Princeton University

Responding to Juvenile Crime: Lessons Learned

Author(s): Peter W. Greenwood
Source: The Future of Children, Vol. 6, No. 3, The Juvenile Court (Winter, 1996), pp. 75-85
Published by: Princeton University
Stable URL: http://www.jstor.org/stable/1602595
Accessed: 14-12-2015 00:19 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

Princeton University is collaborating with JSTOR to digitize, preserve and extend access to The Future of Children.

http://www.jstor.org

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
 75

to Juvenile
Responding
Crime: Lessons Learned
PeterW.Greenwood

Abstract
Although the dispositional options for delinquent youths have diversified over the past
20 years, the debate about the most effective treatment of young offenders continues. Peter W Greenwood,
This article reviews the existing evaluations of variousjuvenile corrections alternatives. Ph.D., is directorof the
In contrast to the conclusions of scholars in the late 1970s that "nothing worked" in criminaljusticeprogram
juvenile corrections, MarkLipsey's 1992 meta-analysisof more than 400 evaluations of
at R4Nl)
Cporation
juvenile programs reported an average 10% improvement in recidivism rates for all
in SantaMonica,CA.
the programs evaluated. Lipsey's meta-analysisfound a significant advantage in com-
munity-based programs run by private providers compared with large custodial insti-
tutions, such as traditional state training schools. The most effective privately run
community programs have high levels of intensity and duration, multiple modes of
intervention, and a great deal of structure. Nonresidential, dawn-to-duskeducational
programs that work with both juveniles and their families are good examples of this
type of programming. Boot camps for juvenile offenders have proliferated because of
their political popularity. A recent evaluation of eight of these programs showed that
four had no effect on recidivism, one resulted in higher recidivism rates, and three
showed improvements in some recidivism measures. Probation and diversion, the
most common of all juvenile court dispositions, are not effective options for youths
with multiple risk factors. Juvenile courts need an array of dispositional options, the
ability to monitor program effectiveness, and the flexibility to find the appropriate
placement for each juvenile offender.

delinquents in most states consisted of large congregate training

schools. In 1971, Jerome Miller, then head of the Department of
Youth Services (DYS) in Massachusetts, took a significant step when he
abruptlyremoved most of the youths from that state's training schools and
placed them in a variety of small community-based institutions and pro-
grams.1Those youths requiring secure care were placed in a number of
small (typically20- to 30-bed) facilities. Miller'sMassachusettsreforms were
motivated as much by concerns for the welfare of the youths as for the
impact these reforms might have on recidivismrates.
To this day,the juvenile corrections systemin Massachusettscontinues to
operate with only a fraction of the secure beds utilized in many other states,
and the majorityof its youths are still placed in a varietyof small, privately
TheFutureof ChildrenTHE COURTVol.6 * No. 3 - Winter1996
JUVENILE

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
76 THEFUTURE - WINTER
OFCHILDREN 1996

operated programs. Over the past two decades, a number of other states,
including Maryland,Pennsylvania,and Utah, have adopted similarreforms
in their systems.2
An evaluationof the Massachusettsreforms,which compared outcomes
for samples of youths committed to DYS before and after the reforms
occurred, found higher averagerecidivismrates3for the postreformyouths.
These resultswere partlyexplained by a decrease in the number of less seri-
ous offenders being committed to DYS.However,in those partsof the state
where the new reformswere most successfullyimplemented, as indicated by
the characterand diversityof their programs on an objective rating scale,
postreform recidivismrates appeared to be lower.'A more recent attempt
to evaluate the Massachusettsmodel4 compared recidivism rates for all
youths released by the MassachusettsDYSin 1985 with those reported for
severalother states (California,Florida,Pennsylvania,Utah, and Wisconsin)
and found the Massachusettsrates to be among the lowest. The 1985
Massachusettsrelease cohort also had lowerrecidivismratesthan the groups
studied in the mid-1970s by Coates and colleagues. However, the signifi-
cance of these comparisons is obscured by systematicdifferences in the
characteristicsof youths committed to these state programs and by varia-
tions in the reporting ofjuvenile police and court contacts.

TrendsinJuvenile ofjurisdictionsare expanding theiruse of in-

home tracking, day treatment, and family
Corrections
Policy treatment programs first pioneered in
Twenty-fiveyears after the Massachusetts Massachusetts.
experiment, the debate over the most effec-
tivemethod of dealingwith chronic and seri- in Evaluation
Difficulties
ous juvenile offenders continues. Maryland Little is known about which correctional
recently attempted to copy some of the approaches are best suited to particular
Massachusettsreformsby closing one of two typesof youths at least in part because there
state-runjuvenile institutions,the Montrose have been only a limited number of careful
Training School, which at that time had a programevaluations.Although manyjudges
population of approximately320 youths. A and correctional practitioners appear to
postrelease sample of youths who probably believe they can assessa program'seffective-
would have been sent to Montrosehad it not ness by observingand talkingwith its partic-
been closed had higher recidivismratesthan ipants, such anecdotal evaluation is often
groups who completed Montrose or were clearlywrong. A youth's improved manners
there when it closed.5 While proponents and respectfuldemeanor may quicklydisap-
continue to urge other states to adopt pear when he returnsto the influence of his
Massachusetts-typereforms, a number of old neighborhood. Moreover,the fewyouths
states, with California the most prominent who do come back to visit a programor the
example, continue to rely heavily on large juvenile court are probablynot representa-
custodialinstitutionsfor handling their most tive of the typicalparticipant.
seriouslydelinquent youths.
The effectiveness of any one particular
The most popular recent reform in the program is unknown unless it is evaluated.
handling of serious or chronically delin- At a minimum, follow-up recidivism data
quent youths is the increasinguse of waiver (rearrest or self-reports) for an adequate
or transferofjuvenile cases to adult criminal sample of participantsmust be compared
court. Between 1988 and 1992, the number with those for an appropriatecontrol group.
of juveniles waived to criminal courts Because the primarygoal of juvenile correc-
increasedby 68%,from 7,000 to 11,700.6For tions is community safety,recidivismis used
less seriously delinquent youths, a number as the minimal measure. Other measures,

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
Respondingto JuvenileCrime:LessonsLearned 77

such as subsequent attainment of a general ishments associatedwith a particularpattern

educational development (GED) certificate of behavior.
or employment, are used less frequently.
Many leading theorists now favor inte-
To be useful to policymakersand practi- grated or interactional models10which com-
tioners outside the program, the results of bine the effects of strwain,control, social
any programevaluationmust include a com- learning, and other theories based on inher-
plete description and documentation of the ited physiologicalor psychologicalphenom-
program'sprocedures and methods of treat- ena. Many also believe that the chain of
ment. Administratorstend to incorporate a causality between primary risk and protec-
varietyof currentlypopular methods in any tive factors (substance abuse, association
one program, rather than testing one single with delinquent peers, attachment to school,
well-defined approach, creating an addition- and so on) is not unidirectional or consis-
al impediment to program evaluation. tent over time. For example, not only will the
weakening of bonds increase delinquency,
To reach a definitive conclusion about
the effectiveness of a particularapproach, it
is best if results from several individual eval- The best
uations are pooled. The positive effects
predictorsoffuture delinquency
observed in a particularstudy may be attrib- are past involvement in delinquency, the
utable to chance, an unusuallyeffective pro- presenceof otherrelatedproblembehaviors,
gram administrator, staff, or particularly and associationwith delinquent peers.
ineffective programming applied to the
comparison group. The preferred method
for assessing results from a number of stud- but participation in delinquent acts is also
ies involves the statistical techniques of likelyto reduce the strength of bonds to con-
meta-analysis,which allow the reviewer to ventional institutions.As youths mature, the
compare results across programs in a com- influence of parental bonding declines
mon frameworkwhile controlling for differ- while that of peers is likely to increase.
ences in sample characteristicsand research
methods. Results from several recent meta- A consistent finding acrossseveralrecent
of
analyses juvenile corrections evaluations epidemiological studies is that the best pre-
are discussed later in this article. dictors of future delinquency are past
involvement in delinquency, the presence of
TheDevelopmentof other related problem behaviors (for exam-
ple, drug or alcohol use, problems at school,
DelinquentBehavior truancy,early sexual experience), and asso-
To think seriously about intervention pro- ciation with delinquent peers.11Youthsfrom
grams for juvenile delinquents, it is neces- impoverished homes, homes with only one
sary to understand how patterns of delin- parent or guardian, homes in which one or
quency develop. Over the past half century, both parents exhibit some kind of problem
delinquency theorists have developed a behavior such as substance abuse, alco-
number of increasingly complex causal holism, or mental illness, or homes in which
models to explain the involvement of youths the parents exhibit poor parenting practices
in delinquent activities. Strain theorists7 are more likely to become delinquent.
believe that delinquency occurs primarily
among lower-income youths who are frus- Somewhere between 30% and 40% of all
tr-atedby their to
inability satisfy needs and boys growingup in urban areasin the United
aspirations through legitimate means. States will be arrested before their 18th
Controltheorists8argue that the restraining birthday.12 Most of those arrestedwill not be
influences exerted by ajuvenile's bonds with arrested again. For those that are, each suc-
conventional community institutionssuch as cessivearrestwill place them at a higher level
family, school, and church prevent delin- of riskuntil, after five or six arrests,they will
quent behavior.Sociallearningtheorists9view have a greater-than-90%chance of being
the decision whether or not to engage in arrested again. Those who reach five arrests
delinquent behavior as the result of the net have been labeled by criminologist Marvin
effect of all the perceived rewardsand pun- Wolfgangas chronic offenders: the 6% of all

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
78 THEFUTURE - WINTER
OFCHILDREN 1996

boys who account for more than 50% of all A number of critics responded to the
arrestsfor their age group.'3 generally defeatist tone of the "nothing
works"reviewsby pointing out that each of
the reviewscontained numerous references
WhatWorksinChanging to programsthat did work and that a more
DelinquentBehavior rigorous screening on therapeutic integrity
One of the primarygoals of thejuvenilejus- and program quality could identify pro-
tice system has alwaysbeen the salvationof grams that were effective.'1-20
the minor from a self-destructivelife of
crime. While public protection has always When reviewers began using the new
been an implicitgoal of the system,in recent technique of meta-analysis, systematic
yearsmanystatelegislatureshave also explic- reviews of the evaluation literature began
itly included it as one of the system'sprima- turning up evidence of consistent positive
ry goals. Both of these objectivesrequire the effects for particular intervention strate-
juvenilejustice systemto pursue the difficult gies.21-24The most thorough of these meta-
task of changing individual behavior pat- analysesforjuvenileswasconducted by Mark
terns. Changes in delinquent behavior are Lipsey25and covers more than 400 pub-
pursued by the juvenile justice system lished and unpublished evaluationsof juve-
through the competing mechanisms of nile programs conducted since 1945.
deterrence and rehabilitation. Lipsey'sstudy also goes beyond most of the
others in coding the characteristicsof the
Deterrence refers to the discouraging interventions, the context in which they
effect that threatened punishments have on were administered, the population that
potential offenders. Researchefforts to date received them, and the evaluationmethod-
have not supplied compelling evidence that ology.At this time, Lipsey'sfindings serve as
increasing the severity of sanctions deters the baseline from which all observations
about the treatmentofjuvenile offendersare
made. The findings from that reviewwhich
in
Changes delinquent behavior are pwsued pertain to this articleare summarizedbelow.
bythejuvenilejusicesystemthroughthe TheOverallEffectiveness
of
cwmpeingmechanisms of deterrenaand Treatment
Correctional
rehabilitation. Starting from a baseline recidivismrate of
50%,Lipseyreportsan average5-percentage-
point reduction in recidivism or a 10%
more crime, but availableevidence suggests improvementfor all the programsevaluated.
that the certaintywith which sanctions are Because effect sizes26are attenuated by the
applied is much more important than their unreliabilityof the outcome measures used
This finding is often used tojustify
severity.14 (recorded arrests),it is likely that the actual
the imposition of more consistent sentenc- effects of treatmentare closer to 10 percent-
ing and requiring some sanctions for all age points, or a 20% reduction in recidivism
offenders, regardlessof their prior record. from the comparison groups. As might be
expected, Lipsey found that evaluationsin
Rehabilitation refers to the process of which control groups received some inter-
attemptingto change the propensityof indi- vention showed less contrast (smallereffect
vidual offenders to commit crimes through size) than those in which the comparison
training,education, and other services.The group received no treatmentat all.
rehabilitative role of juvenile correctional
programs was set back considerably by a The remainder of this section summa-
series of scholarly reviewsof the evaluative rizes some of the other key variables that
literaturein the late 1970swhich were inter- appear to be associatedwith effect size.
preted as saying that "nothing worked"or,
more precisely, that there was insufficient U InstitutionalversusCommunity
Placements.
evidence to conclude that treatment Treatmentin public facilities,custodialinsti-
worked or that any one particularmethod tutions, and the juvenilejustice systemwere
of treatment was more effective than any associatedwith smallereffect sizesthan treat-
other.15-17 ment in community-basedprograms.

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
Respondingto JuvenileCrime:LessonsLearned 79

a Treatment Modality.Structured approach- and early 1980s28found that the majorityof

es that attempt to train subjects in new juveniles sentenced in adult courts received
behaviors and skills were associated with probation or some other alternativeto incar-
larger effect sizes than less structured pro- ceration.A studyofjuvenile and young adult
grams such as counseling and general super- robbery and burglary cases disposed of in
vision. three California cities found that juvenile
court dispositions were more severe, after
a Intensityand Exposure.A "modest positive controlling for offense severity and prior
relationship"was found between effect size record.29A similar studys0in an East Coast
and treatment intensity and duration. jurisdiction found no significant differences
in the minimum and maximum terms
a RiskLevelof Clients.There was a slight ten- imposed on similarjuvenileand young adult
dency for studies ofjuveniles with higher risk defendants in juvenile and adult court.
levels, as indicated by their prior records or
the recidivism rate of the control group, to However, a more recent study by the
show larger effects of treatment when com- General Accounting Office31found that the
pared with those whose risk levels were majority of juvenile offenders now being
lower,but this difference was not statisticallytransferred to criminal courts and convict-
significant. ed of serious offenses are receiving prison
terms. The study did not report any results
a Organization,Staffing, and Management. on time served, conviction rates, or recidi-
Higher levels of involvement by the re- vism-all factors that need to be consid-
searcher in the treatment program were
associated with higher effect sizes, suggest-
ing that operating the program in a manner Despiteintensivelobbyingeffortsto dose
consistent with its original intent and design
is important.
lar juveniletraininginstitutions,many
tooperate
statescontinue these
facilities,
a MethodologicalFactors. Greater attrition someof whichwerebuiltbefore1900.
from treatment or control groups was asso-
ciated with smaller effect sizes, as were larger
samples and more outcome variables.Any ered in assessing the impacts of waiver. (For
significant attrition from either experimen- further information on waivers,see the arti-
tal or control groups within an evaluation cles by Ainsworth and by Snyder in this
study can affect the comparabilityof the two journal issue.)
groups and make the interpretation of out-
come resultsproblematic. Some researchers CustodialInstitutions
question the appropriateness of including Despite intensive lobbying efforts to close
nonrandomized studies in meta-analyses. large juvenile training institutions, many
states continue to operate these facilities,
Assessingthe Primary some of which were built before 1900. The
JuvenileJustice primary criticisms leveled against tradition-
al state training schools have been that they
Alternatives
Dispositional offered sterile and unimaginative pro-
Once a youth is adjudicateddelinquent, the grams, were inappropriate places to run
central task of the juvenile court is to order rehabilitativeprograms, and fostered abuse
and oversee the appropriate disposition. and mistreatment of their charges.32,33
The principal dispositionalalternativesavail- Criticsseek to replace them with community-
able in the juvenile justice system are dis- based programs because such programs
cussed below. provide a more realistic setting in which
youths can learn and practice social skills
WaiverorTransfer to AdultCourt and because they allow youths to develop
The increased use of transfersofjuveniles to or maintain contacts with their families,
adult court is usually justified on the schools, and employers. The large institu-
grounds that the criminal courts can impose tions are defended on the grounds that
more severe sanctions than the juvenile community programs have not been
court. Studies conducted in the late 1970s27 shown to be more effective and/or that

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
80 OF CHILDREN
THEFUTURE 1996
- WINTER

advantages of community programs can be

obtained by breaking a large institution's
population and staff down into small self-
contained units which more closely parallel
the size of community programs. For exam-
ple, in Massachusettsand Utah, secure insti-
tutions are built to accommodate client pop-
ulations of no more than 20 or 30 youths.
Other strategies include involving commu-
nity members in the daily activities of the
institutionand simulatingthe social and eco-
nomic systemsthat are found on the outside.
0r The training schools run by the Michigan
Department of Social Services represent an
excellent model of this approach, and the
."-•
:::
privately run Glen Mills School in
:-ii•
Pennsylvaniais the premier example of what
a dedicated management and staff can
achieve in a closed-campussetting.34

WildernessPrograms
A number of privateproviderswho are will-
-- ing to take the kind of serious delinquents
who are usually placed in state training
schools have found that they can operate
U
more open programs in isolated wilderness
5.
camps,where the remoteness of the location
provides the necessary security and over-
coming the challenges of the environment
0'
UP~
C,,?:1 becomes an aspect of the program. An eval-
?~i
uation based on 90 youths committed to
VisionQuest, a prominent provider of this
type of program, found that a significantly
their institutional programs offer more pro- smaller fraction were rearrested during the
tection to the public. first6 to 18 months after their release than a
similargroup who had participatedin a pro-
In terms of postrelease arrest rates, the bation department camp program (empha-
Lipseyanalysisdescribed in the previoussec- sizing primarily custody, classroom educa-
tion found a significantadvantagein favorof tion, and recreation) the previous
year.35,6
community programs. Whether or not this
advantage is sufficient to justify the some- GroupHomes
what greater risk created by letting youths For youths who need to be placed out of
reside in the community while participating their homes but do not represent such a risk
in rehabilitative programs is a function of that they need to be placed in a secure pro-
the specific risksrepresented by the popula- gram, group homes are the placement of
tion under consideration, the intensity of choice.'37 They can be designed to provide
their supervision while in the community, as much structure as necessary,while allow-
and the community's tolerance for the kind ing youths to attend school and/or partici-
of risk they represent. The large growth in pate in other community programs.Because
privatecommunity placements over the past group homes can take advantageof existing
10 years demonstrates that many communi- community resources such as education,job
ties believe the risksare worth it. training, and medical services, they are less
costly to the juvenile justice system.
Where large custodial institutionscontin-
ue to operate, the Lipsey analysis provides Aftercareand Parole
guidance concerning effective program- Assistance with community reintegration,?
ming within these settings. Some of the relapse prevention,s service advocacy, and

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
 to JuvenileCrime:Lessons
Responding Learned 81

surveillancearejust a few of the manyjustifi- challenging outdoor activities were used

cations cited to support the provision of to support team work and increase self-
aftercareservices to graduatesof institution- confidence, found that youths assigned to
al programs. In spite of these arguments, the challenge camp had higher recidivism
very few jurisdictions provide aftercare for rates than similaryouths assigned to regular
any significant fraction of their youth, and training schools. The program included an
there is little in the wayof evaluativeresearch. ambitious nine-month aftercarecomponent
A recent evaluation of intensive follow-up run by private providers.44In the Michigan
supervision in Pennsylvaniafound that this program, 22% of participatingyouths were
effort was effective in reducing postrelease escalated to regular training school place-
arrests, primarily by reincarcerating failing ments during the residential phase, and
youths sooner than would be likely without
such supervision.39Another evaluation of
efforts to provide intensive postrelease seriousaboutworkingwith
Jurisdictions
supervision and assistance to both youths
and their families in Detroit, Michigan, and youths have developed a wide array of
Pittsburgh,Pennsylvania,found no effect of programsofferingvaryingdegreesof
such efforts on recidivism or self-reported supervisionand servwis.
delinquency.40 The Lipsey25meta-analysis
found little or no effectiveness associated
with traditionalparole. another 47% were escalated or went absent
without leave (AWOL) during the commu-
BootCamps nity aftercarephase. Instead of participating
Boot camps for either juveniles or adults in a 90-day residential program with nine
have recently attracted wide political sup- months of aftercare,youths assigned to the
port. Both politicians and the public appear Michigan challenge program averaged 322
to be impressed by the strictdiscipline, phys- daysin some form of custodyduring the first
ical labor, and military drill that such pro- two years after placement in the program.
grams entail. Since their inception in 1983,
more than 29 states, the Federal Bureau of Community Supervision and
Prisons, 9 juvenile systems, and 10 local Programming
jurisdictions have initiated such programs,41 Jurisdictions serious about working with
without any evidence as to their effects on youths have developed a wide arrayof pro-
recidivism. In addition, these programs are grams offering varying degrees of supervi-
supposed to cost only a fraction of a regular sion and services.At the low-intensityend of
prison or training school commitment the spectrum are performance contracts
because they require just 90 days in a resi- and higher-than-usualfrequencies of con-
dential setting.42 Many boot camp pro- tact with volunteer case manager/trackers.45
grams also involve extensive aftercare At the high-intensity end are dawn-to-dusk
components. education and training programs such as
those run by Associated Marine Institutes
A recently completed evaluation of eight (AMI) in more than a dozen (mostly south-
state boot camp programs for young adult ern) states. These programs provide meals,
offenders concluded that four of the pro- work with parents, arrange supplementary
grams had no effect on recidivism, one recreationalor work experience activitieson
resulted in higher recidivismrates (possibly weekends, and can lead to certification as a
because it put little emphasis on treatment), scuba diver.It is primarilyevaluationsof the
and three showed improvements in some earlyforerunners of these rapidlyspreading
recidivismmeasures.43Boot camp programs nonresidential programs, which were often
are difficult to evaluate because of the high developed or initiated by academic
washout or dropout rate of program partici- researchers, that provide the basis for
pants. In the three programs found to be Lipsey'sfinding that noncustodial, nonjuve-
effective by MacKenzie and colleagues, the nilejustice programsare more effective than
average dropout rate was 38%. other forms of treatment.

A recent evaluation of a 90day juvenile Lipsey25argues that no one approach is

wilderness boot camp in Michigan, where clearlybetter than the others. In addition to

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
82 THEFUTURE - WINTER
OFCHILDREN 1996

privatesponsorshipand communitysetting, "probation as usual" was the only regular

the more effective programs all appear to
have multiple modes of intervention,higher
levels of intensityand duration, and greater
structure in common. They are more the-
oretically rigorous, ambitious, and tightly
executed.
juvenile justice intervention that, when
applied to control groups, did not reduce
the magnitude of the difference in effects
between experimental and control groups.
In other words, regular probation is effec-
tivelyno treatmentat all.

and Diversion
Probation A large body of research on intensively
Probationor diversionis the most frequent supervised probation and reduced proba-
dispositionfor alljuvenilearrestsbecauseit is tion caseloads suggests that this is not an
inexpensive. Because the majorityof juve- effective strategy for reducing recidivism
niles who are arrestedone or two times are rates or improving public safety.46And,
never arrested again, as a practical matter, despite a few well-publicized studies that
probationor diversionis the most reasonable found restitution programs effective in
option in most cases. However, for youths reducing recidivism rates,47restitution was
with multiple risk factors (several prior found by Lipsey to make only a small
arrests,arrestsat an earlyage, drug or gang improvement in the effectivenessof proba-
involvement, or parental problems) proba- tion or diversionprograms.
tion and diversionare not effectiveoptions.
Augmented forms of probation can be
Probationand diversioncover a range of effective.Probationcombined with intensive
low-level services, monitoring, and perfor- supervision, restitution, or other types of
mance obligations. Probation is supervised enhancement was found by Lipsey to have
by the juvenile justice department once a had positive effects, albeit considerably
fewer than the more ambitious and struc-
tured multimodalmethods.
An overwo*edprobationofficerwhoseesa
The Lipsey meta-analysisalso suggests
dient oncea monthhas littleabilityto that more intensive behavioral and skill-
monitorthedient'sbehavioror to exertan oriented interventionsmight be more effec-
influenceoverhis lfe. tive than routine probationor diversionwith
high-riskdelinquents if these interventions
focus on solving problems with both the
youth is adjudicateddelinquent while diver- youths and their families.This is the strategy
sion is overseen by social servicesor a com- represented by Eight Percent Solution, a
munity agency in cases where prosecution program established by the probation
was suspended. As a condition of probation department in Orange County, California,
or diversion, a youth may be required to which targets the 8% of juvenile arrestees
refrainfrom certain activitiesand complete who account for 50% of their arrests.
certain programming, community service,
or restitutionobligations.In both instances, MethodsThatDo
Treatment
there is an implied threat of a harsher MoreHarmThanGood
sanction for any subsequent offense. In Some intervention methods appear to have
some jurisdictions, such as the state of negative effects--that is, to lead to worse
Washington,the diversion process is highly outcomes than no treatment at all. Follow-
formalized,with specific guidelines for sanc- up studiesof the Cambridge-Somerville pro-
tions to be imposed by a communityboard. ject, which provided trainedvolunteer men-
In otherjurisdictions,it is routine and infor- tors to at-riskyouth, have clearlyshown that
mal with little effort at customizing or even well-intended interventions can have
enforcing sanctions. harmfuleffects.AsIn Lipsey'sanalysis,nonju-
venile justice individual counseling, voca-
In reality,an overworkedprobation offi- tional programs, and deterrent schemes
cer who sees a client only once a month has such as Scared Straightproduced negative
little ability either to monitor the client's averageeffects.Another method, or possible
behavior or to exert much of an influence side effect, often cautioned againstby delin-
over his life. In the Lipsey meta-analysis,25 quency researchers,is bringing delinquents

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
Respondingto JuvenileCrime:LessonsLearned 83

together in any unsupervised or unstruc- the program will be damaged for all of the
tured setting. Negative peer pressure youths involved.
appears to be particularlystrong in support-
ing continued delinquency.11 As courts determine appropriateindivid-
ual dispositions, considerations of public
Conclusion safetyand appropriatetreatment should not
To achieve an optimal balance between be confused. While a youth's instant offense
effective treatment, accountability,and com- may be a useful indicator of his potential risk
munity protection, the available evidence to the community,it is not a good indicator
suggests that juvenile courts require access of what kind of programming is required to
to a broad array of dispositional options. change his behavior.All things being equal,
These extend from formal diversion to local treatment programs run in community set-
accountability boards or social services, tings are likely to be more effective in reduc-
through in-home supervision and day treat- ing recidivismthan similarprogramsprovid-
ment, to a varietyof out-of-homeplacements ed in institutions.Whether or not the risk of
providing different levels of security and
types of treatment. Methods that concen-
trate on changing individual behavior pat- All thingsbeingequal,treatment programs
terns which lead to delinquency and improv- run in are
ing pro-social skills are likely to be 10% to
community settings likelyto be
20% more effective in reducing subsequent moreeffectivein reducingrecidivismthan
delinquency than less structured programs similarprograms pmrovided in institutions.
that emphasize individual counseling or
general education.
additional crime posed by allowing particu-
The evidence also suggests that a certain lar categories of youths to remain in the
degree of skill and experience is required to community can be justified by the reduction
develop and run both institutionaland com- in their recidivism rate can be determined
munity-basedcorrectionalprogramsthat are only by assessing both of these parameters
effective in changing behaviors.While effec- and weighing them carefully.This same cal-
tive programscan be run by either public or culation can be made for youths who are
privateagencies, meta-analysesand program considered for transfer to adult criminal
surveyssuggest that such programsare more courts, although no state appears to be
likely to be developed by private providers, doing it explicitly. There is currently no
probably because they are usually held to direct evidence regardingwhether increases
higher standards of performance and in potential punishment and immediate
accountability,and are more free to innovate. community protection achieved through
adult court are accomplished at the cost of
To ensure that any program is being run higher recidivismrates.
effectively,ajurisdiction should monitor the
recidivism rate and pro-social accomplish- One of the more promising areas for
ments (GED certificate, employment, and additional juvenile court intervention and
the like) of its graduatesand compare them activismappears to be with younger offend-
with the graduates of other programs han- ers whose characteristicsplace them at high
dling similar youths. Such comparisons risk for further delinquency. Appropriate
should be run for all programs receiving interventionswill usuallyinvolve the parents
more than 5% of the placements from any and provide a mix of activities,services,and
given court and adjustedfor the risklevels of community sanctions.
the youths being placed.
There are no simple solutions or magic
For the juvenile courts, the challenge of cures for thejuvenile court to apply to prob-
finding an appropriate placement for any lems of juvenile delinquency. The most
specific youths may require experimentation effective systems and processes will be flexi-
and adjustments. Programs should not be ble and continuously experimental in their
required to retain youths who will not abide approach, provide a wide range of treatment
by or cannot adjust to the program's rules and placement options, and be accountable
and conditions. Otherwise, the integrity of for their results.

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
84 THEFUTURE - WINTER
OFCHILDREN 1996

1. Coates, R., Miller,A., and Ohlin, L. Diversityin a youthcorrectional

system:Handlingdelinquentsin
Massachusetts.Cambridge, MA:Ballinger, 1978.
2. Krisberg,B., and Austin,J.F. Reinventingjuvenilejustice.NewburyPark,CA:Sage, 1993.
3. Unless otherwise noted, in this article the term recidivism
refers to rearrestfor any new crime
during some specified time period, usuallyone year following release.
4. Krisberg,B., Austin,J., and Steele, P. Unlockingjuvenilecorrections:
EvaluatingtheMassachusetts
Departmentof YouthServices.San Francisco,CA:National Council on Crime and Delinquency,
1989.
5. Gottfredson, D.C., and Barton, W.H. Deinstitutionalizationofjuvenile offenders. Criminology
(1993) 31,4:591-609.
6. Snyder,H.N., and Sickmund, M.Juvenileoffenders and victims:A focus on violence-Statisticssum-
mary.(Prepared by National Center forJuvenile Justice.) Washington,DC: Office ofJuvenile
Justice and Delinquency Prevention, May 1995.
7. Cohen, A.K.Delinquentboys:Thecultureof thegang. New York:Free Press, 1955.
8. Hirschi, T. Causesof delinquency.
Berkeley:Universityof CaliforniaPress, 1969.
9. Bandura,A. Principlesof behaviormodification.
New York:Holt, Rinehartand Winston, 1968.
10. Thornberry,T.P.Towardan interactional theory of delinquency. Criminology (1987) 25:863-91.
11. Elliott, D.S., Huizinga, D., and Menard, S. Multipleproblem substanceuse, and
youth:Delinquency,
mentalhealthproblems. New York:Springer-Verlag,1989.
12. Blumstein,A., Cohen,J., Roth,JA, and Vishers, C.A., eds. Criminalcareersand careercriminals.
Vol. 1. Washington,DC: National Academy Press, 1986.
13. Wolfgang,M., Figlio, R.M., and Sellin, T. Delinquency
in a birthcohort.Chicago:Universityof
Chicago Press, 1972.
14. Blumstein,A., Cohen,J., and Nagin, D. Deterrenceand incapacitation:
Estimatingtheeffectsof crim-
inal sanctionson crimerates.Washington,DC: National Academy of Sciences, 1978.
15. Lipton, D., Martinson,R., and Wilks,J. Theeffectiveness
of correctional A surveyof treat-
treatment:
mentevaluationstudies.New York:Praeger, 1975.
16. Greenberg, D.E The correctional effect of corrections:A surveyof evaluations.In Corrections
and punishment.D.F Greenberg, ed. BeverlyHills, CA:Sage, 1977.
17. Sechrest, L., White, S.O., and Brown,E.D., eds. Therehabilitation
of criminaloffenders:
Problems
andprospects. Washington, DC: National Academy of Sciences, 1979.
18. Palmer,T. Martinsonrevisited.JournalofResearchin Crimeand Delinquency (1975) 12:133-52.
19. Palmer,T. The effectiveness issue today:An overview.FederalProbation(1983) 46:3-10.
20. Gendreau, P., and Ross, B. Effectivecorrectional treatment:Bibliotherapyfor cynics. Crime&
Delinquency(1979) 25:463-89.
21. Garrett,C.J.Effects of residential treatment on adjudicated delinquents:A meta-analysis.
JournalofResearchin Crimeand Delinquency (November 1985) 22,4:287-308.
22. Kaufman,P. Meta-analysisof juvenile delinquency prevention programs. Unpublished mas-
ter's thesis. Claremont GraduateSchool, 1985.
23. Gottschalk,R., Davidson,W.S.,Gensheimer, L.K, and Mayer,J. Community-basedinterven-
H.C. Quay,ed. New York:Wiley,1987.
tions. In Handbookofjuveniledelinquency.
24. Andrews, D.A., Zinger, I., Hoge, RD., et al. Does correctional treatment work?A clinically-
relevant and psychologically-informedmeta-analysis.Criminology (1990) 28:369-404.
25. Lipsey,M.W.Juvenile delinquency treatment:A meta-analyticinquiryinto the variabilityof
explanation.T. Cook, H. Cooper, D.S. Cordray,et al., eds. New York:
effects. In Meta-analysisfor
Russell Sage Foundation, 1992, pp. 83-126.
26. Effectsizeis a technical term used to interpret the value of a gain achieved through a new
treatment or intervention. It is the ratio of the gain to the standarddeviation. For a further
description of effect size, see TheFutureof Children(Summer/Fall 1995) 5,2:120.
27. Hamparian,D.L., Estep, L., Muntean, S., et al. Youthin adultcourts:Betweentwoworlds.Major
issuesin juvenilejusticeinformationand training.Washington,DC: Office ofJuvenileJustice and
Delinquency Prevention, 1982.
28. Bortner,MA., Traditionalrhetoric, organizationalrealities:Remand ofjuveniles to adult
court. Crime& Delinquency(1986) 32:53-73.

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions
Respondingto JuvenileCrime:LessonsLearned 85

29. Greenwood, P.W.,Lipson, A., Abrahamse, A., and Zimring, E Youthcrimeandjuvenilejusticein

California.Santa Monica, CA:RAND Corporation, 1983.
30. Fagan,J. Thecomparative impactsofjuvenileand criminalcourtsanctionsfor adolescent
felonyoffend-
ers.Washington, DC: National Institute ofJustice, 1991.
31. U.S. General Accounting Office. Juvenilesprocessedin criminalcourtsand casedispositions.
GAO/GGD-95-170. Washington, DC: U.S. Government Printing Office, 1995.
32. Bartollas, C., Miller, S.J., and Dinitz, S.Juvenilevictimization:Theinstitutionalparadox.New York:
Wiley, 1976.
33. Feld, B. Neutralizinginmateviolence: Juvenileoffendersin institutions.
Cambridge,MA:Ballinger,1977.
34. For further information about these programs, contactJ. Otis Davis, Michigan Department of
Social Services, (517) 335-3489; Glen Mills School, Glen Mills Road, Concordville, PA 19311,
(610) 459-8100.
35. Greenwood, P.W.,and Turner, S. The VisionQuest program:An evaluation.Santa Monica, CA:
RAND Corporation, 1987.
36. The Associated Marin Institute's Last Chance Ranch is a prominent provider of these types of
services and is cited as an exemplary program in Howell, J.C., ed. Guidefor implementing the
comprehensive strategyfor serious,violent,and chrnic juvenileoffenders.
Washington, DC: U.S.
Department ofJustice, Office ofJuvenile Justice and Delinquency Prevention, 1995.
37. See note no. 36, Howell, p. 14.
38. Altschuler, D.M., and Armstrong, T.L. Intensive aftercare for the high-riskjuvenile parolees:
Issues and approaches in reintegration and community supervision. In Intensiveinterventions
withhigh-riskyouths:Promisingapproachesin juvenileprobationand parole.Troy L. Armstrong, ed.
Monsey, NY:Willow Tree Press, 1991.
39. Sontheimer, H., Goodstein, L., and Kovacevic,M. Philadelphiaintensiveaftercare probationevalua-
tionproject.Harrisburg,PA:PennsylvaniaCommission on Crime and Delinquency, 1990.
40. Deschenes, E.P., Greenwood, P.W.,and Adams,J. An evaluation of the Nokomis Challenge
Program in Michigan.Journalof Contemporary (May 1993) 9,2:146-67.
CriminalJustice
41. Cronin, R.C., Bootcampsfor adult andjuvenileoffenders:
Overviewand update.Washington, DC:
National Institute ofJustice, 1994.
42. MacKenzie, D.L., and Parent, D. Boot camp prisons for young offenders. In Smartsentencing:
Theemergence of intermediate
sanctions.J.M. Byrne, A.J. Lurigio, andJ. Petersilia, eds. Newbury
Park, CA: Sage, 1992.
43. MacKenzie, D.L., Brame, R., McDowall, D., and Souryal,C. Boot camp prisons and recidivism
in eight states. Criminology
(1995) 33,3:327-57.
44. Deschenes, E.P.,Greenwood, P.W.,and Marshall,G. TheNokomisChallengeProgramevaluation.
Santa Monica, CA: RAND Corporation,July 1995.
45. Davidson, W.S., Redner, R., Amdur, R.L., and Mitchell, C.M. Alternativetreatments
for troubled
youth:Thecaseof diversionfrom thejusticesystem.New York:Plenum Press, 1990.
46. Petersilia,J., and Turner,S. Intensive probation and parole. In Crimeandjustice:A reviewof
Vol. 17. M. Tonry,ed. Chicago: Universityof Chicago Press, 1993, pp. 281-336.
research.
47, Schneider, A. Restitution and recidivism rates ofjuvenile offenders: Results from four experi-
mental studies. Criminology(1986) 24,3:142.
48. McCord,J. A thirty-yearfollow-up of treatment effects. AmericanPsychologist
(1978) 33:284-89.

This content downloaded from 67.66.218.73 on Mon, 14 Dec 2015 00:19:42 UTC
All use subject to JSTOR Terms and Conditions