing in better classification than in a new broad-based policy.
Perhaps not. Perhaps better reentry planning for released inmates would
emerge as especially critical. Or perhaps too much or too little treatment of
sex offenders exists. Again, perhaps not. The point is that effective planning
should include systematic and empirical assessment of the risks associated
with addressing some needs and not others.

Are Current Criminal Justice Policies Needed?

Let us return to the question of whether the prominent policies that populate
the criminal justice landscape today are needed. In this section, I begin with
the focus in recent decades on “get tough” reforms and then turn to the two
illustrations (mass incarceration and sex crime laws) and a number of other
prominent criminal justice policies.

At the broadest level, a central question is whether the plethora of “get
tough” reforms in recent decades was needed. It would appear not. Primarily,
this assessment stems from the fact that states typically do not accurately
estimate the true amount of crime nor do they explicitly articulate how much
of it should be addressed. Unless one has a foundation for determining
whether prior efforts were sufficient, no basis exists for knowing whether
to increase or decrease them. Assume, for example, that a given state is

88 AMERICAN CRIMINAL JUSTICE POLICY

spending more than it needs and then crime rates increase. Arguably, no
additional funding should be invested because the state was overinvesting in
crime-fighting efforts. Similarly, assume that a given state is spending more
than it needs and crime rates decrease. It would make sense that crime
reduction spending should be reduced.

Many observers might argue that funding for crime policies is never ade-
quate. As long as any crime exists, the argument goes, resources should be
dedicated to fighting it. That argument, however, assumes a definition of
need. What, for example, is the proportion of crime that merits a response,
and how much of it can be addressed with available funding? Answers to both
questions require an assessment of the magnitude of the crime problem. Yet
few jurisdictions or states compile independent self-report offending or vic-
timization data on crime, and so have little idea how much crime occurs.
(Law enforcement data invariably capture only a part of all crime, and so are
of questionable use in assessing the scope of or trends in a crime problem
because the data may reflect law enforcement practices as much as they do
crime.77)

The problem confronting criminal justice is broader, however, than one
of not knowing how much crime of various types exists in particular states
and localities or of defining when a level of or change in crime warrants a
response. It also includes a failure to assess the range, scope, and nature
of a wide range of problems within criminal justice. For example, to what
extent do law enforcement officers follow appropriate protocols and laws for
apprehending or questioning subjects or for using force?78 To what extent do
defendants receive adequate legal counsel? To what extent do prosecutors
select and pursue cases in an appropriate manner? How well do judges run
their courtrooms? Are probation or parole officer caseloads excessive?

When we turn our attention to the policies that states have enacted, sim-
ilar questions emerge. The first illustration focused on mass incarceration.
Here, the question is whether this policy has resulted from careful and stud-
ied analysis of crime patterns and trends, deliberation about actual versus
ideal prison sanctioning and sentence lengths, assessment of the causes of
crime and likely returns from incarceration, or the relative need for incar-
ceration versus other policy responses to crime or such problems as prison
overcrowding? It appears not. When one canvasses the literature on legisla-
tive decisions to expand prison system capacity, whether directly through
increased funding for prison expansion or indirectly by enacting laws requir-
ing lengthier sentences that in turn require more prison capacity, little evi-
dence of such stock taking can be found. Put differently, little evidence exists
to suggest that legislatures and correctional systems have systematically

NEEDS EVALUATIONS 89

and carefully conducted needs evaluations for their incarceration invest-
ments.79

Other prison system examples related to mass incarceration abound.
Prison privatization has increased dramatically in recent years, and yet it
frequently occurs in the absence of any clear information about whether
public prisons are operating less efficiently than could be expected from
private prisons. Viewed another way, there remains little evidence to date
that private prisons operate cheaper than public prisons and do so while
achieving comparable outcomes.80

Similarly, supermax prisons have emerged in almost every state over the
past several decades, yet there is little evidence that states undertook needs
evaluations to document whether they were needed.81 The title of an article
on the topic – “The Rise and Rise of Supermax: An American Solution in
Search of a Problem?” – captures this view.82 One common justification is
that they serve to house the “worst of the worst” inmates. States typically do
not, however, provide clear, measurable criteria for identifying such inmates;
consequently, they cannot say how many supermax beds they need or, in
turn, whether existing beds are too few, adequate, or too much.

Sentencing laws provide the central platform for increasing incarceration.
Here, one of the most prominent changes in recent decades has been the
emergence of sentencing guidelines. They have been justified as providing a
way to eliminate unwarranted disparities in sentencing.83 Judges, for exam-
ple, were viewed as being too inconsistent in the sentences they gave, and
the hope was that sentencing guidelines would result in “like” cases being
treated in “like” manner. This argument depends fundamentally on establish-
ing what “like” cases are and how much variability in sentencing such cases
is “warranted.” Notably, however, few states have empirically estimated the
amount of variability in sentencing prior to enacting guidelines.84

Of course, tougher sentencing may emerge for purely politically expedi-
ent reasons, such as to help a prosecutor, judge, or legislator get elected, or
it may result from a myriad of social forces.85 Even so, a needs evaluation
would want to tackle such issues. And it would want to tackle the question
of need on the most straightforward grounds available – to wit, what exactly
is the need for the policy? How much crime of various types deemed to be
worthy of incarceration exists? How much retribution, as measured in years
of incarceration, is appropriate? Retribution is central to any punishment
system. Even so, and as Andrew von Hirsch has highlighted, precious little
empirical research has focused on the “quanta of punishments” that best
reflect or put into effect specific theories of punishment, including retribu-
tion, or the preferences of the public.86

90 AMERICAN CRIMINAL JUSTICE POLICY

Still other criminal justice policy examples exist. The chapter’s second
illustration focused on sex crime laws, which have proliferated and done
so in the absence of any systematic, empirically based needs evaluations.
These laws generally assume that recidivism is greater among sex offenders,
that such offenders are less amenable to treatment and, more generally,
that widely publicized cases reflect a pervasive problem. Each assumption
is incorrect. As a result, states have little credible foundation on which to
assert a need for tougher sanctioning of sex offenders.

Many other examples exist where prominent policies have been pursued
in the absence of clear assessments of need. Drug sentencing and other
“get tough” laws and programs aimed at stiffening penalties for drug-related
offending have proliferated. Here, again, scant attention has been given to
establishing the need for such laws. Certainly, drugs are implicated in much
crime, but that is different from defining and quantifying the need for a
law. In addition, the causal relationship between drugs and crime remains in
question.87 To the extent it is not causal, one might well argue that a need
exists but that it is for treatment, not punishment.88

Mandatory arrest laws have proliferated because of a perceived need to
address a critical problem – domestic violence. An important question, how-
ever, is whether a need for tougher law enforcement was needed more so
than some other response. In answering this question, policy makers would
want to consider such factors as the distribution of domestic violence and
the possibility that it occurs more in some groups than in others and for
different reasons.89 Any significant variation would highlight the need for
a nuanced approach to domestic violence rather than a single one-size-
fits-all policy like a mandatory arrest law. Notably, lawmakers in the 1970s
faced a range of possible policies for reducing domestic violence and an
emphasis on arrest constituted but one of them. A systematic needs evalua-
tion, which was not conducted, would have included consideration of other
legal interventions (e.g., protective orders, mandated treatment), social
service interventions (e.g., shelters for victims, advocacy services), and health
care interventions (e.g., identification of domestic violence in medical set-
tings and reporting of such violence to legal authorities or social service
agencies).90 Today, it remains unclear whether mandatory arrest laws are
more or less needed than these other types of interventions.

In the juvenile justice system, almost all states have recently embraced a
wide array of laws enabling youths to be transferred to the criminal justice
system for processing and sanctioning.91 Again, empirical evidence in sup-
port of the need for such laws is lacking. Few states, for example, have sys-
tematically evaluated how much punishment youths receive in their juvenile

NEEDS EVALUATIONS 91

justice systems and how these youths likely would be punished in the adult
system. An empirical analysis might well reveal, as some studies have, that
youths would receive less severe sanctions.92 Were that true, the need for
transfer laws would be eliminated.

Such examples, while illustrative, appear to be representative of a general
failure of federal, state, and local jurisdictions to identify the need for their
policies. Thomas Blomberg and Gordon Waldo have argued as much, noting
that “juvenile justice policy has rarely been guided or significantly influenced
by research,”93 an assessment equally relevant to criminal justice. The con-
sequences of this failure are at once simple and critical – resources are
likely to be expended on problems that may not warrant attention and other
problems may go largely unaddressed.

Conclusion

Needs evaluations provide many benefits. They can help determine whether
a problem exists and its scope and nature. They can assist in debates about
whether to retain or eliminate a particular policy. More generally, they can
provide guidance about which types of needs should be prioritized and
which policy responses should be pursued. They also can identify the critical
assumptions or research gaps that put discussions of particular “problems”
on a shaky foundation. They help establish the relevant criteria for evaluating
a policy’s effectiveness. And, more generally, they can spur more informed
and constructive discussions about social problems and how best to address
them.

Ideally, needs evaluations occur prior to implementation of any policy.
That simply stands to reason – why expend scarce resources on a nonexis-
tent problem? Moreover, needs evaluations can inform the development of
policies by pointing to critical areas of need that may be especially amenable
to policy influence.

To achieve these benefits, needs evaluations must be informed by empiri-
cal information about the size of a problem as well as trends in the problem,
its distribution socially and geographically, and its causes. They also should
focus on the extent to which a need exists for different types of policy
interventions. Specific social problems, including crime, frequently can be
addressed in numerous ways. Which ones are the most needed may well
depend on the specific nature of a problem as it exists in particular towns,
cities, and states. It may be that existing efforts are sufficient or simply need
to be expanded. But it also may be that new efforts should be undertaken.
In either event, a needs evaluation should also identify the potential benefits

92 AMERICAN CRIMINAL JUSTICE POLICY

and harms of the different options. Doing so allows policy makers to make
more informed decisions about the relative trade-offs associated with dif-
ferent strategies for addressing crime or some problem within the criminal
justice system.

Unfortunately, when one canvasses the criminal justice policy landscape,
it becomes clear that many of the most prominent policies there – including
mass incarceration and sex crime laws – lack any well-established foundation
in an assessment of need. That does not mean the laws should not have
been implemented or necessarily are or will be ineffective. However, it does
greatly diminish their likely success. At the same time, it raises the specter
that considerable amounts of resources have been and are being inefficiently
used in a context where the costs, including the risk of more crime, are great.

Discussion Questions

What are the benefits of evaluating the need for a criminal justice policy?
What are the problems of not evaluating the need for a criminal justice
policy?

How do you conduct a needs evaluation?
How do you define the need for a given policy? For example, what level

of a crime problem is needed before policy makers should respond?
Does there simply have to be any amount of crime? If so, how should
policy makers decide how to allocate funds to specific types of crimes?
To specific types of policies?

What amount of increase in a problem (e.g., crime) is needed before policy
makers should take action? Should specific policies be reduced or elim-
inated if a social problem decreases? For example, if crime decreases,
should policy makers reduce criminal justice funding because of the
reduced need?

Many types of problems exist in criminal justice, and crime is but one
of them. Law enforcement officer, defense counsel, prosecutor, or cor-
rectional personnel may engage in misconduct. Courts may suffer from
case overload. The public may hold incorrect views about the preva-
lence of crime or the conditions of confinement. What priority should
be given to these and other types of problems? On the basis of what
criteria?