ContentslistsavailableatChildAbuseNeglectjournalhomepagewwwelseviercom - PDF document

1 ContentslistsavailableatChildAbuse&Negle
ContentslistsavailableatChildAbuse&Neglectjournalhomepage:www.elsevier.com/locate/chiabuneg ResearcharticlesnotJustpoverty:Educational,social,andeconomic
programs,whichredirectoftenasubsetofneglectreferralstoanon-investigative,oftenvoluntarysystemre-sponsethateschewsvictim/perpetratorlabels(Hughes,Rycus,Saunders-Adams,Hughes,&Hughes,2013).However,lackofclearmeasuresofthecontextsofneglectinCPSrecordsmakeitdiculttoascertainwhetherCPSreportsofneglectarelargelyandprimarilyreectmaterialdeprivationduetopoverty.Moroever,theevidencebasethatseekstodierentiatetheeectsofneglectfromthoseofpovertyandabuseislimitedbyconsiderabledataandmeasurementissues.Currentestimatessuggestthat37%ofallU.S.children(andmorethanhalfofallU.S.Blackchildren)experienceaCPSin-vestigationbeforetheir18thbirthday(Kim,Wildeman,Jonson-Reid,&Drake,2017).Anestimated85%offamiliesinvestigatedbyCPSarebelow200%ofthefederalpovertyline(Dolan,Smith,Casanueva,&Ringeisen,2011),suggestingabroadreachofCPSintothelivesofpoorfamilies.Bycomparinglong-termoutcomesofchildrenreportedtoCPSforneglectwiththosewhoexperiencevaryingdegreesanddurationsofpoverty,wecanbetterunderstandwhethertheallegationsofneglectinvestigatedbyCPSaredistinctfrompovertyintheireectsonchildrenslaterfunctioning.Bycomparingchildrenwithneglectallegationsandchildrenwithabuseallegations,wefurtherassesswhetherabuseisassociatedwithworsesocial,educational,andeconomicandbehavioralout-comesthanneglect.Specically,weaddresstworesearchquestions:(1)Dothesocial,educational,andeconomicoutcomesofchildrenexposedtoallegedneglectdierfromoutcomesofchildrenexposedtoallegedabuseortopovertyalone,netofothersociodemographicconfounds?;and(2)Howdotheestimatedeectsofchildneglectandabusecomparetotheestimatedeectsof2.BackgroundPovertyiscorrelatedwithneglect(Berger&Waldfogel,2011Sedlaketal.,2010Slack,Holl,McDaniel,Yoo,&Bolger,2004andasmallbodyofevidenceindicatescausaleectsofincomeonmaltreatmentandneglectspecically(Berger,Font,Slack,&Waldfogel,2017Cancian,Yang,&Slack,2013Fein&Lee,2003).Whetherpovertycontributestoneglectisnotwidelydisputed;whatremainscontentiousiswhetherwhatisoftenreportedorlabeledasneglectsimplypoverty.Indeed,oneneednotlookfartondarticlesinmainstreammediaoutletswrittenbyorquotingacademicscholarsassertingthatthesysteminvestigatesorremoveschildrenonthebasisofpoverty(Chill,2018Dewan,2018Wexler,2019).StatepolicyandlocalpracticesgenerallyinstructCPScaseworkersnottomakeandingofneglectinsituationsthataroseforreasonsofpoverty(ChildWelfareInformationGateway,),irrespectiveofwhetherthechildwasharmedorlikelytobeharmedinthefuture.Yet,theextenttowhichCPScaseworkersareabletodistinguishbetweenpovertyasacontributingfactorversuspovertyasthesolefactorinneglectisquestionable(Eamon&Kopels,20042.1.DeningneglectAbasicdenitionofneglectisthefailureofaparentorcaregivertomeetachildbasicneedssuchthatachildshealthorsafetyisthreatened(ChildWelfareInformationGateway,2014).Statescivilstatutescommonlydelineatethoseneedstoincludefood,clothing,shelter,medicalcare,supervision,andprotection.However,itisrarelypossibletoparsealloftheformsofneglectthatcouldbeoccurringinanenvironmentbasedonthedocumentationpracticesofmostCPSagencies,wherethedetailsarecontainedinnarrativeformandthesub-categoriesofneglect,ifused,arebroad.Thus,thevastmajorityofresearchonneglectusesasingularconstruct,thoughsomestudies,typicallythoserelyingonparent-reportedneglect,havedierentiatedphysicalneglect(lackoffood,shelter,medicalcare,orclothing)fromallotherformsofneglect(Bergeretal.,2017Font&Berger,2015Yang&Maguire-Jack,Non-physicalformsofneglectarecommonlytermedgeneralneglectorsupervisionneglect,andofteninvolvesituationswherechildrenareleftaloneorleftinthecareofanunsafeorinappropriateperson,aswellaschildrensexposuretoparentalsubstanceabuse,domesticviolence,orcriminality.Theseformsofneglectofteninvolverole-modelingofantisocialbehaviorbyparents(e.g.,intoxication,violencetowardothers,criminalacts).Giventhatparentsarethepredominantunitofsocializationforchildren,childrenmayinternalizeandpracticethevaluese

2 xpressedintheirparentsbehaviors.Thoughcu
xpressedintheirparentsbehaviors.Thoughcurrentnationalmeasures(DepartmentofHealth&HumanServices,2017)arenotconsideredreliable,smallerstudieshaveestimatedthatparentalsubstanceabusewasacontributingfactorinatleastone-third,andperhapsasmanyas80%,ofCPScases(Murphyetal.,1991Semidei,Radel,&Nolan,2001Young,Boles,&Otero,2007).Inadditiontoadverseeectsofin-uterodrugandalcoholexposure,substanceabusemaycompromiseparentalcompetenceinnumerousways.Intoxicatedparentsarelessawareoforattentivetotheirchildrensneeds;theymaybehavemoreerraticallyoraggressively;and,whendependentonsubstances,parentsmaybefocusedontheacquisitionofdrugsoralcoholratherthancaregivingresponsibilities(Mayes&Truman,2002).Althoughsubstanceabuseisnotconstrainedtothoseoflowersocioeconomicstatus,substanceabusecananancialstabilitybyinterferingwithemploymentanddirectlyconsumingresourcesthatcouldbespentonhousing,food,andotherbasicneeds.Similarly,exposuretodomesticviolenceisbelievedtonegativelyaectchildrennotonlyduetotheimpactofwitnessingharmtoothers,butalsobecausethenon-oendingparentcapacitytoconnectwith,discipline,andattendtotheneedsoftheirchildreniscompromised(Holt,Buckley,&Whelan,2008).Incasesrelatedtolackofsupervision,whichmayormaynotinvolvesubstanceabuse,childrenaredeprivedofcognitivestimulationandemotionalconnection,andmayhavegreateropportunityfordelinquentbehavior.2.2.Evidenceonpoverty,neglect,anddevelopmentThecausalimpactsofchildhoodpovertyaremostclearlydemonstratedforchildrensacademicandcognitiveoutcomes(S.A.FontandK.Maguire-Jack Magnuson,&Votruba-Drzal,2017),andareatleastpartiallyexplainedbydierencesincognitiveenrichment(Reeves&Howard,).Cognitiveenrichmentcantakeavarietyofformsreadingtochildren,takingthemtomuseumsorotherplacesthatfacilitatelearning,andregularage-appropriateinteractionandcommunication.Itcanalsobeprovidedthroughmultiplesourcesextendedfamily,daycareproviders,orschoolpersonnel.Althoughpovertymaylimitthequalityofdaycareprovidersandschoolsavailable,parentsinpovertyprovidevaryingdegreesofotherformsofstimulation(Sperry,Sperry,&Miller,2019).Becauseneglectoftenmanifestsinpoorsupervisionandlackofattentiontochildrenswell-being,parentswhoareneglectfulmayprovidetheleastcognitiveenrichment.Hence,neglectedchildrenwouldbeexpectedtoexhibit,onaverage,worseacademicoutcomesthantheirnon-neglectedlow-incomepeers.Thesamefactorsmayberelevantforunderstandingdierencesbetweenpoornon-neglectedandneglectedchildreninsocial-behavioraloutcomes.Thelackofsupervisionandconsistentexpectationsamongneglectedchildrenmayleavethemill-equippedtofollowtherulesandstructureofotherenvironments,includingschoolandwork,whereasexposuretoinappropriatecaregiverbehavior(aggression,druguse,unreliability)providesbehavioralrolemodelsthatarenotconducivetoadaptivesocialfunctioning.Childrenwithimpoverishedbutnon-neglectfulparentsmayhavehigherexposuretoinappropriaterolemodelsduetothecon-centrationofpoorfamiliesindisorganized,high-crime,neighborhoods(Voisin,2007);theymayalsohavelessconsistentlystructuredenvironmentsduetomoreschedulinginstabilityinlow-wagejobs.IfCPSreportsofchildneglectprimarilyreectstructuralcon-straintsfacedbypoorfamilies,thenminimaldierencesinearlyadulthoodoutcomeswouldbeexpectedbetweenchildrenwithneglectallegationsandchildrenfrompoorfamilieswithoutsuchallegations.Incontrast,ifimpoverishednon-neglectfulparentsarebetterabletocontroltheirchildrensenvironmentsandprovideconsistentnurturingthanparentssuspectedofneglect,thentheirchildrenwouldbeexpectedtoexhibitcomparativelybettersocial-behavioraloutcomesthantheirneglectedcounterparts.However,incomepovertyisnotsucientforunderstandingyouthoutcomes;rather,itiswhenpovertyleadstomaterialhardshipordeprivationthatparentingbehaviorandchildwell-beingaremostimpacted(,Aber,Raver,&Lennon,2007).Thus,itmaybethatfamiliesreportedforneglectaresimplythemostseverelyandchronicallyeconomicallydisadvantaged,butthattheprimaryissueremainseconomicinnature.Regardless,economicdeprivationaectschildreninpartbecauseeconomically-distressedparentsarelessattentive,moreharsh,andhavemoreintraf

3 amilialconict(Conger,Ge,Elder,Lorenz,&Si
amilialconict(Conger,Ge,Elder,Lorenz,&Simons,1994Newland,Crnic,Cox,&Mills-Koonce,2013).Itispossiblethattheparentswhohavetheleastsuccessfulcopingmechanismsaremostlikelytomaltreatunderconditionsofeconomicstress.ComparingchildrenwithCPSneglectallegationswithchildrenexposedtovaryinglevelsofpovertyseverityanddurationprovidesinsightintowhetherallegedneglectisdistinctivelyharmful.Notably,neglectcanalsoariseindependentlyofpoverty,butthereislimitedunderstandingaboutthecontextinwhichthisoccurs.Itispossiblethatneglectmanifestsdiandhasdierenteinnon-poorfamilies.Parentswhohavethere-sourcestomeettheirchildsbasicneedsbutaretodosomayhaveamorenegativeimpact,atleastpsychologically,ontheirchildrenthanparentswhoareduetonancialproblems.Additionally,povertymayexacerbatethenegativeimpactsofneglect.Neglectedbutwealthierchildrenmayreceivebettereducationalsupportsduetohigherqualityschoolsthatcompensateforinadequatestimulationandsupportathome.Additionally,therisksofalackofsupervisionmaybegreaterinpoorfamilies.Giventhecorrelationbetweenindividualandneighborhoodpoverty,andbetweenneighborhoodpovertyandcrime(Sampson,Raudenbush,&Earls,1997),neglectedyouthlivinginpovertymayhavegreaterexposuretodeviantpeersthatcompound(orarenotosetby)parentalinuences.Thus,neglectmayincreasetheriskofadverseacademicandsocial-behavioraloutcomesoverall,butespeciallyforyouthinpoverty.2.3.LimitationsofpriorstudiesPriorstudiessuggestthatpoorfamiliesaremorelikelytoperpetrateabuseandneglect(Sedlaketal.,2010)andreachCPSatratesbecausetheyfacegreaterrisks(Jonson-Reid,Drake,&Kohl,2009).Yet,studiesexaminingtheeectsofneglecttypicallycontrolforormatchrespondentsoncurrentparent-reportincome(Chen,Propp,&Corvo,2011Englishetal.,2005Manly,Lynch,Oshri,Herzog,&Wortel,2013Whitaker,Phillips,Orzol,&Burdette,2007).Individualsoftenunderreporttotalincome(Rodin,Beebe,&Call,2005especiallyincomefrombenetprograms(Meyer&Mittag,2015orrefusetodiscloseincomeEpstein,2006).Moreover,householdsmaytransitioninandoutofpoverty,andasnapshotofincomedoesnotaccountfortheuniqueimpactofpersistentchildhoodpoverty(McLeod&Shanahan,1993).Absenthistoricaldataonfamiliesincome,eectsofneglectwillbeoverstatedifchronicpovertyismorecommonamongneglectfulfamiliesthannon-neglectfulfamilies.Studiesalsooftencombineallformsofchildmaltreatmentintoonevariable(Lauetal.,2005),whichcanbeproblematicifindividualtypesorcombinationsofmaltreatmenthavedierentialetiologiesorsequalae.Similarly,researchfocusingonasingletypeofmaltreatmentwilllikelyoverstateitseects,duetothehighrateofco-occuranceofmultiplemaltreatmenttypes(Turner,Vanderminden,Finkelhor,&Hamby,2019Vachon,Krueger,Rogosch,&Cicchetti,20152.4.CurrentstudyThecurrentstudyleveragesadministrativedatatoidentifytheassociationsofpovertyandneglectwithahostofyoungadultoutcomes.Bymeasuringneglectandpovertylongitudinally,weconstructamorerelevantcomparisongroupforneglectedchildrenandtherebyimproveidenticationofneglecteects.Toaddresstheco-occurrenceofmultipleformsofmaltreatment,andtoin-vestigatewhetherabuseismoreharmfulthanneglect,wecomparefourgroupsofchildren:thoseexposedtopovertybutnotmaltreatment,thoseexposedtoneglectbutnotabuse,thoseexposedtophysicalorsexualabusebutnotneglect,andthoseexposedtoacombinationofabuseandneglect.WefurtherconsiderwhethertheassociationsofabuseandneglectwithyoungadultoutcomesS.A.FontandK.Maguire-Jack varybydurationordepthofpovertyexposureduringchildhood.Lastly,weacknowledgethatCPSinvolvementcanitselfimpactchildren,netoftheeectsofmaltreatment.TheeectsofCPSinvolvement,thoughestimatedtobelargelynull(,McTavish,Turner,MacMillan,&Wathen,2018White,Hindley,&Jones,2015)couldbepositiveornegative,andtheimpactsofCPSin-volvementareconfoundedbytheseverityofmaltreatment,giventhatmoreseveremaltreatmentismorelikelytoresultininter-vention.Thus,wealsoconductsubsampleanalysesthatexcludechildrenwhoreceivedaCPSintervention,inordertoascertainwhetherchildrenwithostensiblylowerriskCPScasesneverthelessexperiencemoreadverseoutcomesinyoungadulthoodthanchildrenfromlow-incomefamilies.3.Method3.1.DataandsampleThisst

4 udyusedtheWisconsinAdministrativeDataCor
udyusedtheWisconsinAdministrativeDataCore,alinkedlongitudinaladministrativedatasethousedattheInstituteforResearchonPovertyattheUniversityofWisconsin-Madison,combinedwithrecordsfromtheWisconsinDepartmentofPublicInstruction(DPI)fortheyears20052016.TheDataCoreincludesindividual-leveladministrativerecordsfromstate-administeredpublicsocialwelfareprogramdatasystems,whichhavebeenlinkedacrossprogramsandovertime.Wespecicallyusedrecordsfromthechildwelfaresystem(CPS;whichincludesfostercaredata),SupplementalNutritionAssistanceProgram(SNAP;formerlyfoodstamps),UnemploymentInsurancesystem(UI;whichincludesearningsdata),SupplementalSecurityIncomeprogram(SSI),cashwelfare,DepartmentofCorrections(DOC),andtheMilwaukeeCountyJail.ChildwelfarerecordswerenotcompletelyelectronicinallWisconsincountiesuntilmid-2004.Thus,theanalysisfocusedonchildrenfromMilwaukeeCounty,whichhadelectronicchildwelfarerecordsbeginninginJanuary2000.ChildrendidnotneedtoremaininMilwaukeeCountyasadultstobeincludedintheoutcomemeasureswerecapturedstatewide.Toensurethatchildrensallegedmaltreatmentwasobservedfromarelativelyearlyage,weincludedonlythosebornbetween1993and1996;thus,thesamplewaslargelybetweentheagesof3and7whenCPSrecordsbecameavailable(January2000).Becausesomeoftheearlyadulthoodoutcomes(teenparenthoodandhighschoolcompletion)canoccurbeforelegaladulthood,wefocusedonpovertyandmaltreatmentoccurringpriortoage16.Outcomedataforthesamplewereobserveduntiltheir20Ofthosebornbetween1993and1996,youthwereincludedinthesampleiftheyresidedinMilwaukeeCountyatsomepointduringchildhood,andreceivedSNAPorweretheallegedvictimonaCPSinvestigationpriortoage16.Therewere39,102youthidentiunderthistargetpopulation.Youthwereexcludediftheywerereportedtohavediedbeforeage20(=164),wereknowntohavemovedoutofstate(=2676);theironlymaltreatmentallegationswerenotdenedasabuseorneglectoroccurredatorafterage161259);theyspenttimeinout-of-homecareforanyreason(4495);ortheyweremissinginformationontheirbiologicalmother=1354).Thenalsamplesizewas29,154individuals.Weexcludedchildrenwhospenttimeinout-of-homecarebecauseitwouldnotbepossibletodierentiatetheeectsoftheirmaltreatmentexposurefromtheeectsoftheirout-of-homecareexperiences.Thisexclusionmayunderstateeectsofmaltreatment,asthemostseverelymaltreatedchildrenwouldlikelyexperienceout-of-homeplacementathigherrates.TheCPSportionofthesample(=9278)wasdividedintotwogroups:thoseforwhomtheCPSsystemprovidedin-homeservicesorpetitionedthefamilycourt(=1477)andthosewhoseallegationsledtonointervention(=7801).Becausemoreseverecasesaremorelikelytoresultinintervention,theeectsofinterventionconfoundtheeectsofmaltreatmentinseverecases.However,byfocusingonsuspectedmaltreatmentthatwasconsideredinsucientlyseveretowarrantintervention,weaddresswhethernon-severecasesofneglectaredistinctfrompoverty.3.2.Measures3.2.1.OutcomesWefocusedoutcomemeasurementprimarilyonthersttwoyearsofadulthood,atages18and19.DatafromtheMilwaukeeCountyJailandallWisconsinstateprisonswereusedtoconstructbinaryindicatorsofincarceration.Becausethosewhoappearinprisonsarelikelytohavecommitted(orbeenchargedwith)moreseriousoensesthanthosewhoappearinjailonly,wemodeledjailincarcerationandprisonincarcerationseparately.Parenthoodbyage20wasbasedonmultiplesourcesofdatatolinkchildrentotheirbiologicalparents,includingMedicaidandchildsupportrecords,aswellasotherpublicassistanceandchildwelfarerecords.Theparenthoodmeasureincludedchildrenborntoparentsundertheageof18aswell,thoughamajorityofbirthsoccuratages18and19.HighschoolgraduationwasabinaryindicatordrawnfromDepartmentofPublicInstructionrecords;itdidnotincludeyouthwhoattendedprivateorchoiceschools,oryouthwhowerehome-schooled;thus,modelsofthisoutcomewerebasedonasmallersampleof24,161.Stableemploymentwasequalto1ifayouthreportedearningsinsixormorequartersinthersteightquartersfollowingtheirbirthday(i.e.,wereemployed75100%ofquartersintheirrsttwoyearsofadulthood).Earningsweremeasuredastheaverageearningsperquarterinthersteightquartersaftertheir18birthday,adjustedforation.Employmentandearningswerenotavailab

5 leforindividualsmissingavalidsocialsecur
leforindividualsmissingavalidsocialsecuritynumber;analysisforthoseoutcomeswasbasedonaslightlyreducedsampleof28,701.Together,thesixoutcomemeasurescapturedmultipledomainsoffunctioning:educational,economic,andsocial-behavioral.S.A.FontandK.Maguire-Jack 3.2.2.PovertyTomeasurepoverty,wereliedondatafromSNAP(foodassistance).AlthoughthethresholdforSNAPeligibilityisslightlyabovethefederalpovertyline(130%),itsusageamongpoorfamiliesisfarhigherthanforothermeans-testedprograms;thus,itisamongthemostinclusivemeasuresofpovertythatcanbegleanedfromadministrativedata.RateofSNAPparticipationamongeligiblepersonswasapproximately83%in2015,withhighestparticipationamongthefamiliesingreatesteconomicneed(U.S.DepartmentofAgriculture,2017).Tocapturechildhoodpoverty,weidentiedthosewhowereeverenrolledinSNAPpriortoage16.However,familiesmayenrollinSNAPduetoatemporaryeconomicshockorprolongedeconomicneed,andfamiliesenrolledinSNAPhavedierentlevelsofneed.Ifneglectcasesarecomprisedofmorechronicormoreseverepoverty,thenthosewhowereeverSNAPrecipientswouldnotbeanadequatecomparisongroup.Thus,wecreatedmeasuresofdurationanddepthofpoverty.Durationofpovertywasequaltothepercentofasubjectschildhoodmonths(ages016)inwhichtheyreceivedSNAP.Povertydepthwasequaltotheaveragepercentofthemaximumbenetthatthesubjectsfamilyreceived,conditionalonreceivingabenet.ThepercentofmaximumbenetwasequaltotheamountreceivedinamonthdividedbythemaximumbenetfortheirhouseholdsizepermittedunderUSDArules,whereahigherpercentageindicatesthatthefamilyhadlessincometocontributetotheirownfoodcosts.Monthsinwhichnobenetwasreceivedwerenotincludedinthecalculationofpovertydepth.3.2.3.MaltreatmentMaltreatmenttypesthatwereinvestigatedorconrmedbyCPSwereusedtocreatefourcategories:Noallegedorconmaltreatment(NM),allegedorconrmedneglectonly(NO),allegedorconrmedabuseonly(AO),andallegedorconrmedabuseandneglect(AN).ThemainindicatorofneglectwasbasedoninvestigatedallegationsofneglectbyCPSthatoccurredbetweenyear2000andthechilds16birthday.Weusedthe16birthdayasacutotoavoidmeasuringoutcomessuchasteenparenthoodandhighschooldropoutthatweredeterminedpriortheonsetofmaltreatment.Theprimarycode,Neglect-generallackofcare,fairlygenericandthusdidnotprovideindicationsoftheunderlyingcircumstances.Abuseallegationsincludedphysical,sexual,andemotionalabuseallegations.TypesofmaltreatmentweredrawnbothfromtheallegationsidentiedintheinitialCPSreportandfromsubstantiationdispositionfollowingtheinvestigation,whereapplicable.Inthemainmodels,wedidnotdierentiatebetweenin-vestigationsthatweresubstantiated(conrmed),becauseitisnotclearsuchdistinctionsareindicativeoftheveracityofallegationsHusseyetal.,2005Kohl,Jonson-Reid,&Drake,2009).Rather,investigationsweredierentiatedonthebasisofintervention.BecauseCPSintervenesinalimitednumberofcaseswhereinterventionisperceivedasnecessarytoprotectthechild,itislikelythatthesecasesinvolvehigherriskormoreharmfulsituationsthatcasesinwhichinterventionwasnotprovided.Weexcludedchildrenremovedfromthehomebecausethatisaninterventionthatdirectimpactschildwell-beinginwaysthatarebeyondthescopeofthisstudy,whereasin-homeservicesarelargelyfocusedonthechildscaregiversandeectsonchildwell-beingwouldbeindirect(&Font,2015).Interventionwasdenedasprovidingservicesorcasemonitoringorinvolvingfamilycourt.Itwasnotpossibletoerentiateeectsofinterventionfromseverityofmaltreatmentnoindependentmeasuresofseverityorharmwereavailableinthedata.3.2.4.CovariatesWecontrolledforseveralfactorsthatmayconfoundassociationsofpovertyandmaltreatmentwitheducation,economic,andsocialoutcomes.Youthcharacteristicscontrolledinallmodelswereyearofbirth,sex,race/ethnicity,andreceiptofSSI(asaproxyfordisabilitystatus).Familycharacteristicsweremothersageatrstbirth,numberofchildrenborntomother,numberoffatherstoschildren,andmaritalstatusatchildsbirth.Wealsocontrolledforsubjectsreceiptofcashassistance,employment,earnings,andincarcerationwhenthesubjectwasages03.3.AnalysisLogisticregressionmodelswereusedtoregresstheoutcomemeasuresonthemaltreatmentmeasure,thepovertymeasures,andtheinteract

6 ionsbetweenmaltreatmentandpoverty.Thepov
ionsbetweenmaltreatmentandpoverty.Thepovertymeasuresweremean-centeredandmodelsincludetheinteractionofpovertydurationanddepth.Toeaseinterpretationofmulti-categoryinteractionterms,Weplottedthepredictedprobabilitiesofeachoutcomebymaltreatmentandpovertyduration,withallothervariablesheldconstant.Modelsfocusedoncomparisonsbetweenyouthfromlow-incomefamilies(i.e.,youthwhoreceivedSNAPbenets)withoutCPShistoryand(1)allyouthwithaCPSin-vestigation,(2)youthwithaCPSinvestigationbutnointervention,and(3)youthwithaCPSintervention.4.Results4.1.SampledescriptionTable1containsadescriptionofthefullsamplebygroup(noallegationsofmaltreatment[NM],allegedneglectonly[NO],allegedabuseonly[AO],allegedabuseandneglect[AN]).YouthintheNMandAOgroupsexperiencedashorterdurationofpoverty(40and38%ofchildhoodmonths)thanthoseintheNOorANgroups(49and53%ofchildhoodmonths).NMyouthhadahigherestimateddepthofpovertywhilereceivingSNAPbenetsthanthemaltreatmentgroupsNMyouthreceivedabout63%ofthemaximumbenetwhileonSNAP,versus58%forNOyouth,60%forANyouth,and52%forAOyouth.Similarly,NOandANyouthweremorelikelytohavemotherswhoreceivedcashwelfareinearlychildhood(NOyouth:76%,ANyouth:79%)thanNMyouth(64%)orAOyouth(68%),andhadmotherswhowerelessfrequentlyemployedandhadlowerearnings.MothersofNOandANS.A.FontandK.Maguire-Jack childrenwerealsoyoungeratrstbirththanmothersofAOandNMchildren.Similarly,thereweresignicantbivariatedierencesintheoutcomesofyouthbymaltreatmentexposure.YouthintheNOandANgroupswerelesslikelytocompletehighschool(64%and59%)thanyouthintheAOorNMgroups(73%and78%).NOandANyouthwerealsomorelikelytohaveachildbeforeage20(18%and23%,vs.16%and13%).YouthintheANgroupweremorelikelythanallotheryouthtohavebeeninjail(18%)orprison(5%)byage20;NOyouthandAOyouthhadsimilarrisksofjail(14%vs.12%)andprison(4%vs.3%),whereasNMyouthhadthelowestrisk(jail9%,prison2%).NOandANyouthwerelesslikelytobestablyemployedandhadloweraveragequarterlyearningsthanAOorNMyouth.4.2.Regressionresults4.2.1.Ismaltreatmentassociatedwithyoungadultoutcomes,netofpovertyexposure?Inourrstsetofmodels,weestimatedtheoddsofeachoutcomeasafunctionofmaltreatment,povertyduration,averagepovertydepth,andsociodemographiccontrols(coecientsforcontrolvariablesnotshown).Inthefullsamplewefoundthat,comparedwithNMyouth,allyouthwithmaltreatmentallegationshadloweroddsofhighschoolgraduationandregularemployment,loweraverage Table1SampleDescriptionbyMaltreatmentExposure(fullsample).NoAllegedMaltreatmentAllegedneglectOnly(NO)AllegedneglectandabuseAllegedabuseonly(AO)N19,8763,0602,2483,970%/MeanSD%/MeanSD%/MeanSD%/MeanSDPovertyMeasures(SNAPreceipt,age010.49%5.46%15.10%Duration(%monthsreceived)40.1029.2349.1832.4652.8230.0038.4731.38Depth(mean%maxbenetwhen62.7818.4257.5626.5559.5722.4151.5527.23Outcomesthroughage20Graduatedhighschool77.53%63.59%59.10%72.28%Teenparenthood12.85%18.10%23.49%15.99%MilwaukeeCountyJail8.52%13.76%17.93%12.04%Stateprison1.81%3.92%4.85%2.92%Regularemployment35.54%29.58%30.65%35.47%Averagequarterlyearnings$1036.391284.24$930.181226.56$914.191200.99$1035.011243.87CharacteristicsofYouth'sBiologicalMotherAgeatrstbirth20.184.3619.244.0719.284.0220.154.41Numberofchildren3.771.984.512.264.202.083.511.90Incarceratedatchildages050.42%0.88%1.38%0.81%Receivedcashwelfareatchildages063.61%76.34%79.31%67.91%Avg.quartersemployedatchildages01.661.381.621.271.701.262.011.35Avg.annualearningsatchildages7015.349060.266232.749517.436265.958477.969296.4211106.12YouthFamilyStatusChildwasamaritalbirth12.49%12.78%14.99%17.23%Childwasanonmaritalbirth66.33%71.90%75.62%67.10%Child'sbirthstatusunknown21.18%15.33%9.39%15.67%Paternitynotestablished32.79%23.86%17.17%22.14%Paternityestablished,nosupport12.50%11.67%11.03%12.29%Supportorder54.71%64.48%71.80%65.57%YouthDemographicsWhitenon-Hispanic14.69%18.59%22.51%22.29%Blacknon-Hispanic55.25%59.71%53.20%51.01%Hispanic(anyrace)17.89%11.44%15.61%15.39%Otherrace/unknownrace12.17%10.26%8.67%11.31%Female48.40%47.35%53.02%53.25%Born199324.55%24.28%21.09%24.23%Born199424.99%24.54%25.04%24.51%Born199524.83%24.58%26.20%24.71%Bo

7 rn199625.63%26.60%27.67%26.55%ReceivedSS
rn199625.63%26.60%27.67%26.55%ReceivedSSIaschild7.04%12.03%18.28%11.66%:SNAP=SupplementalNutritionAssistanceProgram.SSI=SupplementalSecurityIncome(means-testedcashassistanceprogramforchil-drenoradultswithdisabilities).S.A.FontandK.Maguire-Jack earnings,andhigheroddsofteenparenthoodandjailorprisonincarceration.ANyouthwerelesslikelytograduatehighschoolandmorelikelytobecometeenparentsthanNOorAOyouth.AOyouthhigheroddsofincarcerationandNOyouthhighloweroddsofstableemployment.TherewerenostatisticallysignicantdierencesbetweenNOandAOyouthintheoddsofhighschoolgra-duation,teenparenthood,orearnings.Bothpovertydurationandpovertydepthwerenegativelyassociatedhighschoolgraduation,andpositivelyassociatedwithteenparenthoodandimprisonment.Povertydurationduringchildhoodwasassociatedwithslightlyhigheroddsofregularemploymentandwithhigheraverageearnings,whereaspovertydepthwasassociatedwithloweroddsofregularemploymentandlowerearnings.Notably,wecannotaccountforwhetheryouthwereenrolledinpostsecondaryeducation,whichmayreduceemploymentandearnings.Givenlowerratesofhighschoolcompletion,itislikelythatmaltreatedyouthwouldalsohavelowerratesofcollegeenrollment;failuretoaccountforcollegemaydownwardlybiasestimatesofdierencesinem-ploymentandearnings.Thesubsamplemodels(PanelsBandC)showthatyouthwithCPSallegationsthatdidnotleadtoanyinterventionwereneverthelessdistinctintheiroutcomeswhencomparedtoNMyouththecoecientsaresimilartothosereportedinthefullsamplemodels.WhenfocusingonlyonCPSyouthwhoreceiveanin-homeintervention(PanelC),thesizeofcocientsdiersfromthenon-interventiongroup,thoughnotinaconsistentdirection.Again,morenegativeoutcomesintheinterventiongroupmaybeattri-butabletoeectsoftheinterventionsthemselves,butthereisnoevidencethatin-homeservicesnegativelyimpactyouthoutcomes.4.2.2.EectsizesWeestimatedthepredictedprobabilitiesormeans(earningsonly)ofvariousoutcomesacrossdierentvaluesofpovertyexposureandmaltreatment.PredictedvalueswerebasedonlogisticregressionmodelsforPanelBthatallowedassociationsbetweenmal-treatmentandeachoutcometobemoderatedbypovertydepthandduration.Wedonotshowtheseinteractionmodelsintableformbecauseoflimitationsintheinterpretabilityofinteractiontermsinthecontextoflogisticregressionandinsteadfocusonthepredictedvalues.Wespecicallyexaminedthepredictedprobability(ormean,forearnings)ofeachoutcomeataverage,low,andhighpovertyexposure.Averagewasdenedasthemeansamplevaluefordurationanddepth(duration=40%ofchildhoodmonthsor6.4yearsofpoverty,depth=average63%ofmaximumbenet).Lowpovertyexposurewas-1standarddeviationonbothmetrics(duration=10%ofchildhoodmonthsor1.6yearsofpoverty,depth=average42%ofmaximumbenet),andhighpovertyexposurewas+1standarddeviationonbothmetrics(duration=70%ofchildhoodmonthsor11.2yearsofpoverty,depth=average84%ofmaximumbenet).Covariateswereheldconstantatthemeanforquantitativemeasuresandatthebasevalueforallcategoricalcontrolvariablesexceptraceandsex,whichweretreatedasbalancedacrossgroups.Wefocusedoncomparisonsofthosewithnoknownmaltreatment(NM)tothosewithmaltreatmentallegationsthatdidnotresultinanyCPSinterventiontoexcludeanyectsofCPSinterventiononrisk.WaldtestswereusedtodeterminethestatisticalsignicanceofgroupdiFig.1displayspredictedprobabilitiesofhighschoolgraduationandteenparenthood.Amongthosewithaverage(samplemean)povertydurationanddepth,thepredictedprobabilityofhighschoolgraduationforNMyouthwas0.83,about7percentagepoints(PP)higherthanforNOandAOyouth(.001),and14PPhigherthanANyouth(.001).Incomparison,thedierenceinprobabilityofhighschoolgraduationforNMyouthwithhighversusaveragepovertyexposurewas5PP,statisticallysigni Fig.1.ProbabilityofHighSchoolGraduationandTeenParenthoodbyAllegedMaltreatmentandPovertyExposure,PanelB.Estimatesarepredictedprobabilitiesand95%condenceintervalsfromlogitmodelsincludingallvariablesfromTable2plusinteractionsbetweenmaltreatmentandpovertyduration,andbetweenmaltreatmentandpovertydepth.Predictedprobabilitiesareestimatedtreatingraceandsexasbalancedacrossgroups,andholdingallothercategoricalcovariatesconstantatthebasecategoryandqu

8 antitativecovariatesatmean.S.A.FontandK.
antitativecovariatesatmean.S.A.FontandK.Maguire-Jack Table2LogisticRegressionResults.GraduatedhighschoolTeenparenthoodRegularlyEmployedAveragequarterlyearningsb(SE)b(SE)b(SE)b(SE)b(SE)b(SE)PanelA.FullSampleMaltreatmentExposure(reference:noneknown)Allegedneglect.512(.053)***.222(.054)***.316(.063)***.448(.113)***.308(.045)***107.709(21.621)***Allegedneglectandabuse.793(.059)***.505(.057)***.750(.067)***.821(.121)***.338(.052)***160.404(27.800)***Allegedabuse.429(.049)***.236(.051)***.503(.060)***.631(.113)***.186(.039)***106.389(21.829)***Povertyexposure.005(.001)***.011(.001)***.014(.001)***.017(.002)***.006(.001)***3.006(.331)***.007(.001)***.003(.001)**.002(.001).006(.003)*.004(.001)***4.321(.371)***PanelB.NMandCPSNon-InterventionMaltreatmentExposure(reference:noneknown)Allegedneglect.491(.057)***.163(.059)**.360(.068)***.430(.122)***.299(.048)***90.495(26.600)***Allegedneglectandabuse.724(.068)***.490(.067)***.643(.080)***.676(.144)***.292(.059)***139.182(32.391)***Allegedabuse.420(.051)***.235(.053)***.514(.062)***.622(.119)***.191(.041)***104.678(22.884)***Povertyexposure.005(.001)***.011(.001)***.014(.001)***.018(.002)***.006(.001)***3.157(.340)***.006(.001)***.003(.001)*.002(.001).007(.003)*.004(.001)***4.264(.382)***PanelC.NMandCPSIn-HomeInterventionMaltreatmentExposure(reference:noneknown)Allegedneglect.617(.119)***.496(.115)***.112(.149).525(.235)*.359(.114)**202.141(58.730)***Allegedneglectandabuse.950(.102)***.522(.098)***1.001(.112)***1.109(.188)***.456(.096)***219.383(50.117)***Allegedabuse.495(.135)***.263(.142).425(.173)***.709(.310)*.200(.116)127.885(65.079)*Povertyexposure.005(.001)***.011(.001)***.015(.001)***.018(.002)***.006(.001)***3.626(.389)***.006(.001)***.001(.001).002(.002).004(.003).008(.001)***p.05**p.01***p.001PanelA:NOAN***ANAO***;PanelB:NOAN**ANAO***;PanelC:NOAN*ANPanelA:NOAN***ANAO***;PanelB:NOAN***ANPanelA:NOAN***NOAO**ANAO*;PanelB:NOAN**;PanelC:NOAN***ANPanelA:NOAN**;PanelC:NOPanelA:NOAO*ANS.A.FontandK.Maguire-Jack smallerthandierencesassociatedwithneglectonly,abuseonly,orcombinedabuseandneglect.Thatis,thedecreaseinprobabilityofhighschoolgraduationassociatedwithaCPSinvestigationisgreaterthanthedecreaseassociatedwithspendinganadditional4.8yearsinpoverty.Ataveragepovertydurationanddepth,theprobabilityofteenparenthoodforNMyouthwas0.10,similartotheriskforNOyouth(=.107),and3PPlowerthanAOyouth(.001)and4PPlowerthanANyouth(.001).Incomparison,riskforhighpovertyNMyouthwas.13,statisticallyequivalenttotheratesofaveragepovertyAOandANyouth,buthigherthanaveragepovertyNOyouth(Predictedprobabilitiesofexperiencingjailandprisonbyage20areshowninFig.2.Thepredictedprobabilityofjailbyage20was0.03forNMyouthataveragepovertyexposure,1PPlowerthanNOyouth(.001),3PPlowerthanANyouth(.001),and2PPlowerthanAOyouth(.001).Incomparison,NMyouthwithhighpovertyexposurewere2PPmorelikelytoexperienceofa1SDincreaseinpovertydurationanddepthwasnotsignicantlydierentfromthatofneglect(NO,=.150),abuseandneglect(AN,=.052),orabuse(AO,=.203).Findingsforriskofprisonweresimilartondingsforriskofjail.Lastly,thepredictedprobabilitiesofstableemploymentandpredictedaveragequarterlyearningsareshowninFig.3.Thepredictedprobabilityofstableemploymentinthe8quartersafterreachingage18was.41forNMyouthataveragepovertyexposure;6PPlowerthanforNOyouth(.001)andANyouth(.001),and4PPlowerthanAOyouth(.001).Moreover,allmaltreatmentgroupsataveragepovertyhadalowerprobabilityofstableemployment(=.008)andlowerearnings=.026)thanNMyouthwithhighpovertyexposure.4.2.3.SensitivityanalysesWetestedadditionalmodelswithaninteractionbetweenpovertydepthandpovertydurationandanonlineartermforpovertyduration.Neitherchangesubstantivelyalteredtheresults.Wealsoacknowledgepossibledierencesintheeectsofsubstantiatedandunsubstantiatedinvestigations.However,becausethevastmajorityofchildrenwhoneverreceivedaCPSinterventionalsodidnothavesubstantiatedinvestigations,itwasnotpossibletoseparateeectsofsubstantiationandintervention.5.DiscussionThisstudysoughttoassesswhether,andhow,theeectsofCPS-investigatednegl

9 ectdierfrompovertyandfromCPS-in-vestigat
ectdierfrompovertyandfromCPS-in-vestigatedabuse.Ofcourse,causaleectsofmaltreatmentarediculttoidentify,giventhatexperimentaldesignswouldbeun-ethicalandtherearemanyconfoundingfactors.However,inthisstudy,weimproveonpriorestimatesbyaccountingforarichsetoflongitudinalchildhoodeconomicandfamilycharacteristics.WefoundthatyouthwithCPS-investigatedneglecthavesubstantiallyworseoutcomeslowerratesofhighschoolgraduationandregularemployment,andhigherratesofteenparenthoodandin-thanyouthwithoutmaltreatmentallegationswhowereexposedtosimilardurationanddepthofpoverty.Moreover,worseoutcomeswereobservedevenforyouthwithallegationsthatledtonoCPSinterventionostensiblylow-riskcases.Lastly,whereasyouthexposedtobothabuseandneglectfareworsethanyouthexposedtoasingleformofmaltreatment,wefoundnoconsistentindicationthatabusewasassociatedwithworseoutcomesthanneglect.Rather,abusewasmorepredictiveofjailand Fig.2.PredictedProbabilitiesofJailandPrisonIncarcerationbeforeAge20byAllegedMaltreatmentandPovertyExposure,PanelB.Estimatesarepredictedprobabilitiesand95%condenceintervalsfromlogitmodelsincludingallvariablesfromTable2plusinteractionsbetweenmaltreatmentandpovertyduration,andbetweenmaltreatmentandpovertydepth.Predictedprobabilitiesareestimatedtreatingraceandsexasbalancedacrossgroups,andholdingallothercategoricalcovariatesconstantatthebasecategoryandquantitativecovariatesatmean.S.A.FontandK.Maguire-Jack prisonincarceration,whereasneglectwasmorepredictiveoflackingstableemployment.Thisisconsistentwithpriorresearchshowingthatneglectisstronglyassociatedwithcognitionandlearning,whichmayaectemployment,whereasabuseismorepredictiveofexternalizingbehavior,whichmaymanifestincriminalactivity(Font&Berger,2015Manly,Kim,Rogosch,&Cicchetti,Aboutninetypercentofourmaltreatmentsample(88%forNeglectonly,93%forAbuseandNeglect,84%forAbuseonly)receivedfoodassistanceatsomepointpriortoage16,conrmingthattheCPSsystemisoverwhelmingcomprisedofchildrenfromlow-incomefamilies.Moreover,thechildrenthatreachtheattentionofCPSspendmoretimeinpovertythanotherchildrenonpublicassistance.However,ourstudysuggeststhatallegationsofneglectmatterbeyondtheeectsofpoverty,andthattheprocessofreportingandscreeningforinvestigationis,onaverage,eectivelycapturingadistinctlyat-risksubsetofimpoverishedyouth.Indeed,childrenidentiedasatriskofneglecthaveworseoutcomesthanimpoverishedchildrenacrossmultipledomains,evenathighlevelsofpoverty.Totheextentthattherisksofneglectfulfamiliesaredistinctfromnon-neglectfulimpoverishedfamilies,interventionstargetingeconomicneedsalonemaybeinadequate.ndingsofthisstudyalsochallengetheperceptionthatneglectislessharmfulthanabuse.PriorresearchershavespeculatedthatthisperceptionmaybeduetotheconsequencesofabusebeingmoreimmediatelyobservablethantheconsequencesofneglectDubowitz,2007).Nevertheless,thedivergenceinoutcomesforneglectedyouthandimpoverishednon-neglectedyouthissig-cant,atleastbyearlyadulthood.Giventheprevalenceofneglectandtheincreasedriskofadverseoutcomesassociatedwithneglect,targetedeortstopreventandtreattheeectsofneglectwarrantgreaterpriority.ndingsofthisstudyalsohaveimplicationsfortheCPSsystem.Ourresultssuggestthatcurrentchildprotectionpractice(whatisreportedtoCPSandscreenedinforinvestigation)is,onaverage,identifyingasubsetofimpoverishedchildrenatdis-proportionatelyhighriskofadverselifeoutcomesitisnottargetinglow-incomefamiliesindiscriminately.Atthesametime,allegationsthatledtonointerventionwereneverthelessassociatedwithadverseoutcomes.Inotherwords,allegedmaltreatmentthatiseitherunabletobeprovenorisdeemedinsucientlyseveretowarrantinterventionisneverthelessasignicantpredictorofahostofoutcomesnetofpovertyanddemographics.Notably,however,CPSmaynotbetheappropriateagencytoaddresstherisksfacedbytheseyouth,giventhattheservicesprovidedorprocuredbyCPSpredominatelyfocusonreducingtheriskthatparentswillperpetratenewmaltreatment,notonchildrenshealthordevelopment(Berger&Font,2015).Asaresult,thesocial-emotionalandeducationalneedsofchildrenmaynotbedirectlyaddressed.Increasedaccesstohighqualityservicesfor

10 allfamiliescomingtotheattentionofCPS,reg
allfamiliescomingtotheattentionofCPS,regardlessofscreeningandsubstantiationdecisionsiscriticalforimprovingoutcomesforallchildren.Somestateshaveimplementedtracksinwhichfamilieswhoarescreenedoutorunsubstantiatedareoeredvoluntaryservicestohelpimprovefamilycircumstancesinaneorttopreventfuturemaltreatment.However,preventionoffuturemaltreatmentmaynotbesutoimprovechildrensdevelopmentaloutcomes.Giventhatfactorsliketeenparenthood,loweducationalattainment,andcriminalityareriskfactorsforperpetratingmaltreatment,greaterattentiontothedevelopmentalneedsofchildreninvestigatedbyCPSbywayofproperassessmentandreferraltoservicesforchildrenmaypaydividendsviareducedintergenerationaltransmission. Fig.3.ProbabilityofStableEmploymentandMeanQuarterlyEarningsatAges18and19,byMaltreatmentandPovertyExposure,PanelB.Estimatesarepredictedprobabilities(foremployment)ormeans(earnings)and95%condenceintervalsfromlogit(employment)andlinear(earnings)regressionmodelsincludingallvariablesfromTable2plusinteractionsbetweenmaltreatmentandpovertyduration,andbetweenmaltreatmentandpovertydepth.Predictedvaluesareestimatedtreatingraceandsexasbalancedacrossgroups,andholdingallothercategoricalcovariatesconstantatthebasecategoryandquantitativecovariatesatmean.S.A.FontandK.Maguire-Jack 5.1.LimitationsandimplicationsforfutureresearchWenoteseverallimitations.First,thedataarefromresidentsofasinglecounty,and,givenvariabilityinhowmaltreatmentisreportedandscreenedacrossstates,thesendingsmaynotgeneralizenationally.Forexample,agenciesthatdeclinetoinvestigateagreaterproportionoflower-riskreferralsmayobservelargerdierencesbetweenitsCPSpopulationandthebroaderpopulationofimpoverishedchildren,whereasagenciesthatinvestigatemostorallreferralsmayobserveweakerdierences.MilwaukeeCounty,thesiteofthisstudy,investigatesapproximately80%ofreports(Dumas,Elzinga-Marshall,Monahan,vanBuren,&Will,2015higherratethanthenationalaverage(U.S.DepartmentofHealth&HumanServices,2017Second,themeasuresofmaltreatmenttypesaresimplistic,giventhebreadthanddepthofallegationsthatmaybeinvestigated.Thisstudydemonstratedonlythat,atvariousthresholdsofpoverty,thosewithneglectallegationshaveahigherprobabilityofadverseoutcomesthanthosewithoutanymaltreatmentallegations,andaresimilartothosewithabuseallegations.Yet,itremainspossiblethat,forexample,allegationsinvolvinginadequatesupervisionduetoparentalsubstanceabuseareuniquelypredictiveofadverseoutcomes,whereasallegationspertainingtoinadequatehousingarenot.AlthoughmanystudiesofneglectrelyonCPSallegationcodes,thisremainsaseriouslimitationtothestudyofneglect.Themeasuresalsocontainnoindicationofseverityofharmanditisimpossibletoascertainwhethertheabuseandneglectgroupswereexposedtothesamedegreeofharmorriskofharm.Itispossiblethatthereisadierentthresholdforremovaltoout-of-homecareincasesofabusethanneglectbecausetheimpactsaremoreimmediatelyobservable(e.g.,physicalinjuries),andthat,asaresult,thein-homeCPSabuseandneglectgroupsaresub-stantivelydissimilar.Third,ourcomparisongroupsuerssomecontaminiation(e.g.,someofthosewithoutidentiedmaltreatmentallegationsex-periencedmaltreatment)fortworeasons:notallmaltreatmentisinvestigatedbyCPSandelectronicCPSrecordswerenotavailablepriorto2000.ContaminationconcernsaresomewhatdiminishedbythenumerousyearsofCPSdata,asthecumulativelikelihoodthatamaltreatedchildreachestheattentionofCPSshouldincreaseovertime.Nevertheless,contaminationislikelytodownwardlybiasourestimates.StudiescombiningsurveyandadministrativedatatoaccountforsamplecontaminationwouldaddressthisFourth,theremaybeomittedvariablesthatbiasourndings.Forexample,familieswithneglectallegationsmaybemorelikelytohavespecicformsofmaterialhardship,suchashousinginstability,thanfamiliesinourpovertycomparisonconditions.HousingproblemsareverycommonamongfamiliesinvolvedwithCPS(Courtney,McMurtry,&Zinn,2004)andmaybeauniqueriskfactorforadverseoutcomes(Slack,Font,Maguire-Jack,&Berger,2017Lastly,weacknowledgelimitationstotheuseofSNAPreceipttomeasurepoverty.Althoughthevastmajorityofthoseeligiblereceiveassistance,especiallythosea

11 thigherratesofpoverty(U.S.DepartmentofAg
thigherratesofpoverty(U.S.DepartmentofAgriculture,2017),notalldo.Moreover,theremaybemisreportinginincomeandassets,whichcouldresultininaccurateestimatesofpovertydepthforsomechildren.However,becausethesemeasuresareconsistentlystructuredovertime,havesomedegreeofvalidationduetoprogrameligibilityverianddonotrelyonself-report,wearguethesemeasuresarepreferabletothoseusedinotherstudies.Nevertheless,giventheselimitations,researchersshouldcontinuetoexaminetheeectsofneglecttoconrmorchallengeourndings,andourresultsidentifyareaswheremeasurementcouldbeimproved.Ourresearchpointstotheneedtoaccountformultipleformsofmaltreatmentexposure.Agreatdealofexistingresearchfocusesonasingleformofmaltreatment,ormeasuresmaltreatmenttypebasedonwhatisallegedatasinglepointintime.Ourndingsshowthatyouthexposedtobothabuseandneglectfaredsignicantlyworsethanyouthexposedtoabuseorneglectalone;thus,studiesthatfocusonabusewithoutaccountingforneglect,orassignahierarchyofmaltreatmentinwhichabuseisassumedtobeprimary,willlikelyoverstatetheeectsofabuse.Inaddition,ourstudyemphasizestheneedtodisaggregatethemultiplesubtypesofneglect.Fordecades,researchershavecalledforimprovementtothemeasurementofneglect,inordertobetteridentifyitsantecedentsandconsequences(Dubowitz,Pitts,&Black,Slack,Holl,Altenbernd,McDaniel,&Stevens,2003)butfederaldatacollectionthroughtheNationalChildAbuseandNeglectDataSystemcontinuestocombineallformsofneglect,andotherprominentsurveysofchildmaltreatmentalsoincludelimitedmeasuresofneglect.Moreresearchisneededtodeterminethelongtermeectsofneglectsubtypes,and,particularly,whethertheyaredistinctfrompoverty.5.2.ConclusionTheCPSsystemreceivesalargevolumeofreferralseachyear,mostofwhichinvolveneglectandresultinnointervention.Yet,severalyearsaftertheallegedincidentofneglect,youthareathighriskforadverseoutcomes,netofpovertyexposureandfamilycharacteristics.Indeed,evenamongthoseexposedtolong-termpoverty,thosewhoalsohaveneglectallegationsarelesslikelytograduatehighschoolortoberegularlyemployed,aremorelikelytoexperienceincarceration,andhavelowerearnings.Notably,ndingslargelyholdevenwhenfocusingonlyoncases(thoseforwhichnointerventionwasprovided).Formanyoutcomes,therearenodierencesbetweenthosewithneglectallegationsandthosewithabuseallegations,althoughthosewithbothabuseandneglectallegationsareathighestriskofadverseoutcomes.Insum,despitethatchildhoodpovertyandneglectarefrequentlycomorbid,neglecthasdistinctlynegativeassociationswithyouthoutcomes,similartoabuse.Giventhatdenitionsofneglectarebroadandadministrativemeasuresaresimplistic,wecannotdeterminewhatexactlychildrenreportedforneglectareexperiencingthatplacesthematincreasedrisk.However,thisstudysuggeststhatitisnotpovertyalone.S.A.FontandK.Maguire-Jack ThisworkwasfundedbytheEuniceKennedyShriverNationalInstituteofChildHealthandHumanDevelopment(R21HD091459)andwithsupportfromtheCapstoneCenterTranslationalCenterforChildMaltreatmentStudies(P50HD089922)andthePopulationResearchInstituteatPennStateUniversity(P2CHD041025).WethanktheWisconsinDept.ofChildrenandFamilies,Dept.ofHealthServices,Dept.ofCorrections,Dept.ofPublicInstruction,andDept.ofWorkforceDevelopmentfortheuseofdata,butacknowledgethattheseagenciesdonotcertifytheaccuracyoftheanalysespresented.AppendixA.SupplementarydataSupplementarymaterialrelatedtothisarticlecanbefound,intheonlineversion,atdoi:,T.O.,McTavish,J.,Turner,S.,MacMillan,H.L.,&Wathen,C.N.(2018).TherelationshipbetweenchildprotectioncontactandmentalhealthoutcomesamongCanadianadultswithachildabusehistory.ChildAbuse&Neglect,79,22Berger,L.M.,&Font,S.A.(2015).Theroleofthefamilyandfamily-centeredprogramsandpolicies.TheFutureofChildren,25(1),155Berger,L.M.,Font,S.A.,Slack,K.S.,&Waldfogel,J.(2017).Incomeandchildmaltreatmentinunmarriedfamilies:Evidencefromtheearnedincometaxcredit.ReviewofEconomicsoftheHousehold,15(4),1345Berger,L.M.,&Waldfogel,J.(2011).Economicdeterminantsandconsequencesofchildmaltreatment.RetrievedfromOECDwebsite:social-issues-migration-health/economic-determinants-and-consequences-of-child-maltreatment_5kgf09zj

12 7h9t-enCancian,M.,Yang,M.-Y.,&Slack,K.S.
7h9t-enCancian,M.,Yang,M.-Y.,&Slack,K.S.(2013).Theeectofadditionalchildsupportincomeontheriskofchildmaltreatment.TheSocialServiceReview,87Chen,W.-Y.,Propp,J.,&Corvo,K.(2011).ChildneglectanditsassociationwithsubsequentjuveniledrugandalcoholoChildandAdolescentSocialWorkJournal,28(4),273ChildWelfareInformationGateway(2014).nitionsofchildabuseandneglect.RetrievedfromU.S.DepartmentofHealthandHumanServiceswebsite:Chill,P.(2018).HundredsofU.S.Childrentakenfromhome.Hartfordcourant.RetrievedfromConger,R.D.,Ge,X.,Elder,G.H.,Lorenz,F.O.,&Simons,R.L.(1994).Economicstress,coercivefamilyprocess,anddevelopmentalproblemsofadolescents.Development,65(2),541Courtney,M.E.,McMurtry,S.L.,&Zinn,A.(2004).Housingproblemsexperiencedbyrecipientsofchildwelfareservices.ChildWelfare,83(5),393Davern,M.,Rodin,H.,Beebe,T.J.,&Call,K.T.(2005).Theeectofincomequestiondesigninhealthsurveysonfamilyincome,povertyandeligibilityestimates.HealthServicesResearch,40(5p1),1534Dewan,S.(2018).Familyseparation:ItsaproblemforU.S.citizens,too.November2)RetrievedfromTheNewYorkTimesDolan,M.,Smith,K.,Casanueva,C.,&Ringeisen,H.(2011).NscawIIbaselinereport:IntroductiontonscawIInalreportOPREreport2011-27aWashington,DC:OceofPlanning,ResearchandEvaluation,AdministrationforChildrenandFamilies,U.S.DepartmentofHealthandHumanServicesH.(2007).Understandingandaddressingtheneglectofneglect:diggingintothemolehill.ChildAbuse&Neglect,31(6),603Dubowitz,H.,Pitts,S.C.,&Black,M.M.(2004).Measurementofthreemajorsubtypesofchildneglect.ChildMaltreatment,9(4),344Dumas,A.,Elzinga-Marshall,G.,Monahan,B.,vanBuren,M.,&Will,M.(2015).ChildWelfareScreeninginWisconsin:Ananalysisoffamiliesscreenedoutofchildprotectiveservicesandsubsequentlyscreenedin.RetrievedfromDepartmentofChildrenandFamilieswebsite:Duncan,G.J.,Magnuson,K.,&Votruba-Drzal,E.(2017).Movingbeyondcorrelationsinassessingtheconsequencesofpoverty.AnnualReviewofPsychology,68Eamon,M.K.,&Kopels,S.(2004).Forreasonsofpoverty:Courtchallengestochildwelfarepracticesandmandatedprograms.ChildrenandYouthServicesReview,(9),821English,D.J.,Upadhyaya,M.P.,Litrownik,A.J.,Marshall,J.M.,Runyan,D.K.,Graham,J.C.,etal.(2005).Maltreatmentswake:Therelationshipofmaltreatmentdimensionstochildoutcomes.ChildAbuse&Neglect,29(5),597Epstein,W.M.(2006).ResponsebiasinopinionpollsandAmericansocialwelfare.TheSocialScienceJournal,43(1),99Fein,D.J.,&Lee,W.S.(2003).TheimpactsofwelfarereformonchildmaltreatmentinDelaware.ChildrenandYouthServicesReviewFont,S.A.,&Berger,L.M.(2015).Childmaltreatmentandchildrensdevelopmentaltrajectoriesinearlytomiddlechildhood.ChildDevelopment,86(2),536,E.T.,Aber,J.L.,Raver,C.C.,&Lennon,M.C.(2007).Incomeisnotenough:Incorporatingmaterialhardshipintomodelsofincomeassociationswithparentingandchilddevelopment.ChildDevelopment,78(1),70Holt,S.,Buckley,H.,&Whelan,S.(2008).Theimpactofexposuretodomesticviolenceonchildrenandyoungpeople:Areviewoftheliterature.ChildAbuse&Neglect,(8),797Hughes,R.C.,Rycus,J.S.,Saunders-Adams,S.M.,Hughes,L.K.,&Hughes,K.N.(2013).Issuesindierentialresponse.ResearchonSocialWorkPractice,23Hussey,J.M.,Marshall,J.M.,English,D.J.,Knight,E.D.,Lau,A.S.,Dubowitz,H.,etal.(2005).Demaltreatmentaccordingtosubstantiation:distinctionwithoutadiChildAbuse&Neglect,29(5),479Jonson-Reid,M.,Drake,B.,&Kohl,P.L.(2009).Istheoverrepresentationofthepoorinchildwelfarecaseloadsduetobiasorneed?ChildrenandYouthServicesReview,,422Kim,H.,Wildeman,C.,Jonson-Reid,M.,&Drake,B.(2017).LifetimeprevalenceofinvestigatingchildmaltreatmentamongUSchildren.AmericanJournalofPublicHealth,107(2),274Kohl,P.L.,Jonson-Reid,M.,&Drake,B.(2009).Timetoleavesubstantiationbehind:Findingsfromanationalprobabilitystudy.ChildMaltreatment,14(1),17Lau,A.S.,Leeb,R.T.,English,D.,Graham,J.C.,Briggs,E.C.,Brody,K.E.,etal.(2005).Whatsinaname?Acomparisonofmethodsforclassifyingpredominanttypeofmaltreatment.ChildAbuse&Neglect,29(5),533Manly,J.T.,Kim,J.E.,Rogosch,F.A.,&Cicchetti,D.(2001).Dimensionsofchildmaltreatmentandchildrensadjustment:ContributionsofdevelopmentaltimingandDevelopmentandPsychopathology,13(04),759S.

13 A.FontandK.Maguire-Jack Manly,J.T.,Lynch
A.FontandK.Maguire-Jack Manly,J.T.,Lynch,M.,Oshri,A.,Herzog,M.,&Wortel,S.N.(2013).Theimpactofneglectoninitialadaptationtoschool.ChildMaltreatment,18(3),155Mayes,L.C.,&Truman,S.D.(2002).Chapter13:Substanceabuseandparenting.InM.H.Bornstein(Vol.Ed.),Handbookofparentingvolume4:Socialconditionsandappliedparenting:Vol.4,(pp.329359).Mahwah,NJ:LawrenceErlbaumAssociatesMcLeod,J.D.,&Shanahan,M.J.(1993).Poverty,parenting,andchildrensmentalhealth.AmericanSociologicalReview,Meyer,B.D.,&Mittag,N.(2015).Usinglinkedsurveyandadministrativedatatobettermeasureincome:Implicationsforpoverty,programeectivenessandholesinthesafetyNationalBureauofEconomicResearchMurphy,J.M.,Jellinek,M.,Quinn,D.,Smith,G.,Poitrast,F.G.,&Goshko,M.(1991).Substanceabuseandseriouschildmistreatment:Prevalence,risk,andoutcomeinacourtsample.ChildAbuse&Neglect,15(3),197Newland,R.P.,Crnic,K.A.,Cox,M.J.,&Mills-Koonce,W.R.(2013).Thefamilymodelstressandmaternalpsychologicalsymptoms:Mediatedpathwaysfromeconomichardshiptoparenting.JournalofFamilyPsychology,27(1),96Pelton,L.H.(2016).Separatingcoercionfromprovisioninchildwelfare:Preventivesupportsshouldbeaccessiblewithoutconditionsattached.ChildAbuse&Neglect,(1),427Reeves,R.,&Howard,K.(2013).Theparentinggap.RetrievedfromRoberts,D.E.(2014).ChildprotectionassurveillanceofAfricanAmericanfamilies.TheJournalofSocialWelfare&FamilyLaw,36(4),426Sampson,R.J.,Raudenbush,S.W.,&Earls,F.(1997).Neighborhoodsandviolentcrime:AmultilevelstudyofcollectiveeScience,277(5328),918Sedlak,A.J.,Mettenburg,J.,Basena,M.,Peta,I.,McPherson,K.,Greene,A.,etal.(2010).Fourthnationalincidencestudyofchildabuseandneglect(NIS-4).DC:USDepartmentofHealthandHumanServicesRetrievedonJuly,9,2010Semidei,J.,Radel,L.F.,&Nolan,C.(2001).Substanceabuseandchildwelfare:Clearlinkagesandpromisingresponses.ChildWelfare,80(2),109Slack,K.S.,Font,S.,Maguire-Jack,K.,&Berger,L.M.(2017).Predictingchildprotectiveservices(CPS)involvementamonglow-incomeUSfamilieswithyoungchildrenreceivingnutritionalassistance.InternationalJournalofEnvironmentalResearchandPublicHealth,14(10),1197Slack,K.S.,Holl,J.,Altenbernd,L.,McDaniel,M.,&Stevens,A.B.(2003).Improvingthemeasurementofchildneglectforsurveyresearch:Issuesandre-ChildMaltreatment,8(2),98Slack,K.S.,Holl,J.L.,McDaniel,M.,Yoo,J.,&Bolger,K.(2004).Understandingtherisksofchildneglect:Anexplorationofpovertyandparentingcharacteristics.ChildMaltreatment,9(4),395D.E.,Sperry,L.L.,&Miller,P.J.(2019).Reexaminingtheverbalenvironmentsofchildrenfromdierentsocioeconomicbackgrounds.ChildDevelopmentinpress.Turner,H.A.,Vanderminden,J.,Finkelhor,D.,&Hamby,S.(2019).Childneglectandthebroadercontextofchildvictimization.ChildMaltreatmentU.S.DepartmentofAgriculture(2017).Trendsinsupplementalnutritionassistanceprogramparticipationrates:Fiscalyear2010toscalyear2015.RetrievedfromU.S.DepartmentofHealthandHumanServices(2017).Childmaltreatment2016.RetrievedfromAuthor,AdministrationforChildrenandFamilies,AdministrationonChildren,YouthandFamilies,ChildrensBureauwebsite:Vachon,D.D.,Krueger,R.F.,Rogosch,F.A.,&Cicchetti,D.(2015).Dierentformsofchildmaltreatmenthavecomparableconsequencesamongchildrenfromlow-incomefamilies.JAMAPsychiatry,72(11),1135Voisin,D.R.(2007).Theeectsoffamilyandcommunityviolenceexposureamongyouth:Recommendationsforpracticeandpolicy.JournalofSocialWorkEducation,(1),51Wexler,R.(2019).LittleAJFreundshorricdeathisnotanargumentfortearingupmorefamilies.23RetrievedfromChicagoSun-TimesWhitaker,R.C.,Phillips,S.M.,Orzol,S.M.,&Burdette,H.L.(2007).Theassociationbetweenmaltreatmentandobesityamongpreschoolchildren.ChildAbuse&Neglect,31(11),1187White,O.G.,Hindley,N.,&Jones,D.P.(2015).Riskfactorsforchildmaltreatmentrecurrence:Anupdatedsystematicreview.Medicine,Science,andtheLaw,55Yang,M.-Y.,&Maguire-Jack,K.(2016).Predictorsofbasicneedsandsupervisoryneglect:EvidencefromtheIllinoisFamiliesStudy.ChildrenandYouthServicesReview,67,20Young,N.K.,Boles,S.M.,&Otero,C.(2007).Parentalsubstanceusedisordersandchildmaltreatment:Overlap,gaps,andopportunities.ChildMaltreatment,12S.A.FontandK.Maguire-Jac