ContentslistsavailableatSciVerseScienceDirect
Accident
Analysis
and
Prevention
jo u r n al h om ep ag e :w w w . e l s e v i e r . c o m / l o c a t e / a a p
Effects
of
admission
and
treatment
strategies
of
DWI
courts
on
offender
outcomes
Frank
A.
Sloan
a,∗, Lindsey
M.
Chepke
a, Dontrell
V.
Davis
a, Kofi
Acquah
a,b,
Phyllis
Zold-Kilbourn
caDepartmentofEconomics,DukeUniversity,POBox90097,Durham,NC27708,UnitedStates bBrownUniversity,69BrownStreet,EconomicsBoxB,Providence,RI02912,UnitedStates cBuyTheNumbers,Inc.,6151ColdSpringsTrail,GrandBlanc,MI48439,UnitedStates
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:
Received31October2011 Receivedinrevisedform 19December2012 Accepted22December2012 Keywords: Alcohol-impaireddriving Drunkdriving Courtintervention DWIoffenders Motorvehiclecrashes Treatment
a
b
s
t
r
a
c
t
Purpose:ThepurposeofthisstudyistoclassifyDWIcourtsonthebasisofthemixofdifficultcases participatinginthecourt(casemixseverity)andtheamountofinvolvementbetweenthecourtand participant(serviceintensity).Usingourclassificationtypology,weassessedhowcasemixseverityand serviceintensityareassociatedwithprogramoutcomes.Weexpectedthatholdingotherfactorsconstant, greaterserviceintensitywouldimproveprogramoutcomeswhilearelativelyseverecasemixwouldresult inworseprogramoutcomes.
Methods:Thestudyuseddatafrom8DWIcourts,7fromMichiganand1fromNorthCarolina.Usinga 2-wayclassificationsystembasedoncourtcasemixseverityandprogramintensity,weselectedparticipants in1ofthecourts,andalternatively2courtsasreferencegroups.Referencegroupcourtshadrelatively severecasemixesandhighserviceintensity.Weusedpropensityscorematchingtomatchparticipants intheothercourtstoparticipantsinthereferencegroupcourtprograms.Programoutcomemeasures weretheprobabilitiesofparticipants’:failingtocompletethecourt’sprogram;increasingeducational attainment;participantsimprovingemploymentfromtimeofprogramenrollment;andre-arrest. Results:Formostoutcomes,ourmainfindingwasthathigherserviceintensityisassociatedwithbetter outcomesforcourtparticipants,asanticipated,butacourt’scasemixseveritywasunrelatedtostudy outcomes.
Conclusions:Ourresultsimplythatdevotingmoreresourcestoincreasingdurationoftreatmentis pro-ductiveintermsofbetteroutcomes,irrespectiveofthemixofparticipantsinthecourt’sprogram. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction
Externalitiesfromdrinkinganddrivingbehaviorsarewell
doc-umented.Whiletherehasbeenadramaticdecreaseinthenumber
of alcohol related fatal crashes over the last decade, the total
numberof fatalcrashesoverallhasdecreased aswell(National
HighwayTrafficSafetyAdministration,2009).Nationally,the
per-centageofalcohol-impaireddrivingfatalitieshasremainedat32%;
between2000and2009,thepercentofalcohol-impaired
passen-gervehicledriversinvolvedinfatalcrashesremainedpractically
unchanged(NationalHighwayTrafficSafetyAdministration,2009,
2011).Giventheexternalitiesfromrecklessdriving,governments
haveenactedandenforcedlawstopromotesafedriving,invested
inroadways,promulgatedregulationstopromotevehiclesafety,
controlledentryofalcoholsellers,andotherlawsdirectlyaimed
∗ Correspondingauthor.Tel.:+19196139358;fax:+19196817984. E-mailaddress:[email protected](F.A.Sloan).
atdiscouragingdrivingwhileintoxicated(DWI).1Amorerecent,
thoughcertainlynotnew,policyapproachisimplementationof
DWItreatmentcourts.
Modeledontheconceptofdrugcourts,DWIcourtshavebeen
establishedthroughouttheU.S.tointegratepenaltiesforDWI
vio-lationswithtreatmentforunderlyingalcoholaddiction.Arationale
forDWIcourtsisthattheconventionaltrafficcourtmodelof
pros-ecutingDWIoffendersfailstoaddresstheaddictioncomponent
ofDWIcases.Because treatmentcourtsutilize avarietyof
ser-vices,manyof thesecourtsaretailoredtodealwithmorethan
onedimensionofDWIbehavior.
TherehavebeenseveralevaluationsofDWIandhybridcourts’
effectiveness(Eibneretal.,2006;MacDonaldetal.,2007;Moore
etal.,2008;Bouffardetal.,2010;BouffardandBouffard,2011).
In general, the outcome measure is re-arrest for DWI or for
anyoffense. Based in part onunpublishedstudies, the general
1Foreaseofreading,werefertodrinkinganddrivinggenerallyasDWI,regardless
ofwhatanindividualstatecallsthisoffense.Thistermvariesbystateandjurisdiction andisalsocallede.g.,DUI,OWI,OUI,OMVI,DUIL,DWAI,orDWUI.
0001-4575/$–seefrontmatter © 2012 Elsevier Ltd. All rights reserved.
conclusionfromexistingevidenceisthatDWI/hybridcourtsdeter
repeatoffenses(HuddlestonandMarlowe,2011).Nevertheless,not
allfindingsareentirelypositive.Bouffardetal.(2010)foundthat
DWIcourtsareineffectiveinpreventingrecidivismforindividuals
previouslychargedwithaDWIbut areeffectivefor individuals
withnon-DWIoffenses.Inastudyonthecosteffectivenessofa
CaliforniaDWIcourt,Eibneretal.(2006)concludedthattheDWI
courtwaslesscostlythantraditionalcourtsforthird-time
offen-ders,butmorecostlyforsecondtimeoffenders,whichimpliesthat
theDWIcourtwascomparativelymoreeffectiveindealingwith
hard-coreoffenders.Irrespectiveoftheprioroffense record,the
authorsconcludedthattheDWIcourtimprovedoffenderoutcomes
acrossseveralmeasures(e.g.,alcoholproblemsindex,drinksper
day,stressfullifeeventsindex,recidivism).
Manyofthestudiessufferfromimportantmethodological
lim-itations.Somestudieslackacontrolgrouporsufficientevidence
thatthecontrolgroupadequatelymatchestheDWIcourtgroup.
Samplesizesareoftensosmallthatthestudiesareunderpowered.
Thevastmajorityofstudiesarebasedonlocalizedsamples.Also,
positiveresultsmaybemuchmorelikelytobepublishedormore
generally,tobepublicized(publicationbias).
Likethepreviousstudies,weconductedthisstudytoassessthe
effectivenessofDWIcourtswithvaryingcasemixesandservice
intensitiesonprogramoutcomes.Wedefinedcasemixseverityas
themix ofdifficult casesparticipatingin thecourt and service
intensityastheamountof involvementbetweenthecourt and
participant. We hypothesized that for a given level of service
intensity,courtswitha mixof participantsfor which favorable
programoutcomesareinherentlymoredifficulttoachieve(higher
casemixseverity)wouldinfacthavepooreroutcomessince
cop-ingwithamoreseverecasemixwouldplacehigherdemandson
afixedquantityofprogramresources.Otherfactorsbeingequal,
wehypothesizedthatagreateramountofservicespercourt
par-ticipant(higherserviceintensity)wouldleadtobetteroutcomes.
Inotherwords,wetestedforwhetherornotagreater
expendi-tureofresourcesinfactproducedbetterprogramoutcomes.Our
studyimprovedonpreviousstudiesspecificallybyhavingan
ade-quatecontrolgroupandsamplesize,andinusingpropensityscore
matching(PSM)tocontrolfortheheterogeneityofindividualDWI
treatmentcourts.
Inthenextsections,weprovideanoverviewofourstatistical
methods,describethecourtsinoursample,describeourstatistical
methodsindetail,andthenpresentareviewofourfindings.Finally,
wediscussourfindingsinthecontextofwhatisknownaboutDWI
andhybridcourtsaswellasourconclusions.
2. Methods
2.1. Overview
Weclassifiedcourtsbasedontheircasemixseverityandservice
intensity.Casemixseveritywasdeterminedbyseveral
character-isticsofcourtparticipants,mentalhealthhistory,useofschedule
1or2controlledsubstances,orpriorsubstanceabusetreatment.
Serviceintensitywasbased ontherequirementsofeach court,
e.g., daysof treatmentrequired or useof sanctions and
incen-tives. Beforematching, weclassified courtsonthe basis ofthe
mixofdifficultcasesparticipatinginthecourt(casemixseverity)
andtheamountofinvolvementbetweenthecourtandparticipant
(serviceintensity).Thenaftermatching, wecompared matched
groupsonthebasisoftheirprogramoutcomes.Weuse
propen-sityscorematching (PSM)toaccountfor thedifferencesin our
sample.PSMinvolvesthefollowingsteps.First,logitanalysisis
conductedtoassesscorrelatesofapersonbeinginthetreatment
referenceversusthecontrolgroup.Inourstudy,sinceparticipants
were“treated”inallsamplecourts,werefertothewhatis
typ-icallycalledthe“treatment”groupasthe“reference”groupand
thecontrolgroupisreferredtoasthecomparisoncourtsample.
Second,usingtheparameterestimatesfromthelogitregression,a
predictedprobabilityofbeingaparticipantinthereferencecourt
sampleiscalculated,bothforparticipantsinthereferencecourt
programand(counterfactually)inthecomparisoncourtprogram.
Thesepredictedprobabilitiesareusedformatching.Thismethod
ofPSMpairsreferencecourtparticipantswithcomparisoncourt
participantswhosepropensityscores(probabilitiesofbeinga
par-ticipantinthereferencecourtprogram)differuptoapre-specified
amount(thisisknownasthecaliperwidth)(Austin,2011).Weused
onetoonenearestneighbormatchingwithoutreplacementusing
a0.05 caliper.Finally,oncethegroupsarematched,anaverage
treatmenteffect(ATT)iscalculatedforeachmatchedpairof
partic-ipants.TheATTcomparesoutcomesofparticipantsinthereference
court(s)withthoseinthecomparisoncourtsamples.
2.2. Data
DatacamefromsevenDWIcourtsinMichiganandoneDWI
courtinNorthCarolina.Weincludedtwostatesinouranalysisto
provideresultsthatarenotuniquetotheidiosyncrasiesofan
indi-vidualstate’sDWIcourtprogram.Michiganwaschosenfortwo
mainreasons;it hasawell-establishedDWIcourtprogramand
hasprogramsinvariousgeographicareasofthestate.Asof2010,
Michiganhad24designatedDWIcourts(MichiganDrugTreatment
Courts,2010).IndividualMichiganDWIcourtswereselectedbased
ontheirpriorparticipationinDWIresearchandwillingnesstoshare
theirdata.NorthCarolinawaschosenbecauseofthequalityand
availabilityoftheircourtdata.Atthetimethisstudywasconducted,
North Carolinahadoneoperating DWIcourt.2 Thedatausedin
ouranalysisspanned2004to2010forMichiganand2000to2010
forNorthCarolina.Intotal,oureight-courtsamplecontained3844
observations.Analysisofindividualcourtpairsdifferedinnumbers
ofobservationsduetomissingvalues.
BothMichiganandNorthCarolinamaintainDWIcourtdataat
thecourtlevelwhilegeneralcriminalcourtdataaremaintained
atthestatelevel. NorthCarolina criminalcourt datawere
pro-videdfortheentirestateandincludesDWIarrestdataforall100
counties. Duetorestrictionsin Michigan datapolicies,criminal
courtdatawereonlyobtainedforthesevencountieswithwhich
wehadexistingdatauseagreementswithDWIcourts.Asaresult,
ourrecidivismanalysisislimitedtoarrestsoccurringintheseven
Michigancountiesusedinourstudy.Thislimitationismitigatedby
thefactthatmostpersonsarearrestedforDWIintheircountyof
residence;aminorityofrearrestswouldoccurincountiesoutside
ofwheretheinitialarrestoccurred.Tosupportthis,inaseparate
analysisofNorthCarolinacriminalcourtdatafor2001–2011from
all100countiesinthestatewefoundthat70percentofarrestsfor
DWIoccurredinthearrestee’scountyofresidence.Thisprovidesa
roughestimateoftheshareoftotalDWIarrestswecouldmeasure
inMichigan.
The court data contained information on substance abuse
andmentalhealthhistoryatentry,informationondemographic
characteristicsofprogramparticipants,andseveralmeasuresof
outcomes, including whetheror not theparticipant completed
theprogram,andwhetherornottherewerechangesin
educa-tional attainment and in employment status between thedate
ofentryandthetimetheindividualcompletedtheprogram.We
2Asof2012,TheMecklenburgSTEPcourtwastheonlyactiveDWIcourtinNorth
Carolina.However,thiscourtisdividedintotwoseparatecourtswithonecourt beingprimarilySpanishspeaking.Forpurposesofouranalysis,wehavetreatedthe courtsasasinglecourt.
Fig.1. Classifyingcourts.Note:thetwoaxesdenotetherespectivemeans.
alsoobtainedinformationonDWIarrestsandconvictionsinthe
countiesinwhichtheDWIcourtswerelocatedinMichiganand
fortheentirestateofNorthCarolina.Inaccordancewithourdata
agreementwiththecourts,eachcourtwasassigned alabel
cor-respondingtoitspositionin termsof serviceintensityand the
severityofcasesitaccepted(seeFig.1).Themeanvaluesofthe
casemixandintensityindexeswere0.40and0.00,respectively,the
latterhavingbeennormalizedtozero.Thecasemixindexvaried
from0.25to0.45;theintensityindexvariedfrom−0.5to1.5.
Thedataformats and contentwerethesame forthe
Michi-gancourtsinthesample.Whiletheformatandcontentdiffered
betweenNorthCarolinaandMichigan,thevariablesusedinour
empiricalanalysisweregenerally comparable betweenthetwo
states.However, there werea few differences in databetween
thestates.Forexample,theNorthCarolinacourtdidnotobtain
informationonchangesineducationalattainmentorrecidivism.
CourtsinMichiganobtaineddataonrecidivism,butseveralcourts
reportedratesofre-arrestforanyoffenseincludingDWIwellbelow
1%.Foratleast1DWIcourt,thelowre-arrestratereflectedashort
follow-upperiod.Insomecases,suchratescouldreflectpoor
ascer-tainmentofrearrests,andforthisreasonwedidnotanalyzedata
withre-arrestratesfromMichigancourtsthatreportedratesbelow
1%.
EnrollmentcriteriaforDWIoffendersvariedbyDWIcourt,
how-ever,thereweresimilarrequirementsacrosscourts.Participants
must:bedeemedtohaveasubstancedependency,aDWIcharge,
residenceinthecountywherethecourtislocated,andnoprior
vio-lentcrimesorfelonies.3Allcasesinthecourtdatawereincludedin
ouranalysis,withtheexceptionsnotedbelowintheresultssection.
Additionally,whilesomecourtshadmorecases,thisisreflectiveof
thesizeoftheDWItreatmentcourt,notouranalysis.
OursampleofparticipantsinMichiganDWIcourtswas
repre-sentativeofDWIcourtsinMichiganasof2009–2010intermsof
importantpublishedattributes.InMichigan,60%ofparticipants
completedthe program statewide;compared witha 61%
com-pletionrateinoursample(comparisonofdatainMichiganDrug
TreatmentCourts,2010withourdata). In boththeuniverse of
Michigan DWI courtsand in our sample, 9% of participants at
entrywereblack(Table1); oursamplecontained8% Hispanics
versus6%for DWIcourts inMichigan overall. Personsof other
race/ethnicitywere3%inbothsamples.Oursampleconsistedof
3 Theeligibilitycriteriaweregatheredfromparticipanthandbooksfromthe
indi-vidualDWIcourts.
24%femalesversus27%forMichiganDWIcourtsoverall.Inboth
samples,medianeducationalattainmentatentrywas12years.25%
ofoursamplewasnotemployed(eithernotinthelaborforceor
unemployedatentrytotheprogram);comparedwitharateof32%
notemployedforMichiganDWIcourtsoverall.
2.3. Classifyingcourts
Thecourtsdifferedin both thetypesofcasestheyaccepted
andintheintensityoftreatment.Wedevelopedacourt-specific
casemixindexasfollows.First,wespecifiedand,usingdatafromall
courtparticipantsintheentiresample,weestimatedanequation
usinglogitanalysiswiththeindividualprogramparticipantasthe
observationalunit.Thedependentvariablewasabinaryvariableset
equalto1ifthepersondidnotcompletetheprogramandwassetto
0otherwise.Explanatoryvariablesmeasuredsubstanceabuseand
mentalhealthhistoryatentryintotheprogramanddemographic
characteristics.Mentalhealthhistorywasdefinedasabinary
vari-ablethatmeasuredwhetheranindividualhadamedicaldiagnosis
fromtheDiagnosticandStatisticalManualofMentalDisordersIV
(DSM-IV)beforeentryintothecourt.Substanceabusewasdefined
bythecourtaswhethertheindividual:hadanypriorsubstance
abusetreatment;hadanypriorexperiencewithintravenousdrugs;
useddrugsoralcoholbeforeage16;theperson’smainaddiction
wastoalcohol,eitherreportedassuchorinferredfromaDSM-IV
code;thepersonusedSchedule1(e.g.,heroin,LSD,marijuana)or
Schedule2(opium,morphine,cocaine)drugs.Weclassifieddrug
usersintoSchedule1or2basedonwhichtypeofdrugtheperson
wasmostdependent.
Thedemographiccharacteristicswere:femalegender;ageat
entry; race/ethnicity —white (omitted reference group); black;
otherrace;Hispanic;martialstatus—marriedcurrently,other
mar-italstatus (omittedreferencegroup);educationalattainment—a
countvariablefor<highschool;highschoolorequivalent;
com-munitycollegeortradeschool;collegegraduate;postgraduate;
advanced degree.Finally,we includeda binaryvariablefor not
employedatcourtprogramentry(employedomittedgroup).With
thepredictedvaluesforeachperson,wecalculatedthemean
prob-abilityforeachcourt.Thisyieldedacasemixindexforeachcourt.
Theindexvaluesvariedfrom0.30(mostfavorablecasemix)to0.44
(leastfavorablecasemix).
Tomeasureserviceintensity,we usedfactoranalysiswitha
Varimax rotation of these variables: numbers of days in a
12-stepprogram(a12-stepprograminvolvesnon-medicalsupportive
treatment);numberofdaysinaDWIorhybridcourt,andnumbers
ofdrugtests,scheduleddrugcourtreviews,sanctions,and
incen-tives.DWIcourtsalsouseformaltreatment,however,notallcourts
recordthisinformation,soasaproxy,weuseddaysin12-step
pro-gram.Whilenotaformalinpatientoroutpatientformoftreatment;
12-stepprograms,suchasAlcoholicsAnonymous(AA),havebeen
showntohavesimilartreatmentoutcomesasformaltreatment
programs(Timkoetal.,2000;MoosandMoos,2006;Kellyetal.,
2009).Incentiveswereofferedasacarrottoinducedesired
behav-ior,includingareductioninthepenalty.Weusedloadingsfromthe
firstfactorforourindexofserviceintensity.Thisfactoraccounted
for40.7%ofthevariationinthemeasuresandwaspositivelyrelated
toeachindividualmeasureintensity—mostcloselyrelatedtothe
numberofscheduleddrugcourtreviewsandleastrelatedtothe
numberofdaysina12-stepprogram.
2.4. Selectingreferencegroups
Weselectedareferencecourtorcourtstowhichwematched
casemixesofindividualcourts.Thereferencecourtorcourtswas
thetreatmentgrouptowhich wecompared severalalternative
Afterobtainingcasemixandserviceintensityindexesforeach
court,weplottedcourtcasemixindexesagainstcourtservice
inten-sityindexes(Fig.1).Asrequiredbythetermsofourdataagreement
withthecourts,individualcourtnameshavebeenreplacedwith
numberscorrespondingtotheirplacementinFig.1.Horizontaland
verticallinescorrespondtomeanvaluesofcasemixandservice
intensity,whichsplittheareaintoquadrantswithquadrantII
con-taining2courts,quadrantIVwith1,quadrantIwith2,1onthe
border—groupedinIduetoitsproximitytotheothercourtinI,and
quadrantIIIwith3courts,respectively.QuadrantIcontainscourts
withrelatively severecasemixes, and lowerservice intensities;
QuadrantII,relativelyseverecasemixesandhigherservice
inten-sities;QuadrantIIIhascourtswithrelativelylessseverecasemixes
andlowerserviceintensities;andQuadrantIVhascourtswith
rel-ativelylessseverecasemixesandgreaterserviceintensities.
Weexpectedgreaterserviceintensitywouldimprovecourt
out-comes.Ontheotherhand,amorecomplexcasemixmayrequire
moreresourcestoachievethesameoutcome,evenaftermatching
onindividualcharacteristics.
2.5. Propensityscorematching
Theroleoftheaboveclassificationschemewastodecideon
whichcourtstomatch.Wenowdescribehowweimplemented
PSM.Ratherthanmatchontheindexesforcasemixseverityand
serviceintensity,whichonlyaffectedthechoiceofwhichcourtsto
match,onaperparticipantbasis,wematchedthecharacteristics
ofparticipantsinacomparisoncourtorcourtswithcharacteristics
of participantsina reference court orcourts.4 The
characteris-ticsconsistedofthesamevariablesasinthelogitanalysisused
topredictthecasemixseverityindex foreach court. Giventhe
PSMmethodweused,thenumberofobservationsinthe
result-ingmatchedsampleswerelessthantheoriginalcourt samples,
becauseofmissingvaluesorbecausethematchdidnotsatisfythe
abovecalipercriterion.Sincethereferencecourtshadarelatively
severecasemix,thematchingprocesstendedtofindmatchesfor
moreseverecasesinthecomparisoncourtsamples.
Thefinalresultwasanaveragetreatmenteffectonthetreated
(ATT),whichcomparesoutcomesofparticipantsinthereference
court(s) withthose in the comparison court samples. The ATT
effectwascalculatedasthedifferencebetweenthevalueforan
outcome for a participant in a referencecourt(s) program and
thecorrespondingvalueforthematchedparticipantinthe
com-parisoncourt(s)program.WecomputedATTsfortheseoutcome
probabilities: program completion; improved employment
sta-tus; improvededucationalattainment; and anotherDWI arrest
within:oneyear,twoyears,andthreeyearsfollowingadmission
tothecourtprogram.Improvementsinemploymentand
educa-tionalattainmentweremeasuredusingstatusattimeofentryinto
theDWIcourtprogram versusstatuspostprogram.Tomeasure
improvementinemploymentstatus,werankedemployment
sta-tusinthreecategoriesindecreasingorder:employedover30hper
week;employedlessthan30haweek;andunemployed.
Individ-ualsidentifiedasnotinlaborforceatentrywereexcludedfrom
theanalysisduetotheinabilitytoimproveemployment.Ifthere
wasanyimprovement in employment status, a binaryvariable
forimprovedemploymentstatuswassetequalto1.Educational
attainmentwasclassifiedintosixcategoriesrangingfromunder12
yearstoagraduatedegree.Wesetabinaryvariableforimproved
educationalattainmentequalto1ifthepersonwasinahigher
educationalattainmentcategoryatgraduationthanatadmission
4GenerallyinPSM,matchingisbasedoncharacteristicsofthetreatmentgroup,
herethereferencegroup.However,sincewewantedtocompareacontrolgroupto severalreferencegroups,wematchedoncharacteristicsofthecontrolgroup.
Table1
Meanvaluesofindividualcharacteristics.
Variablename Observations Mean Std.Dev.
Outcomes
Failedtocompleteprogram 2,967 0.39 0.49
Improvedemploymentstatusafterprogram 2,701 0.18 0.38 Improvededucationalstatusafterprogram 2,566 0.11 0.32 Substanceabuseandhealthhistory
Mentalhealthhistory 3,844 0.17 0.37
Drughistory–schedule1 3,844 0.13 0.33
Drughistory– schedule2 3,844 0.03 0.18
Drughistory– alcohol 3,844 0.86 0.35
Pre-16addiction 3,844 0.35 0.48
AnyIVdrugexperience 3,844 0.03 0.18
Priorsubstanceabusetreatment 3,844 0.51 0.50 Demographiccharacteristicsandemploymentstatus
Female 3,844 0.24 0.43 Age 3,841 32.82 11.84 Married 3,844 0.17 0.37 Hispanic 3,844 0.08 0.28 Black 3,844 0.09 0.29 Otherrace 3,844 0.03 0.17
Educationalattainment(index) 3,844 2.03 0.94
Notemployed 3,844 0.25 0.44
Explanatoryvariablesformissingvaluesnotshown. Educationalattainmentvariesfrom1–6.
tothecourtprogram.Wecomputedre-arrestratesforanycharge
for1yearand2yearfollowingthedateofprogramentryforthose
courtsforwhichthedataappearedtobereliable.
Weperformedthreesetsofmatches.Thefirstinvolved
com-parisonsofthereferencecourtwitheachoftheothercourts.The
referencecourt(II-B)hadthemostseverecasemixandthehighest
serviceintensity.Inasecondmatch,wegroupedcourtsby
quad-rant.ThecomparisonsforstudyoutcomesbetweenII-Bandcourt
participantsincourtprogramsineachoftheother3quadrantsis
illustratedinFig.1.Inathirdgroupofcomparisons,participants
inbothcourtprogramsinquadrantII(moreseverecasemixand
greaterserviceintensity)werecomparedwithparticipantsinthe
programsofcourtsineach oftheotherquadrants.Thefirstset
hastheadvantageofnotaggregatingparticipantsfromdifferent
courtswithinaquadrant,whichhavedifferentpoliciesand
prac-tices.However,samplesizesofparticipantsinthefirstsetareoften
muchlowerthaninthesecondandthirdsets.
3. Results
Overall,18%ofparticipantsimprovedtheiremploymentstatus
afterparticipatingintheDWIcourtprogramwhile39%of
partic-ipantsfailedtocompletetheirprograms(Table1).Whilethevast
majorityofparticipantshadhistoriesofalcoholuse,aminorityhad
historiesofillicitdruguse.Halfofparticipantshadreceived
sub-stanceabusetreatmentpriortoentryintotheprogram.Fewerthan
afifth(17%)hadahistoryofmentalillness.Themeanparticipant
hadahighschooldiplomaorequivalent,butafewparticipantshad
graduatedegrees.25%werenotemployedatprogramentry.
On average, sample persons spent 362 days in DWI courts
(Table2).Themediannumber of dayswasalmostthesameas
themeanvalue,364days.However,whilethemeannumberof
daysina12-stepprogramwas129,overhalfofthesampledid
notparticipateina12-stepprogram.Therewasalsosubstantial
variabilityinnumbersofdrugtestswhileenrolledandinuseof
sanctionsandincentives.Afewpersonswereintreatmentafter
theywerenolongerofficiallyenrolledinthecourtprogram.Some
personsremainedinthecourtprogramandwereundertreatment
foryears.Themaximumvaluesof893and897fordaysincourtand
ina12-stepprogramaremaximumsbutnotoutliers.Individuals
Table2
Frequencydistributionofprogramvariables.
DaysinDWIcourt Daysin12stepprogram No.ofscheduledDWIcourtreviews No.ofdrugtests No.ofsanctions No.ofincentives Precentile 1 13 0 0 0 0 0 25 200 0 4 28 0 0 50 364 0 9 105 1 1 75 505 189 19 224 3 3 100 893 897 116 828 30 23 Mean 362.40 128.78 13.35 142.95 1.84 2.02 Std.Dev. 197.84 208.05 13.26 138.77 2.42 2.83 Observations 3,844 3,844 3,839 3,838 3,814 3,814
wereeliminatedfromtheanalysis,asthemaximumamountoftime
allowedinaDWIcourtis2.5years(e.g.,maximumvaluesof2177
fordaysincourtinouroriginalsampleand4682daysina12-step
programwerenotacceptablevalues).
We performed logit analysis on whether or not the
par-ticipant completed the program with and without court fixed
effects (Table3). Withfew exceptions, e.g., theodds ratios for
race/ethnicity,addingthecourtfixedeffectshadlittleeffectonthe
estimatedoddsratiosandassociatedconfidenceintervals.Thus,we
discusstheresultsforthespecificationthatexcludescourtfixed
effects.Parameterestimatesfromthisspecificationwereusedto
constructourcourt-specificcasemixindex.
Factorsassociatedwithahigherprobabilityofprogramfailure
includedhavingamentalhealthhistory(oddsratio(OR)=1.39;95%
confidenceinterval(CI):1.12–1.73),havingbeenaschedule2drug
user(OR=1.83;95%CI:1.12–3.00),andhavingstartedalcoholuse
beforeage16(OR=1.26;95%CI1.04–1.52).Holdingotherfactors
Table3
Failuretocompletedrugcourtprogram(Logitanalysis).
Variables Fixedeffects?
No Yes
Substanceabuseandhealthhistory
Mentalhealthhistory 1.39 1.43
(1.12–1.73) (1.14–1.79)
Drughistory–schedule1 1.19 1.22
(0.87–1.63) (0.88–1.69)
Drughistory–schedule2 1.83 1.97
(1.12–3.00) (1.19–3.25)
Drughistory–alcohol 0.75 0.73
(0.52–1.08) (0.50–1.06)
Pre-16addiction 1.26 1.36
(1.04–1.52) (1.12–1.65)
AnyIVdrugexperience 0.94 0.88
(0.60–1.47) (0.56–1.39)
Priorsubstanceabusetreatment 1.10 1.10
(0.93–1.30) (0.92–1.30) Demographiccharactersticsandemploymentstatus
Female 0.79 0.76 (0.65–0.95) (0.62–0.92) Age 0.98 0.98 (0.98–0.99) (0.98–0.99) Married 0.75 0.74 (0.59–0.96) (0.58–0.95) Hispanic 1.15 1.69 (0.84–1.58) (1.20–2.37) Black 1.16 1.59 (0.86–1.55) (1.19–2.24) Otherrace 1.01 1.18 (0.61–1.68) (0.70–1.97) Unemployed 1.88 1.97 (1.57–2.26) (1.64–2.38)
Educationalattainment(index) 0.66 0.65
(0.59–0.73) (0.58–0.73)
Observations 2,967 2,967
95%confidenceintervalsinparentheses.
Explanatoryvariablesformissingvaluesnotshown.
constant,priorsubstanceabusetreatmentwasunrelatedto
fail-uretocompletetheprogram.Withtheothercovariatesincluded,
race/ethnicitywasunrelatedtoprogramcompletion.(However,in
thespecificationwithcourtfixedeffectsincluded,beingHispanicor
blackwasassociatedwithahigherprobabilityoffailingtocomplete
theprogram.)Factorsassociatedwithalowerprobabilityof
pro-gramfailureincludedbeingfemale(OR=0.79;95%CI:0.65–0.95),
older(OR=0.98;95%CI:0.98–0.99foreachyearofage),currently
married(OR=0.75;95%CI:0.59–0.96),andhavingahigher
edu-cationalattainment(OR=0.66:95%CI0.59–0.73,a34%reduction
intheoddsofprogramfailureforeacheducationalcategory(1–6)
attained).
Beforematching, there weresubstantial differencesin
char-acteristics of participants by court. Even after matching, some
differencesincharacteristicsremained.However,standardized
dif-ferencesexceededthe10%thresholdinonlyaminorityofcases
(Table 4).Matcheswerebetterfor courtslocatedinQuadrantI
comparedtoQuadrantII(moreseverecasemixandlowerservice
intensitycomparedwithcourtswithrelativelyseverecasemixes
withagreaterserviceintensity)andQuadrantIII—IIcomparisons
(lessseverecasemixandlowerserviceintensitycompared with
moreseverecasemixwithagreaterserviceintensity)thanthey
werefortheQuandrantIV—QuadrantIIcomparisons(lesssevere
casemixandgreaterserviceintensitycomparedwithmoresevere
casemixwitha greaterservice intensity).Court I-Bwasonthe
boundaryofQuadrantsIandIII,butweplaceditinQuadrantIfor
purposesofouranalysis.
Priortomatching,13%ofparticipantsinCourtsII-AandII-Bhad
amentalhistoryascomparedtocourtswithlowerservice
inten-sity(QuadrantI)—CourtsI-AandI-B.20%ofparticipantsinthelatter
courtshadamentalhealthhistory.ThetotalsampleforCourtsII-B
andII-Acombinedwas819andforCourtsI-BandI-A,thecombined
samplewas1574.Aftermatching,16%ofparticipantsfromCourts
II-BandII-Ahadamentalhealthhistorywhile14%fromCourtsI-B
andI-Adid.Thestandardizeddifferenceformentalhealthhistory
fellfrom19.1%to3.56%,thelatterbeingwellwithinusualcriterion
foragoodmatch,astandardizeddifferencepercentageofunder10%
inabsolutevalue.Themeanpercentagewithmentalhealth
histo-riesforCourtsII-BandII-Aincreasedfrom13%to16%duetothe
matching,whichreducedthesamplefromthesecourtsfrom819
to396.Nomatchwithinthe0.05calipercriteriacouldbefoundfor
theremainingCourtII-BandII-Aparticipants.Therewerealso396
personsfromCourtsI-BandI-A,givenone-to-onematching.For
thecomparisonofthesecourts,nostandardizeddifferencesexceed
10%,althoughbeforematching,standardizeddifferencesabovethis
thresholdsexistedfor almostallthecovariates.Thelargest
dif-ferencewasforpercentblack;beforematchingthestandardized
differencewas−55%.Aftermatching,thedifferencewas−1.15%.
Evenaftermatching,thereweresubstantialdifferencesin
char-acteristicsbetween CourtsII-B, II-A, and IV-A in mental health
history,age,blackrace,andineducationalattainmentatprogram
entry.Onereasonfor thepoorer matchisthattherewere
F.A. Sloan et al. / Accident Analysis and Prevention 53 (2013) 112– 120 117
Standardizeddifferences:pre-andpost-matching.
Courts Pre-match Post-match Pre-match Post-match Pre-match Post-match
I-Bvs. I-A II-Bvs. II-A Std. Diff.(%) I-B– I-A II-B– II-A Std. Diff.(%) IV-A II-B– II-A Std. Diff.(%) IV-A II-B– II-A Std. Diff.(%) III-A–III-B–III-C IIB–IIA Std. Diff.(%) III-A–III-B–III-C II-B– II-A Std. Diff.(%) Substanceabuseandhealthhistory
Mentalhealth history 0.20 0.13 19.09 0.14 0.16 −3.56 0.28 0.13 38.01 0.26 0.22 10.40 0.11 0.13 −6.25 0.14 0.14 0.11 (0.40) (0.33) (0.35) (0.36) (0.45) (0.33) (0.44) (0.42) (0.31) (0.33) (0.35) (0.35) Drughistory– schedule1 0.19 0.084 31.15 0.12 0.13 −2.93 0.064 0.084 −7.74 0.069 0.11 −15.57 0.083 0.084 −0.60 0.10 0.11 −4.54 (0.39) (0.28) (0.35) (0.34) (0.25) (0.28) (0.25) (0.32) (0.28) (0.28) (0.31) (0.32) Drughistory– schedule2 0.038 0.034 2.16 0.028 0.033 −2.88 0.030 0.034 −2.28 0.033 0.041 −4.27 0.023 0.034 −6.64 0.028 0.034 −3.24 (0.19) (0.18) (0.17) (0.18) (0.17) (0.18) (0.18) (0.20) (0.15) (0.18) (0.17) (0.18) Drughistory– alcohol 0.85 0.85 0.00 0.83 0.82 3.28 0.90 0.85 15.09 0.91 0.87 12.86 0.85 0.85 0.00 0.81 0.82 −0.58 (0.36) (0.36) (0.38) (0.39) (0.30) (0.36) (0.29) (0.34) (0.36) (0.36) (0.39) (0.39) Pre-16addiction 0.32 0.50 −37.08 0.44 0.43 2.02 0.45 0.50 −10.19 0.47 0.60 −26.28 0.24 0.50 −55.10 0.38 0.43 −11.07 (0.47) (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) (0.49) (0.43) (0.50) (0.49) (0.50) AnyIVdrug experience 0.038 0.021 10.40 0.023 0.023 0.00 0.080 0.021 27.70 0.045 0.041 1.98 0.024 0.021 2.24 0.025 0.025 0.00 (0.19) (0.14) (0.15) (0.15) (0.27) (0.14) (0.21) (0.20) (0.15) (0.14) (0.16) (0.16)
Priorsubstanceabuse treatment
0.41 0.59 −35.32 0.56 0.60 −7.45 0.73 0.59 30.64 0.67 0.67 1.73 0.50 0.59 −17.64 0.63 0.62 1.73
(0.49) (0.49) (0.50) (0.49) (0.44) (0.49) (0.47) (0.47) (0.50) (0.49) (0.48) (0.49)
Demographiccharacteristicsandemploymentstatus
Female 0.25 0.19 15.20 0.21 0.20 1.25 0.24 0.19 12.96 0.24 0.20 9.79 0.27 0.19 19.76 0.22 0.22 1.05 (0.44) (0.39) (0.41) (0.40) (0.43) (0.39) (0.43) (0.40) (0.45) (0.39) (0.41) (0.41) Age 31.45 33.60 −18.58 31.73 32.55 −7.24 37.10 33.60 32.03 35.67 32.25 31.13 32.70 33.60 −7.95 32.83 32.64 1.76 (12.37) (10.70) (12.01) (10.56) (11.16) (10.70) (11.06) (10.88) (11.82) (10.70) (11.36) (10.64) Hispanic 0.070 0.22 −44.31 0.14 0.14 −2.20 0.0028 0.22 −74.13 0.0041 0.0081 −5.21 0.022 0.22 −64.27 0.050 0.056 −2.67 (0.26) (0.42) (0.34) (0.35) (0.053) (0.42) (0.064) (0.090) (0.15) (0.42) (0.22) (0.23) Black 0.052 0.24 −55.02 0.13 0.13 −1.15 0 0.24 −78.93 0 0.27 −85.59 0.071 0.24 −47.56 0.15 0.14 2.38 (0.22) (0.43) (0.33) (0.34) (0) (0.43) (0) (0.45) (0.26) (0.43) (0.36) (0.35) Other 0.024 0.017 4.57 0.020 0.020 0.00 0.061 0.017 22.72 0.041 0.033 4.27 0.035 0.017 11.40 0.036 0.025 6.50 (0.15) (0.13) (0.14) (0.14) (0.24) (0.13) (0.20) (0.18) (0.18) (0.13) (0.19) (0.16) Married 0.15 0.21 −15.55 0.15 0.17 −6.14 0.22 0.21 2.16 0.22 0.15 16.63 0.14 0.21 −18.36 0.17 0.15 3.95 (0.36) (0.41) (0.36) (0.38) (0.41) (0.41) (0.42) (0.36) (0.35) (0.41) (0.37) (0.36) Unemployed 0.29 0.23 13.08 0.20 0.23 −7.39 0.24 0.23 2.98 0.23 0.29 −13.08 0.23 0.23 0.40 0.23 0.21 4.05 (0.45) (0.42) (0.40) (0.42) (0.43) (0.42) (0.42) (0.45) (0.42) (0.42) (0.42) (0.41) Educational attainment(index) 1.94 1.89 5.12 1.94 1.90 3.59 2.26 1.89 35.98 2.25 2.06 19.58 2.20 1.89 30.95 2.03 1.97 5.54 (0.85) (1.085) (0.84) (1.11) (0.97) (1.085) (0.96) (0.95) (0.91) (1.085) (0.77) (1.12) N 1,574 819 396 396 361 819 246 246 1,090 819 357 357
Explanatoryvariablesformissingvaluesnotshown:mentalhealthhistory,drughistory–schedule1,drughistory–schedule2,drughistory–alcohol,pre-16addiction,anyIVdrugexperience,priorsubstanceabusetreatment, female,currentage,hispanic,black,otherrace,educationalattainment,unemployed.
Table5
Programcompletion,improvedemployment,andimprovededucationalattainment(ATTanalysis). PANELA
Court Failedprogram Improvedemployment Improvededucation
Reference (II-B)
Comparisona Differenceb T-Test Reference
(II-B)
Comparison Difference T-Test Reference(II-B) Comparison Difference T-Test
I-B 0.28 0.38 −0.092 −1.58 0.22 0.14 0.08 1.66 0.096 0.048 0.048 1.47 III-C 0.11 0.28 −0.17 −2.22 0.27 0.15 0.12 1.44 0.11 0.17 −0.057 −0.83 IV-A 0.23 0.47 −0.25 −4.38 0.22 0.23 −0.0081 −0.15 0.090 0.082 0.0082 0.23 III-B 0.26 0.38 −0.12 −1.71 0.17 0.29 −0.12 −1.93 0.075 0.25 −0.17 −3.26 III-A 0.27 0.53 −0.25 −4.75 0.23 0.15 0.074 1.62 0.10 0.045 0.058 1.96 I-A 0.29 0.51 −0.22 −4.48 0.20 0.16 0.045 1.10 0.086 0.10 −0.016 −0.53 II-A 0.28 0.29 −0.011 −0.16 0.17 0.28 −0.11 −1.83 – – – – PANELB
Failedprogram Improvedemployment Improvededucation Reference
(II-B/II-A)
Comparison Difference T-Test Reference (II-B/II-A)
Comparison Difference T-Test Reference (II-B/II-A)
Comparison Difference T-Test
I-B–I-A 0.27 0.45 −0.18 −3.72 0.21 0.15 0.055 1.36 0.089 0.12 −0.031 −1.00 III-A–III-B–III-C 0.26 0.51 −0.25 −4.81 0.22 0.16 0.057 1.28 0.10 0.13 −0.030 −0.87
PANELC
Failedprogram Improvedemployment
Reference(II-B/II-A) Comparison Difference T-Test Reference(II-B/II-A) Comparison Difference T-Test I-B–I-A 0.34 0.39 −0.06 −1.62 0.22 0.15 0.07 2.41
IV-A 0.31 0.42 −0.11 −2.44 0.27 0.22 0.06 1.41
III-A–III-B–III-C 0.32 0.42 −0.10 −2.80 0.23 0.20 0.03 0.88 Boldindicatessignificanceat5%levelorhigher
aComparisonisforcourtindicatedintheleftmostcolumn. b Differenceisbetweenthereferenceandcomparisoncourts.
Notshown,inordertoconservespace,arecomparabledataforthe
othermatches.Ingeneral,PSMworksbetterwhentherearemore
observationsonwhichtomatch.
Table5showsATTeffectsandassociatedstatisticalsignificance
levelsforthreesetsofcomparisons:CourtII-B(reference)andall
oftheothercourts(comparison,PanelA);CourtII-BwithII-A
(ref-erence),withcourtsintheotherquadrants(comparison,PanelB);
andCourtII-BwithII-A(reference)combinedwiththecourtsin
theotherquadrants(comparison,PanelC).Thedifferencebetween
panelsBandCisinthecomparisongroups;PanelCaddsCourt
IV-A.Theoverallfindingisthathigherserviceintensity,evenwitha
moreseverecasemix,leadstohigherratesofprogramcompletion.
Morespecifically,comparedtoCourtII-B,participantsinCourts
IV-A,III-A,I-A,andII-Awerelesslikelytocompletetheirprograms
(PanelA).Thispatternsuggeststhatonaverage,themore
demand-ing,service-intensiveprogramofferedbyCourtII-Bdidnotdeter
completion;infact,ifanything,completionrateswerehigherthan
forcourtswithlowerrequirementsforcompletion.
Combiningindividualcourts byquadrant (Table5,Panel B),
CourtII-Bhadstatisticallysignificantlyhigherratesforcompletion
thanallofthecourtgroups.
WithCourtsII-BandII-Aparticipantsasthereferencegroup,
themajorityofATTsforfailedprogramwerestatisticallysignificant
atthe5%levelorhigher.Differencesinfailureprobabilitieswere
−0.11forcourtIV-Aand−0.10forcourtsIII-A,III-B,andIII-C.
Overall, the empirical evidence for relationships between
courtcasemixseverity/serviceintensitywasweakerforimproved
employmentthanitisforcompletion.Theonlystatistically
signif-icantdifferencesinimprovedemploymentareinPanelC.Courts
II-BandII-Ahadgreatersuccessinimprovingemploymentstatus
ofparticipantsthanCourtsI-BandI-A(0.07).Thedifferencesinthe
probabilitiesofimprovedemploymentwerefairlylargerelativeto
thereferencegroupemployment(0.07and0.03).Thedifference
inprobabilityofimprovedemploymentbetweencourtsII-B/II-A
and courtIV-A is large(0.06)butnot statisticallysignificantat
conventionallevels.
The findings on the relationship between casemix
sever-ity/service intensity and improved educational attainment is
mixed.WecouldnotcompareCourtsII-BandI-Awithothercourts
onimprovingeducationalattainmentsinceCourtII-Adidnot
col-lectdataoneducationalattainment.WefoundthatCourtII-Bhad
moresuccessinimprovingeducationalattainmentofits
partici-pantsthanCourtIII-A,whichisconsistentwiththeviewthathigher
serviceintensityleadstobettereducationaloutcomes.However,
CourtIII-BperformedevenbetterinthisdimensionthanCourt
II-B,afindingthatrunscountertoourhypothesisthathigherservice
intensity,holdingotherfactorsconstant,improvesprogram
out-comes.
Although the resultspresented thus far suggest that courts
with higher service intensity tend to achieve better results in
terms of program completion, we found no statistical
differ-enceinrecidivismmeasuredbyratesofre-arrestforanyoffense
witha 2-yearfollow-upfromthedateofadmissiontothe
pro-gram (Table 6). Court II-B, the more service intensive court,
had a less favorable re-arrest rate than Court I-A for a 1-year
follow-up.
4. Discussion
Formostoutcomes,ourmainfindingwasthathigherservice
intensityisassociatedwithbetteroutcomesforcourtparticipants.
Althoughwematchedonparticipantcharacteristics,wefoundthe
proportionof difficultcasesa courthad (highercasemix
sever-ity)wasunrelatedtotheoutcomeswestudied.Thisgoesagainst
Table6
Recidivism(ATTanalysis).
Court Rearrests 1-yearrearrests 2-yearrearrests
Reference Comparisona Differenceb T-Test Reference Comparison Difference T-Test Reference Comparison Difference T-Test
IV-A 0.11 0.12 −0.0057 −0.17 0.080 0.068 0.011 0.41 0.11 0.11 0.0057 0.17 I-A 0.11 0.091 0.024 0.88 0.083 0.051 0.031 1.42 0.11 0.075 0.035 1.38 Boldindicatessignificanceat5%levelorhigher
aComparisonisforcourtindicatedintheleftmostcolumn. bDifferenceisbetweenthereferenceandcomparisoncourts.
haveanadverseeffectonoutcomesbecausethehardertotreat
participantswouldconsumealargershareofcourtresources.
Our measure of service intensity included both duration of
serviceand quantityof servicesprovided per unit oftime. The
empiricalevidenceontreatmentduration,nottakenfromcourt
settings, generally indicates that programs of longer duration
aremoreeffectivethanbriefinterventions(Wutzkeetal.,2002;
Zhangetal.,2003;McKay,2005).However,therearediminishing
marginalreturnstoextendingtreatmentlength(Zhangetal.,2003).
Anotherpossibilityisthatcasemixseveritydiddiffersufficiently
amongcourtsinouranalysissample.
Theresultsonserviceintensityareimportantinimplyingthat
resources devoted to court programs are productive in terms
ofachievingbetteroutcomesatthemargin. Furthermore,these
resultsimply that creating a DWI court withoutthought asto
thestructureordesignofthetreatmentaspectisnotlikelytobe
productive.Researchonaddictioninterventionshasshownthat
adjustingtreatmentservicestoaccommodateanindividuals’
clin-icalassessment resultsin bettertreatmentoutcomes(Marlowe
etal.,2009).In termsof ourcourts,this meansthat withinthe
observedrangeofserviceintensity,−0.5forCourtIIIAto1.0for
CourtIIB, adding resources makes a difference.Futureanalysis
shouldassesswhetherornotthedifferenceinbenefitsisworth
theaddedcost.Assessmentsofbenefitversuscostshouldinclude
amorecomprehensivemeasureofbenefitthancostsavingstothe
criminaljusticesystemfromreducedrecidivism.Thecalculationof
benefitshouldalsoconsiderbenefitsintermsofimproved
produc-tivity,bothinemploymentsettingsandashouseholdmembers.
Courts areat a disadvantagein treating offenders withalcohol
addictionasresearchhasshownthatearlyinterventionmaybe
moreimportant than theintensityofthe treatment(Moos and
Moos,2003).Nevertheless,apolicyofincreasedtreatmenthasbeen
showntodecreasealcohol-relatedmotorvehicledeaths(Freeborn
andMcManus,2010).
Thisstudymeasuredmixofparticipantsatboththelevelofthe
courtforpurposesofclassifyingcourtsbycasemixseverityand
serviceintensityandattheindividualparticipantlevelforpurposes
of comparing outcomes among courts. Even within quadrants,
there wassubstantialvariation in thecharacteristicsof
partici-pantsincourtprograms.Toobtainpropensityscorematches,we
lostaconsiderablenumberofparticipants.Thisisatleastpartly
attributabletotheheterogeneityofcasemixesamongcourt
pro-grams.Intheend,courtswillopposeoutcome-basedcomparisons
untiltheycanhaveconfidenceinthecasemixadjustmentprocess.
Amoreimportantpracticalimpedimenttooutcome-based
com-parisonsislackofadequatedata.Forthesecomparisonstoserve
asabasisofallocatingscarcepublicresources,dataonanumber
ofrelevantoutcomemeasuresshouldbecollected.Forexample,
thisstudywaslimitedinassessingeffectsofcasemixandservice
intensityonrecidivismbecausethedatawereoftensopoor.Courts
shouldbediligentincollectingrecidivismdatafromparticipants,
particularlyinstateswhereacentralstateagencydoesnotmaintain
criminalcourtrecords.
This study has several strengths. First, rather than treat
court programs as homogeneous entities, this study assessed
heterogeneityoftreatmentcourtsaswellashowdifferencesamong
suchcourtsrelatetoparticipantoutcomes.Second,althoughthe
eight courts in our sample is admittedly a small number, we
includedalargernumberofcourtsthaninpreviousstudies.Our
courtscamefromtwostates. Furtherstudiesshouldattemptto
includecourtsfromagreaternumberofstates;pastresearchhas
notcrossedstatelines.Third,althoughthereisroomfor
improve-ment,ourstudyaccountsforcasemixdifferencesinparticipants
amongcourts.
We acknowledgeseverallimitations.First, mostof the
sam-plecamefromonestate.Includingotherstateprogramswould
substantiallyaddtobothheterogeneityofcourtparticipants,e.g.,
accordingtorace/ethnicity,andtoheterogeneityinprogram
phi-losophy and in details of implementation. Second, due todata
limitations,ourstudywaslimitedtoafewoutcomemeasures.In
particular,itwouldbeimportanttomeasuretheimpactofcourt
programsonfuturedrunkdrivingandDWIviolationsand
convic-tions.Third,thecourtswereselectedbecausetheywerewilling
toparticipate.However,even ifallcourtswillingtoparticipate
inourstudywerebetterthanaverage,weassesseddifferencesin
programoutcomesresultingfromdifferencesinserviceintensity
andinclientcharacteristics.Althoughthedifferencingapproach
weusedeliminatedsomepotentialbiasfromtheselectionprocess,
someselectionbiasmayremain.
Subjecttothecaveatsjustnoted,weobtainedempiricalsupport
forthehypothesisthatdevotingmoreresourcestotreatmentis
productiveintermsofbetteroutcomes,butnotforthehypothesis
thatholdingotherfactorsconstantthatamoreseverecasemixis
associatedwithpoorerprogramoutcomes.Howthecourtswith
moreseverecasemixeslearntocopewiththechallengeofaclient
groupmoreresistanttochangeismostcertainlyanimportanttopic
forfurtherstudy.
Acknowledgments
ThispaperwasfundedinpartbyagrantfromtheNational
Insti-tuteofHealth(NIH1R21AA018168-01A1),specifically,theNational
InstituteonAlcoholAbuseandAlcoholism(NIAAA).Wewouldlike
tothankthecourtsinMichiganforsharingtheirdata.
References
Austin,P.C.,2011.Optimalcaliperwidthsforpropensity-scorematchingwhen esti-matingdifferencesinmeansanddifferencesinproportionsinobservational studies.PharmStat10(2),150–161.
Bouffard,J.A.,Bouffard,L.A.,2011.Whatworks(ordoesn’t)inaDUIcourt?An exam-pleofexpeditedcaseprocessing.JournalofCriminalJustice39(4),320–328. Bouffard,J.A.,Richardson,K.A.,Franklin,T.,2010.DrugcourtsforDWIoffenders?
TheeffectivenessoftwohybriddrugcourtsonDWIoffenders.JCrimJust38(1), 25–33.
Eibner,C.,Morral,A.R.,Pacula,R.L.,MacDonald,J.,2006.Isthedrugcourtmodel exportable?Thecost-effectivenessofadriving-under-the-influencecourt. Jour-nalofSubstanceAbuseTreatment31(1),75–85.
Freeborn,B.A.,McManus,B.,2010.Substanceabusetreatmentandmotorvehicle fatalities.SouthEconJ76(4),1032-1032-1048.
Huddleston,W.,Marlowe,D.B.,2011.Paintingthecurrentpicture:anationalreport ondrugcourtsandotherproblem-solvingcourtprogramsintheUnitedStates. NatlDrugCourtInstUS,p,28.
Kelly,J.F.,Magilli,M.,Stout,R.L.,2009.Howdopeoplerecoverfromalcohol depen-dence?Asystematicreviewoftheresearchonmechanismsofbehaviorchange inalcoholicsanonymous.AddictResTheory17(3),236–259.
MacDonald,J.M.,Morral,A.R.,Raymond,B.,Eibner,C.,2007.TheefficacyoftheRio HondoDUIcourt.EvaluationReview31(1),4–23.
Marlowe,D.B.,Festinger,D.S.,Arabia,P.L.,Dugosh,K.L.,Benastutti,K.M.,Croft,J.R., 2009.Adaptiveinterventionsmayoptimizeoutcomesindrugcourts:apilot study.CurrPsychRep11(5),370–376.
McKay,J.R.,2005.Isthereacaseforextendedinterventionsforalcoholanddruguse disorders?Addiction100(11),1594–1610.
MichiganDrugTreatmentCourts,2010AnnualReportandEvaluationSummary. Lansing,MI,MichiganSupremeCourtandtheStateCourtAdministrativeOffice. Moore,K.A.,Harrison,M.,Young,M.S.,Ochshorn,E.,2008.Acognitivetherapy
treat-mentprogramforrepeatDUIoffenders.JCrimJust36(6),539–545.
Moos,R.,Moos,B.,2003.Long-terminfluenceofdurationandintensityoftreatment onpreviouslyuntreatedindividualswithalcoholusedisorders.Addiction98, 325–337.
Moos,R.,Moos,B.,2006.Participationintreatmentandalcoholicsanonymous:a 16-yearfollow-upofinitiallyuntreatedindividuals.JournalofClinicalPsychology 62(6),735–750.
NationalHighwayTrafficSafetyAdministration,2009.2008TrafficSafetyAnnual Assessment—Highlights.TrafficSafetyFacts.U.S.DepartmentofTransportation. NationalHighwayTrafficSafetyAdministration,2011.PassengerVehicles.Traffic
SafetyFacts:2009Data.U.S.DepartmentofTransportation.
Timko, C.,Moos, R.,Finney, J.,Lesar, M.,2000. Long-term outcomesof alco-holusedisorders:comparinguntreatedindividualswiththoseinalcoholics anonymous and formal treatment. Journal of Studies onAlcohol 61 (4), 529–540.
Wutzke,S.,Conigrave,K.,Saunders,J.,Hall,W.,2002.Thelong-termeffectivenessof briefinterventionsforunsafealcoholconsumption:a10-yearfollow-up. Addic-tion97,665–675.
Zhang,Z.,Friedman,P.,Gerstein,D.,2003.Doesretentionmatter?Treatment dura-tionandimprovementindruguse.Addiction98,673–684.
