R

esearch findings have solidly situated unstable

employment and a lack of conventional ambition
as important risk and need factors among offenders
(i.e., linked to an increased likelihood of recidivism
when not effectively addressed). However, many
methodological deficits have been noted in the research
methodology exploring the impact of employment
on offender reintegration. Given these constraints,
we cannot unequivocally claim that employment
interventions systematically reduce recidivism. The
current research was not undertaken to examine
the impact of an employment program in reducing
recidivism. Rather, the intent of this study was to
explore the specific relationship between employment
status and community outcomes for groups of federal
offenders: those employed while on conditional release,
and a matched comparison group of offenders who
were unemployed. Results are presented, and implications
discussed as they relate to future research and
community employment initiatives for offenders.
Background
R
esearchhasidentifiedunstableemploymentand
lackofconventionalambitionasimportantneed
factorsamongoffenders
2
withasmanyas75%
ofoffendersidentifiedwithemploymentneeds
uponentrytofederalinstitutions.
3
Furthermore,
researchershavereportedthereintegrativeeffectof
skilledemployment,orahistoryofemployment
priortoincarceration,foroffendersreleasedtothe
community.
4
Thesefindingsillustratetheimportance
ofassessingfactorsconstruedasemployment
deficits(e.g.,lackofemploymentskills)and
competencies(e.g.,strongemploymenthistorypriorto
incarceration)fortheircontributionstocommunity-
basedoutcomesforoffenders
5
andalsodemonstrate
thepotentialroleofemploymentinterventionin
contributingtosuccessfulcommunityreintegration
foroffenders.Insum,whetheremploymentis
viewedasaprotectivefactororadeficit,empirical
evidencesupportstheroleofemploymentin
contributingtocommunityoutcomesforoffenders.
P
E
R
S
P
E
C
T
I
V
E
S

T
he impact of community-based employment
on offender reintegration
Christa A. Gillis and Mark Nafekh
1
Performance Assurance, Correctional Service of Canada
Employment programming and outcome
Methodologicalweaknesseshavebeennotedby
numerousresearchersattemptingtoreviewthe
employmentliterature,
6
includingdefinitionalissues
(e.g.,definingvariablesinadichotomousmanner),
whichoverlookimportantfactorssuchasquality
ofparticipation,lengthoftimeintheprogram,and
reasonsforattrition.Additionally,manyprogram
evaluationsfailtoreportimportantinformation
pertainingtooffenderemploymentneedsand
competenciespriortoprogramparticipation.
Moreover,theissueofco-morbidityinoffender
needs,suchasthecombinationofemploymentand
substanceabuseneeds,isimportanttoconsiderfor
itspotentialimpactonworkperformanceand
treatmentgain.
Thelimitationsinresearchmethodologydesigned
toexploretheefficacyofemploymentinterventions
incontributingtoreducedrecidivismwereaptly
andsuccinctlysummarizedbyRyan:
7
problemsin
researchmethodologyandprogramdevelopment,
includingcomparabilityofexperimentaland
controlgroups,selectionofparticipants,tracking
ofex-offenders,differentiationbetweenstructural
andsubculturalvariables,anddefinitionofjob
retention.Acomprehensiveevaluationof
employmentprogrameffectivenessmustthus
consideravarietyoffactorsthatmaymoderatethe
impactoftheprogramonthecriterionofinterest
(e.g.,jobattainmentandretention,successful
communityperformance).
Findingsregardingtheimpactofemployment
traininghavebeenequivocal,
8
withsomestudies
reportingpositiveeffectsofemploymenton
recidivism,andothersreportinglimitedorno
effects.Somereviewers,basedonaqualitative
analysisoftheliterature,haveadoptedafairly
optimisticoutlookontheimpactofemployment
trainingonrecidivism.
9
PearsonandLipton,
10
in
theirmeta-analyticreviewofeducationaland
vocationalprograms,stated:Althoughsome
10
typesofeducationalandvocationalprograms
appearpromising intermsofreducingrecidivism,
duetoalackofstudiesusinghigh-qualityresearch
methodsweareunabletoconcludethattheyhave
beenverified effectiveinreducingrecidivism
(Abstract,italicsinoriginal).
Method
Thepresentstudywasdesignedtoexploretheimpact
ofemploymentonoffenderscommunity-based
outcomes(i.e.,measuresofsustainedreintegration,
includinglengthoftimeinthecommunity),while
controllingforriskandneedvariablesthatimpact
communityreintegration.Forthepurposeofthese
analyses,allavailabledataforfederallysentenced
offenderswereextractedfromtheCorrectional
ServiceofCanadas(CSCs)automateddatabase
(OffenderManagementSystem;OMS).Community
employmentinformationwasavailablefor
23,525federaloffendersreleasedonaconditional
releasebetweenJanuary1,1998andJanuary1,2005.
Approximately95%weremen(N=22,269)and5%
werewomen(N=1,256)*.
Theemploymentexperiencesoffederaloffenders
conditionallyreleasedtothecommunitywas
identifiedthroughCSCsOffenderManagement
System.Uponidentifyingtheemployedgroup,
thematchedgroupwasdevelopedusingSAS
(StatisticalAnalysisSystem)software.Next,offender
identifiersforbothgroupswerelinkedtothose
intheOMSdatacontaininginformationrelevant
tothestudy(demographicinformation,sentence
information,andratingsassociatedwiththestatic
anddynamiclevelsofintervention).
Thepopulationwasdividedintotwogroups:
offendersrecordedasbeingemployedbetweentheir
releasedateandtheendoftheirsentenceandthose
whowereunemployed.Theemployedgroupwas
thenrandomlymatchedtotheunemployedgroup,
withthematchingcriteriacontrollingfortime,
opportunityandtendency.Matchingcriteriaalso
addressedtheissueofco-morbidityinoffenderneeds.
Specifically,thegroupswerematchedongender,risk
level,releaseyear,sentencelength,severaldynamic
factors,**andtheregionalstatisticareaclassification
(SAC)***groupingswhichcorrespondedtothe
offendersdesignatedsupervisionoffice.The
matchingprocessyieldedsamplesof4,640men
and156women.****
Theinformationwassubjectedtosurvivalanalysis,
astatisticaltechniquethatestimatesthetimetaken
toreachaneventandtherateofoccurrenceofthat
event.Thistypeofanalysiswasusedtodraw
comparisonsbetweenemploymentgroupsacross
threeeventsoroutcomemeasures:1)anyreturnto
federalcustodybeforetheendofsentence;2)areturn
tofederalcustodywithanewoffencebeforeendof
sentence;and3)areturntofederalcustodywithout
anewoffencebeforeendofsentence.Comparisons
weredrawnformenandwomenseparately.
Results
Themediantimetooutcomewasusedasameasure
ofcentraltendencyforthesurvivaldata.An
examinationofthereleasecohortrevealedthatthe
survivalcurvesforemploymentweresignificantly
differentformenandwomen[
2
(1,N=24,061)=
19.40,p.001).Themediantimetoemployment
was6monthsformenand10monthsforwomen.
However,asillustratedifFigure1,bothsurvival
curveseventuallyconverged,indicatingthatover
time,therewerefewerdifferencesinemployment
ratesbygender.
Whencomparedtotheirmatchedcounterparts,
employedmenweremorelikelytoremainon
conditionalreleaseuntiltheendoftheirsentence
[
2
(1,N=4,653)=357.40,p.001)].Themediantimeto
returnwasalsolaterfortheemployedgroup(11
monthsversus37months,respectively).Employed
menwerealsolesslikelytoreturntofederalcustody
withanewoffence[
2
(1,N=4,653)=86.71,p.001)]or
technicalrevocation[
2
(1,N=4,653)=128.62,p.001)]
(seeFigures2,3and4).
Forwomenoffenders,theemployedgroupwas
morelikelytoremainonconditionalreleaseuntil
theendoftheirsentence[
2
(1,N=156)=9.09,p.01)].
Anexaminationofthesurvivalcurves(see
Figure5)revealsthatattheendofthestudyperiod,
approximately70%oftheemployedgroupremained
onconditionalreleasecomparedtoapproximately
55%oftheunemployedgroup.Lowbaseratesfor
returnswithanewoffenceprecludedanyestimation
ofthemediantimeinthecommunity.However,the
employedgroupwaslesslikelytoreturnwithanew
offencethantheirmatchedcounterparts[
2
(1,N=156)
=8.54,p.05)].Therewerenosignificantbetween
groupdifferencesfortechnicalrevocations.
P
E
R
S
P
E
C
T
I
V
E
S

11
P
E
R
S
P
E
C
T
I
V
E
S

Survival Curves Males

Months out Any return
Figure 2
0
0 30 40 50 60 70 80 20 10
0.25
0.50
0.75
1
Months out
S
u
r
v
i
v
a
l

D
i
s
t
r
i
b
u
t
i
o
n
F
u
n
c
t
i
o
n

Not yet employed Found employment
Survival Curves Males
Months out Return new offence
Figure 3
0
0 30 40 50 60 70 20 10
0.25
0.50
0.75
1
Months out
S
u
r
v
i
v
a
l

D
i
s
t
r
i
b
u
t
i
o
n
F
u
n
c
t
i
o
n

Not yet found employment Found employment
Survival Curves
Time to Employment
Figure 1
0
0 30 40 50 60 70 80 20 10
0.25
0.50
0.75
1
Months to employment
S
u
r
v
i
v
a
l

D
i
s
t
r
i
b
u
t
i
o
n
F
u
n
c
t
i
o
n

Women Men
12
Survival Curves Males
Months out Technical revocation
Figure 4
0
0 30 40 50 60 70 20 10
0.25
0.50
0.75
1
Months out
S
u
r
v
i
v
a
l

D
i
s
t
r
i
b
u
t
i
o
n
F
u
n
c
t
i
o
n

Not yet found employment Found employment
Survival Curves Women
Months out Any return
Figure 5
0 10 20 30 40 50
0.25
0.50
0.75
1
Months Out
S
u
r
v
i
v
a
l

D
i
s
t
r
i
b
u
t
i
o
n
F
u
n
c
t
i
o
n

Not yet found employment Found employment
Discussion and Implications
Thisresearchprovidesinformationonfederal
offenderscommunityemploymentoutcomes,
yieldingdatathattodate,hasnotbeenavailable.
First,theresultsprovideabaselineestimateofthe
averagelengthoftimeittakesformenandwomen
offenderstofindworkwhileonconditionalrelease.
Whereasmenwhofindworktakeamediantime
of6months,womentendtoobtainworkafter
10monthsinthecommunity.
Additionally,thisstudyprovidesanimportant
contributiontotheemploymentliterature,using
systematicandcontrolleddatatosituatejob
acquisitioninthecommunityasanimportantfactor
inoffenderreintegration.Specifically,thestudy
respondstomanyofthecriticismslevelledat
previousemploymentresearch,namely,thelackof
acomparisongroup.Comparisonsoncommunity-
basedoutcomesfortheoffendersinthecurrent
sampleshowedtheimpactofemploymentin
contributingtoanincreasedlikelihoodofsuccessful
sentencecompletion,alongerperiodoftimeinthe
community,andadecreasedlikelihoodofreturningto
theinstitutionforanewoffenceortechnicalviolation.
Thesefindingshaveimplicationsforcommunity-
basedprogramming,emphasizingtheneedfor
readily-accessibleemploymentinterventionsfor
offendersastheyarereleasedtothecommunity.
Althoughnotaprogramperse,CSCscommunity
employmentcentresprovideemploymentservices
toconditionally-releasedoffenderstopreparethem
tofindwork.Theservicesfocusonproviding
employmentservicestooffenders,including
individualemploymentassessment,counselling,
jobsearchtechniquesandon-the-jobplacements
P
E
R
S
P
E
C
T
I
V
E
S

13
tooffendersreleasedtothecommunity.Apreliminary
explorationofthecentreswasrecentlyconductedby
Gillis&Crutcher.
11
Thisprofile,alongwithrecent
evaluationfindings,
12
demonstratesthatthecentres
aremeetinganimportantdemand,respondingto
theriskandneedprinciplesinprovidingservices
tooffenderswithidentifiedemploymentneeds.For
offenderswithconsiderablebarriers,moreintensive
employmentprogrammingmaybenecessary,and
shouldbeaccessibletooffenderspriortorelease,or
astheyarereleasedtothecommunity.
Employment,asaprogram,hasbeeneclipsed
overthepastdecadeswiththeadventandwide
1
340LaurierAvenueWest,Ottawa,ONK1A0P9
2
Enocksson,K.(1981).Correctionalprograms:Areviewofthevalueof
educationandtraininginpenalinstitutions.Journal of Offender Counseling,
Services and Rehabilitation, 5(1),518.Also,seeFinn,P.(1998).Job
placementforoffendersinrelationtorecidivism.Journal of Offender
Rehabilitation, 28(1/2),89106.Seealso,Gendreau,P.,Goggin,C.,&
Gray,G.(1998).Caseneeddomain:Employment.Forum on
Corrections Research, 10(3),1619.
3
Motiuk,L.(1997).Classificationforcorrectionalprogramming:
TheOffenderIntakeAssessment(OIA)process.Forum on Corrections
Research, 9(1),1822.
4
Markley,H.,Flynn,K.,&Bercaw-Dooen,S.(1983).Offenderskills
trainingandemploymentsuccess:Anevaluationofoutcomes.
Corrective and Social Psychiatry and Journal of Behavior Technology
Methods and Therapy, 29,111.
5
Gillis,C.A.,&Andrews,D.A.(2002).Understandingemployment:
Aprospectiveexplorationoffactorslinkedtocommunity-based
employmentamongfederaloffenders.Forum on Corrections Research,
14(1),36.Also,seeGillis,C.A.,&Andrews,D.A.(2005).Predicting
community employment for federal offenders on conditional release. Research
report,ResearchBranch,CorrectionalServiceCanada,Ottawa,ON.
6
Gaes,G.G.,Flanagan,T.J.,Motiuk,L.L.,&Stewart,L.(1999).Adult
correctionaltreatment.InM.TonryandJ.Petersilia(Eds.),Prisons
(pp.361426).Chicago,IL:UniversityofChicagoPress.Also,see
Gerber,J.,&Fritsch,E.J.(1995).Adultacademicandvocational
correctionaleducationprograms:Areviewofrecentresearch.Journal
of Offender Rehabilitation, 22(1/2),119142.Seealso,Pearson,F.S.,
&Lipton,D.S.(1999).The effectiveness of educational and vocational
programs: CDATE meta-analyses. Paperpresentedattheannualmeeting
Dont be shy
distributionofprogramsdesignedtoaddressother
needareas(e.g.,substanceabuseandviolence).
Employmentinitiativeshaveexistedsincethe
adventofinstitutions,yetasnotedbyAndrews
etal.,
13
itcanbesaidthattheemploymentfactor,
forallofthetraditionalattentionithasreceived
incorrections,hasnotreceivedthequalityof
attentionitdeserves.Itishopedthatthisresearch
willcontributetosolidifyingtheperceptionof
employmentasanimportantfactorinoffenders
communityreintegration,andtofurtheringits
statusasasignificantprogramarea.
oftheAmericanSocietyofCriminology,Toronto,ON.Also,seeRyan,
T.A.(1998).Job retention of offenders and ex-offenders: Review and synthesis
of the literature. Unpublishedmanuscript,CollegeofCriminalJustice,
UniversityofSouthCarolina,Columbia,SC.
7
Op.Cit.,Ryan,1998,p.E5.
8
Op.Cit.,Gaes,1999;Gerber&Fritsch,1995;Pearson&Lipton,1999;
Ryan,1998.
9
Op.Cit.,Finn,1998.Also,seeGerber&Fritsch,1995.
10
Op.Cit.,Pearson&Lipton,1999.
11
Gillis&Crutcher,2005,thisvolume.
12
Op.Cit.,Gillis&Andrews,2005.
13
Andrews,D.A.;Pirs,S.;Walker,J.&Hurge,A.(1980).Atheoretical,
researchandprogramframeworkforemployment-orientedservices
inprobationandparole:(Aninterimreport).PartsI,II,III,IV,
andVI.Ontario:MinistryofCorrectionalServices.(p.3).
* Theanalysesexamineduniquesentences,thusitispossibleforoffenderstoappearmore
thanonceinthepopulation.
** Dynamicfactorswerethoseassessedjustpriortotheoffendersreleasedates.Theyare
comprisedofthefollowingdomains:employment,family/maritalrelations,associates,
substanceabuse,communityfunctioning,personalemotionalorientation,andattitudes.
*** TheSACsidentifygeographiczonesbasedonpopulationcountsanddensitiesresulting
fromthe2001CanadianPopulationCensus.Thezonesareclassifiedasbeingacomponent
of1)CensusMetropolitanAreas(CMA)whichhaveapopulationover100,000,2)Census
Agglomeration(CA)areaswhichhaveapopulationthatislessthan100,000butmore
than10,000and3)RuralCommunities(RC)whichincludeallothertown,villagesbut
excludesreservecommunities.Thisgeographicdesignationwasbasedonthepremise
thatoffendersresidedrelativelyclosetothelocationwheretheywerebeingsupervised.
****Eachsamplewascomprisedofunemployedoffenders(50%),andemployedoffenders(50%).
P
E
R
S
P
E
C
T
I
V
E
S

Feel free to drop us a line and let us know what you think of
FORUM. We are always happy to hear from our readers and
interested in any suggestions about our content, our look and
our approach.
14