Академический Документы
Профессиональный Документы
Культура Документы
963|Ranking:Education&EducationalResearch7outof219
Source:2013JournalCitationReports(ThomsonReuters,2014)
TheEffectofCommunityLinguistic
IsolationonLanguageMinorityStudent
AchievementinHighSchool
1. TimothyArthurDrake1
1
1. VanderbiltUniversity,Nashville,TN
NextSection
Abstract
Researchonlanguageminoritystudentoutcomeshasrevealedsizeableandpersistentachievementgaps.
Thereasonsforthesegapsareoftencloselylinkedwithotherfactorsrelatedtounderperformance,
includinggenerationalstatus,race/ethnicity,andsocioeconomicstatus.Usingsocioculturalsecond
languageacquisitiontheoriesandcommunitylinguisticcapitalasatheoreticalframe,thisstudybuilds
upontheextantliteraturetoexaminetherelationshipbetweencommunitylinguisticisolationandlanguage
minority10thgradestudentachievementoutcomes.Ifindthatlanguageminorityachievementgapsare
threetofivetenthsofastandarddeviationinreadingandmathbutthattheeffectisattenuatedbyincreased
levelsoflinguisticisolation.Theseresultsappearrobusttoanumberofspecifications.Possiblereasonsfor
theattenuationarealsodiscussed.
achievementgap
immigration/immigrants
languagecomprehension/development
sociolinguistics
Thetopicoflanguageminority(LM)studentoutcomeshasreceivedsignificantattentionfromtheresearch
communityinrecentyears.LMstudents,orchildrenwhosenativelanguageisnotEnglish,arepredictedto
be40%oftheU.S.schoolagepopulationbythe2030s(Thomas&Collier,2002).Educationresearch,
policy,andpracticerelatedtothesestudentshavebeenprofoundlyaffectedbymajorlegalandlegislative
decisions,suchasthelandmarkLauv.NicholsSupremeCourtdecision,whichruledthatallschool
practicesthatexcludedchildrenfromtheeducationalprocessbasedonlanguagewerebanned(Moran,
2005),andtheNoChildLeftBehindActof2001,requiringallstudentstomeetminimumlevelsof
proficiencyinEnglishlanguageartsandmathematicsby2014,includingthosecategorizedaslimited
Englishproficient(Abedi&Dietel,2004).Thesedecisions,inconnectionwithdemographicgrowth,have
contributedtoagrowingconcernforLMstudentachievementoutcomes.
ThisconcernforLMstudentperformancehasbeenmostnotablyreflectedintheproliferationof
scholarship,policy,andprogramstargetingLMstudents.Largescaleassessmentdataandpolicyreportson
theachievementoutcomesofthesestudentssuggeststhattheachievementgapswithnonminority
languagepopulationsisbothsizeableandpersistent(Fry,2007;Rampey,Dion,&Donahue,2009).Often
ofcentralconcerninstudiesexaminingLMstudenttrajectoriesistheschoolcontextinwhichLMstudents
findthemselves,includingthelanguageservicesavailabletothem,theracial/ethniccompositionoftheir
schools,theiropportunitytolearn,andthequalityofinstructiontheyreceive(Rolstad,Mahoney,&Glass,
2005;Wang&Goldschmidt,1999).Theseschoolrelatedfactorsnotwithstanding,empiricalstudiessince
the1960shaveconsistentlyshownthatnonschoolfactors,includinghouseholdincome,childhoodhealth,
andparentaleducation,helptoexplainmoreofthevariationinstudentachievementthanschoollevel
factors(Battistich,Solomon,Kim,Watson,&Schaps,1995;Ladd,2012).WithregardtoLMstudents,
indicatorsrepresentingnationalorigin,timeincountry,andgenerationalstatushavealsobeenfoundtobe
importantsourcesofvariationinstudentmathandreadingperformanceattheelementaryandsecondary
levels(Callahan,Wilkinson,&Muller,2010;Reardon&Galindo,2009).
NotablymissingfromempiricalstudiessurroundingLMstudentachievementisanindicationofthe
communitylinguisticenvironmentinwhichLMstudentsfindthemselvesandits(potential)effectontheir
performance.Alonglineofresearchexaminesthesocializationprocessesthatoccurinneighborhoodsand
theireffectonminorityanddisadvantagedstudentachievement(Jencks&Mayer,1990;Wilson,1987).
Lessisknownabouttheseprocessesincommunitieswithlinguisticvariability.
SomelanguageacquisitiontheoristshypothesizethatcontactwiththeEnglishlanguagethroughnative
languagespeakersiscriticalindevelopingproficiencyinasecondlanguage(Gass,Mackey,&Pica,1998).
Thus,schoollanguagepoliciesthatisolateLMandEnglishlanguagelearner(ELL)studentsfromnative
speakershavegenerallybeenfoundtobedetrimentaltostudentoutcomes(Arias&MorilloCampbell,
2008;Gifford&Valds,2006),includingoutcomesinhighschool(Gndara&Orfield,2012;Lillieetal.,
2010).Bycontrast,theinfluenceofoutofschoolisolationonstudentperformancehasonlyrecentlybeen
examinedandsuggeststhatLMstudentsresidinginneighborhoodsofhighlinguisticisolationoutperform
theirLMpeersinareasoflowisolation(Glick,Walker,&Luz,2013).Reasonsforthismightincludethe
communityculturalcapitalfoundindisadvantagedneighborhoods(Yosso,2005).
UsingdatafromtheU.S.Censusandanationallyrepresentativesampleofhighschoolstudents,Iexamine
theextenttowhichvaryinglevelsofcommunitylinguisticisolationisassociatedwithLMstudent
achievement.InordertounderstandthewaysinwhichcommunitylinguisticisolationmightinfluenceLM
studentachievement,Idrawuponaconceptualframeworkrootedinsocioculturaltheoryoflanguage
acquisition(Poehner&Lantolf,2005).Thistheoreticaltraditionisparticularlyvaluableinhypothesizing
aboutthemechanismsthroughwhichlinguisticallyisolatedcommunitiescouldinfluenceLMstudent
performance.Furthermore,thistraditionarguesagainstscholarshipemphasizingstudentdeficits,focusing
insteadontheassetsmarginalizedpopulationsmightleveragetoimprovestudentachievement.Inthis
view,communitylinguisticisolationactsasaproxyforsociallanguageacquisitionprocessesandvarious
formsoflinguisticcapitalnoteasilycapturedbyotherquantitativeindicators.
Inthisstudy,Iaskthefollowingresearchquestion:Whatistheeffectofcommunitylinguisticisolationon
LMstudentachievementinhighschool?
PreviousSectionNextSection
PriorResearchandTheoreticalFrame
AnumberofanalystsofLMachievementoutcomeshavefocusedonstudentgenerationalstatus,school
languageprogramsandtheirimpact,andracialandethnictestscoregaps(Callahanetal.,2010;Hagy&
Staniec,2002;Portes&MacLeod,1999;Rolstadetal.,2005).Theseanalyseshaveyieldedimportant
results,buttheyhavefailedtoaccountforthelinguisticisolationofLMstudentsincontextsoutsideof
schoolandhowthatisolationmightinfluencetheirstudentachievementoutcomes.Anotableexceptionhas
beentherecentworkbyGlickandcolleagues(2013)examiningtheeffectofcommunitylinguistic
isolationonearlychildhoodcognitivedevelopment.Theirstudyfindsthatalthoughchildrenwith
immigrantmothershavelowercognitiveoutcomescomparedtotheirpeerswithnativebornmothers,
communitylinguisticisolationamelioratestheperformancegap.Thereasonsforthisattenuationareonly
brieflydiscussedbytheauthors,whoarguethat[the]findingisconsistentwiththehypothesisthatsome
communitycharacteristicsofferprotectionoratleastpartiallyamelioratetheeffectsofothernegative
communitycharacteristicsforsomechildrenofimmigrants(Glicketal.,2013,p.150).Inordertobuild
onthisfinding,IbrieflyreviewthehistoryoflinguisticisolationasameasureintheU.S.Censusandthen
turntodevelopingaframeworkforunderstandinghowcommunitylinguisticisolationmightbeassociated
withLMstudentperformance.
LinguisticIsolation
Thetermlinguisticisolationoriginatedinthe1980U.S.Censusinresponsetothegrowingneedto
communicatewithindividualswhowerebeingsupportedbylegislationandthe(amended)requirementin
theVotingRightsActrequiringtheprovisionofmaterialsinasecondlanguageforalldistrictswithmore
than5%ofasingleLM(Siegel,Martin,&Bruno,2001).Linguisticallyisolatedhouseholdsarethosein
whichalladult(ages15andolder)residentsspeakalanguageotherthanEnglishandnonereportspeaking
Englishverywell.Inthisstudy,Idefinecommunitylinguisticisolationasthepercentageoflinguistically
isolatedhouseholdsinaschoolszipcode.Broadlyspeaking,U.S.schoolslocatedinneighborhoodsof
highlinguisticisolationtendtobedominatedbyoneortwoLMgroups.Recentworkexaminingthe
segregationofLatinoandELLstudentsinArizonasuggeststhattheprocessofisolatingLMstudentsfor
languageinstructionisdetrimentaltostudentoutcomes(Gndara&Orfield,2012;Lillieetal.,2010).
Neighborhoodlinguisticisolation,however,goesbeyondbeingamerelinguisticidentifier.AsVega,Ang,
Rodriguez,andFinch(2010)argue,although[linguisticisolation]...islargelyademographicmeasureof
individualhouseholds,whichisthenaggregatedtothelevelofaneighborhood,itisanimportantproxy
markerforculturalconditionswithinaneighborhood.Linguisticisolationisamarkerforthepresenceofan
ethnicenclaveandethnicsubgrouphomogeneity.(p.115)
Assuch,neighborhoodlinguisticisolationmightbeviewedasameasureofacommunitysculturalwealth,
orcommunitylevelformsofcapitalthatincludeaspirational,navigational,resistant,familial,social,and
linguisticcapital(Yosso,2005).AlthougheachofthesemightactasaleverforimprovedLMstudent
performance,Ifocusonlinguisticcapital.Todoso,Inowbrieflyreviewsocioculturaltheoriesofsecond
languageacquisitionandtheirrelationshiptostudentperformance.
LinguisticCapitalandSocioculturalTheoriesofSecondLanguage
Acquisition
Linguisticcapitaloccursthroughcommunicativeexperiencesinmorethanonelanguageand/orstyle
(Yosso,2005)andiscloselyrelatedtosocioculturaltheoriesofsecondlanguageacquisition.Sociocultural
philosophyisbroadlyconcernedwiththerelationshipbetweenthesocialworldandhumanactivity,
especiallylearningandbehavior(Shoen,2011).Itsapplicationstoworkinsecondlanguageacquisitionare
variedbutcommonlyregardthesourceoflanguagedevelopmentasenvironmentalratherthanindividual
thatis,shapedbythesocialexperiencesinacommunityandnotbyapredeterminedcognitivecode
(Johnson,2003).
Scholarlyresearchbegantotestthesocialprocessesoflanguageacquisitionaroundtheturnofthecentury.
Donato(1994),forinstance,foundevidencetochallengetheacceptedviewthaterrorcorrectionor
scaffoldedhelpshouldbeprovidedbyamorecapableindividual(i.e.,anativelanguagespeakerorexpert).
Instead,Donatosfindingssuggestthatindividualswithsimilarlevelsofsecondlanguageproficiencyare
quitecapableofprovidingsupport.Infact,inhisstudy,muchoftheparticipantsknowledgehadbeen
acquiredthroughtheprocessofsocialinteraction,withthecollectivescaffoldinglargelydependentonthe
experienceandinteractionofallthesecondlanguagelearningparticipants(Johnson,2003).
Thatlanguagedevelopmentisasocialratherthanindividualprocesshasledsomesecondlanguage
acquisition(SLA)scholarstocallforanewmetaphorforsecondlanguagedevelopment.Ratherthanusing
onlythetermlanguageacquisition,PavlenkoandLantolf(2000)arguethatparticipationmightbean
appropriatecomplement.DrawingupontheworkofSfard(1998),theseauthorsnote,[Theacquisition
metaphor]...compelsustothinkofknowledgeasacommoditythatisaccumulatedbythelearnerandto
construethemindastherepositorywherethelearnerhoardsacommodity.InSLAsuchanapproachallows
ustoseelanguageasasetofrulesandfactstobeacquiredandpermitsustodiscusslearnerlanguageinall
itscomplexity.[Theparticipationmetaphor],ontheotherhand,obligesustothinkoflearningasaprocess
ofbecomingamemberofacertaincommunity(Sfard1998,p.6),whichentailstheabilityto
communicateinthelanguageofthiscommunityandactaccordingtoitsparticularnorms(ibid).(p.156)
Suchareconceptualizationofthelanguagelearningprocesssuggeststhatoneshouldexaminelanguage
developmentasnotonlystructuralbutcontextual,intricatelyrelatedtoissuesofaffiliationandbelonging
inacommunitysetting.Moreover,itisimportanttorecognizetheimportantrolelanguageminoritiescan
playastranslatorsintheirhomesandcommunities(Orellana,Reynolds,Dorner,&Meza,2003).This
processoftranslationservestonotonlystrengthenlinguisticskillsbutalsodevelopothercompetencies,
includingaudienceawareness,crossculturalawareness,metalinguisticawareness,andsocialmaturity,
amongothers.Importanttothecontextofthisstudy,itseemslikelythatstudentsinenvironmentswith
higherlinguisticisolationwouldbecalledonmoreoftentoprovidethesetranslationservicesandtherefore
developsomeoftheseadditionalcompetenciescompetenciesassociatedwithgreateracademic
achievement.
ThesocioculturalframeworkofSLA,therefore,hastwoimportantimplicationsforthepresentstudyofthe
potentialrelationshipbetweencommunitylinguisticisolationandstudentachievement:First,thesocial
aspectoflanguagelearning,particularlyasitrelatestothecontextualenvironmentinwhichLMstudents
findthemselves,isanimportantelementofbothlanguagelearningandhighermentaldevelopment.
Includedinthesesocialcontextsistheimportanceofcollectivescaffoldingoflearnersscaffoldingthat
includesotherLMlearners.Second,thesecontextualizedenvironmentsarenotlimitedsolelytotheformal
spacesoflanguagelearning(i.e.,theclassroom)butnecessarilyincludeparticipationinthecommunityin
whichalanguagelearnerresides.Accordingly,thisframeworkpositsthatthecommunicativeexperiences
ofLMstudentsinenvironmentsoflinguisticisolation(i.e.,thelinguisticcapital)mayservetostrengthen
LMstudentperformance.Thisstudyisdesignedtoexaminethishypothesis.
PreviousSectionNextSection
MethodologyandData
Muchoftheanalysisintheeconomicsofeducationisbaseduponthesimpleproductionmodel,where
inputs,likeschoolresources,teacherquality,andfamilycharacteristics,produceanoutcome,oftenstudent
achievement.Theseeducationproductionfunctionstypicallyemploysomeformofregressionanalysis,
withinputsgenerallydividedintothosethatcanbemanipulatedbypolicymakers,likeschoolresources,
andthosethatareoutsidepolicymakerscontrol,likestudentinnateabilityorfamilybackground
characteristics(Hanushek,1995).Themodelcouldbesummarizedasfollows:
WorksincetheColemanReporthasshownthatnonschoolinputsaccountformostofthevariationin
studentachievement(Battistichetal.,1995;Ladd,2012).
SincetheneighborhoodlinguisticisolationofLMstudentsistheprimaryvariableofinterestinthisstudy,
mygoalistodevelopaproductionmodelthathasarobustsetofnonschoolfactorsthatmightbecorrelated
withboththelinguisticenvironmentofLMstudentsandtheiracademicachievement.Failingtoaccountfor
thesevariableswillbiastheestimatesoftheeffectoflinguisticisolationupward,assumingthatcommunity
characteristics,suchaspoverty,arepositivelycorrelatedwithlinguisticisolationandnegativelyassociated
withacademicachievement.1Inordertodevelopthissetofcommunitycovariates,Imergedatafromthe
EducationalLongitudinalStudy(ELS)withthe2000U.S.Census.Unlikedatacollectedfromeitherthe
schoolordistrict(Bali&Alvarez,2004;Wang&Goldschmidt,1999)orfromsurveys(Callahanetal.,
2010),theU.S.Censusdataallowforcommunitylevelindicatorsrepresentingthepopulationof
individualsinthearea,leadingtomorepreciseestimates.
DataSets
TheELSof2002isdesignedtofollowanationallyrepresentativepanelof10thgradestudentsthrough
theirtransitionfromhighschooltocollegeandintotheworldofwork(ELS,2004).ELSisamultilevel
study,withquestionnairescollectingdatafromstudents,parents,teachers,andadministrators.Iuse
informationfromeachofthesesurveystogatherstudentinformationonschoolandfamilybackground
characteristics.Inaddition,informationontheschoolslocationviaitspostalzipcodeallowsmethe
opportunitytomergedatafromthe2000U.S.Census.2
Inordertomergethesedatasets,Imaketheassumptionthatchangesinthecommunityenvironmentduring
the2.5yearlagbetweenthecollectionoftheCensusdatainthespringof2000andthecollectionofELS
datainthefallof2002willnotbesystematicallydifferentbetweenschools.Thatis,althoughIamnot
assumingthatneighborhoodconditionsdidnotchangeduringtheperiod,Iamassumingthatthesechanges
werenotsystematic.Giventhisassumption,anychangewouldbeconsideredrandomnoiseandwouldnot
biastheestimates.Apotentialthreattothisassumptionareestimated1.5millionpeopleimmigratingtothe
UnitedStatesduringthistimeperiod(Camarota,2010).Thistrend,however,wouldlikelymakethe
estimatesoflinguisticisolationin2000toolow.Assuch,theestimatedcoefficientsonlinguisticisolation
couldrepresentalowerboundofitseffectonstudentachievement.Othercommunityfactors,including
poverty,educationlevel,andincome,arealsoassumedtobelowerin2000than2002,giventheaverage
economicandsocialconditionoftheimmigrantpopulationduringthestudyperiod(SurezOrozco&
SurezOrozco,2009).
Measures
Ioperationalizestudentacademicachievementwithmeasuresofstudentperformanceonmathandreading
tests.ThesemeasurescomefromcognitivetestsdevelopedforELSfromframeworksusedfortheNational
EducationalLongitudinalStudy(NELS:88)anditemsfrompreviousassessments,includingNELS:88,
NationalAssessmentofEducationalProgress,andtheOrganisationforEconomicCooperationand
DevelopmentProgrammeforInternationalStudentAssessment(ELS,2004).Allofthereadingquestions
and90%ofthemathquestionsweregivenasmultiplechoice,withtheother10%givenasopenended
questionsscoredaseitherrightorwrong.Testswerescoredusingbotharawnumberrightprocedureand
onebasedonitemresponsetheory.
Inputsintotheproductionmodelincludestudent,family,school,andcommunitybackground
characteristics.AlthoughIestimatethemodelsusingthefullsampleofstudents,myinterestintheLM
populationmeansthatmyselectionofinputsshouldbeparticularlysensitivetothispopulation.The
primarycovariateofinterestisthepercentagelinguisticallyisolatedintheschoolszipcode.Asmentioned
previously,neighborhoodlinguisticisolationisamarkerthatgoesbeyondthelinguisticmakeupofthe
communityandsuggestsethnichomogeneityandvariousformsofculturalcapitalthatmayberelatedto
LMstudentachievement.The2000U.S.Censusrevealsthattheracial/ethnichomogeneityofLMstudent
neighborhoodsvaried,fromneighborhoodswithhighconcentrationsofLatinostudentstothosewith
relativelyequalnumbersofAsianandLatinostudents.
Inaddition,IincludecontrolsthathavebeenparticularlyrelevanttoLMstudentachievement,including
generationalstatus(Callahanetal.,2010;Portes&MacLeod,1999),timeincountry(Fry&PewHispanic
Center,2005),percentageELLinschool(Conger,2010),andmobility(Fong,Bae,&Huang,2010;
Rumbaut,2008).Additionally,sincelinguisticisolationisatthecommunitylevel,itisimportanttoinclude
neighborhoodlevelcovariatesthatareassociatedwithbothstudentachievementandlinguisticisolationso
astobetteridentifytheeffectoflinguisticisolation.Accordingly,Ialsocontrolforcommunitypercentage
unemployment,medianhouseholdincome,percentagewithabachelorsdegree,percentagebelowthe
povertyline,andpercentagemobility,orthepercentageofthecommunity(ages5andolder)wholivedina
differenthomein1995.Otherstudentandschoolcovariatesarethosetraditionallyincludedineducation
productionfunctions,suchasstudentrace/ethnicity,socioeconomicstatus,andgender,andschool
covariates,suchasstudentenrollment,percentageofmathandreadingteacherswitha
professional/mastersdegreeorhigher,percentagefree/reducedlunch,andpercentageenrolledinremedial
math.
ModelsandAnalysis
Iestimatethefollowingmodelpredicting10thgradeacademicachievement:
whereAisrepresentsatestscoreforstudentiinsubjects,LMrepresentsifthestudentselfidentifiesasan
LMstudent,andLinguisticIsolationrepresentsthepercentageofhouseholdsintheschoolszipcodethat
arelinguisticallyisolated.TheparameterestimatesonthevariablesrepresentingLM,linguisticisolation,
andtheirinteractionwillbetheprimaryestimatesofinterestforthisanalysis.Controlsincludestudent,
school,andcommunitycovariatesdescribedintheMeasuressection.
SinceIestimatetheproductionfunctionusingordinaryleastsquares(OLS)regression,itisimportantto
notethatalthoughthesevariablescanbegroupedatdifferentlevels,eachisassumedtobepartofasingle
functionofstudentachievement.Thatis,unlikemultilevelmodelswhereeachlevelisallowedtohaveits
ownprediction,modelspredictingstudentachievementusinganOLSestimatorassumethatall
characteristics,regardlessoflevel,areattributesoftheindividual.Forexample,inthesemodels,parental
income,communityeducationlevel,andteacherqualityareasmuchapartofthestudentsidentityas
genderorrace/ethnicity.Alternativespecificationsdesignedtoaccountforthenestednatureofthedata
producedstatisticallysimilarresults.Inmyanalyses,Ialsoassumethatthemissingdataoccuratrandom,
andIusedmultipleimputationtoaddressthemissingness(Schafer,1999).Theamountimputedisreported
inAppendixTableA1.Furthermore,inordertoaccountforthesurveydesign,IusefirstorderTaylor
serieslinearizedestimatesofthestandarderrors(Efron,1981).3
PreviousSectionNextSection
Results
Table1containsalistofthedescriptivestatisticsforthevariablesusedinthisanalysis,conditionedonLM
status.Notsurprisingly,theLMstudentpopulationinthissampleispredominatelyLatinoandAsian,with
nearlyhalfbeingfirstgenerationimmigrants.ComparedwiththeirnonLMpeers,theseLMstudentscome
frompoorerhomes,schools,andcommunitiesandarefoundinneighborhoodswithnearlytripletherateof
thenumberofLMindividualsandalmost3.5timesthepercentageoflinguisticallyisolatedhouseholds.
StudentperformanceonthecognitivetestsshowsthatLMstudentsperformathalfastandarddeviation
lowerthantheirnonLMpeersinbothreadingandmath.
Viewthistable:
Inthiswindow
Inanewwindow
DownloadtoPowerPointSlide
Table1
DescriptiveStatistics,byLanguageStatus
Table2showstheresultsfromtheOLSestimationofthetwomodelspredictingstudentachievement.As
commoninthelargerscholarlyliteratureonracial/ethnicachievementgaps,thesemodelsshowthat
AfricanAmerican,Latino/a,NativeAmerican,andmultiracial(nonHispanic)studentsarepredictedto
scoreanywherefromtwotosevententhsofastandarddeviationlowerthantheirWhitepeersinbothmath
andreading,allelsebeingheldequal.Studentsocioeconomicstatusisalsoastrongpredictorofvariation
inachievement,asisgenerationalstatus,timeincountry,andmobility.
Viewthistable:
Inthiswindow
Inanewwindow
DownloadtoPowerPointSlide
Table2
ResultsFromOrdinaryLeastSquaresEstimator,byDependentVariable
HighlightedinboldinTable2arethethreeprimarycoefficientsofinterest.Asindicatedinthesample
descriptivestatisticsandafterconditioningonimportantcovariatesofstudentachievement,LMstudents
arepredictedtoscoreanywherefromthreetofourtenthsofastandarddeviationinmathandbetweenfour
andfivetenthsofastandarddeviationinreadinglowerthantheirnonLMpeers.4Byincludingthe
interactionbetweenLMstatusandlinguisticisolationinthemodel,thecoefficientonlinguisticisolation
representstheeffectofcommunitylinguisticisolationonnonLMstudents.Appropriately,noneofthese
coefficientsisstatisticallysignificantfromzero,suggestingthatafterconditioningonotherfactors,the
linguisticenvironmentofnonLMstudentsdoesnotcontributetotheirvariationinstudentachievement.
SimilartothedirectionpositedbythetheoreticalframeworkandconsistentwiththeworkofGlickand
colleagues(2013),theinteractiontermofLMandlinguisticisolationindicatesthatastheleveloflinguistic
isolationinacommunityincreases,LMstudentperformanceimproves(ceterisparibus),andtheresultis
statisticallysignificantfromzero.Thecoefficientestimatesonthesetermsforbothreadingandmath,
althoughsignificant,arerelativelysmall.Togetasenseofthemagnitude,theestimatesrepresent4.2%and
2%oftheLMachievementgap,respectively.
Nonetheless,accordingtotheGaussMarkovassumptions,inthepresenceofafunctionalform
misspecificationoromittedvariables,theOLSestimateswillbebiased(Wooldridge,2009).Inorderto
examinetherobustnessoftheresult,Iaddresseachofthesepossibilitiesinthefollowingsections.
FunctionalFormSpecifications
Inthissection,Iconsidertwoalternativefunctionalformspecifications.Equation1imposesalinear
relationshipbetweenlinguisticisolationandLMstudentachievement;thatis,foreverypercentageincrease
incommunitylinguisticisolation,thereisanincrementalandconstantresponseinstudenttestscores.
Nevertheless,itisalsopossiblethattherelationshipisquadratic,withhigherlevelsoflinguisticisolation
producinganexponentialchangeinstudentachievement.Columns1and3ofTable3highlightthe
parameterestimatesofinterestforbothreadingandmathscoresaccordingtothishypothesized
relationship.5Althoughtheestimatedcoefficientonthequadratictermsandtheirinteractionsaresmalland
nodifferentfromzero,theparameterestimatesontheinteractionoflinguisticisolationandLMstatus
representsabout11%and7%ofthemathandreadinglanguageminoritygaps,aresultconsistentwiththe
linearrelationshipproposedabove.
Viewthistable:
Inthiswindow
Inanewwindow
DownloadtoPowerPointSlide
Table3
AlternativeOrdinaryLeastSquaresSpecifications,SelectedCoefficients
Yetbothoftheabovespecificationspositadoseresponserelationship.Itispossible,however,thatthe
relationshipismorediscrete,whereinLMstudentsinenvironmentswithhighlevelsoflinguisticisolation
differfromLMstudentsinenvironmentswithlowlevels.Inordertooperationalizethisconstruct,I
decidedtocreatetwodummyvariables,oneindicatinghighlevelsoflinguisticisolationandonelow.More
specifically,Itookobservationsinthetopandbottomquartiles,respectively,andassignedthemavalueof
1forhighandlowlinguisticisolation;allobservationsintheinterquartilerangewereassignedazero.I
thenrerantheoriginalmodelwithfullcontrols,replacingthecontinuousvariableoflinguisticisolation
withthesediscretemeasures.Columns2and4ofTable3,reporttheestimatedresults.Accordingtothese
models,LMstudentsresidingincommunitieswithlowlevelsoflinguisticisolationarenodifferentthan
theirLMpeerslivingincommunitieswithaveragelevelsoflinguisticisolation.Thecoefficientestimates
ontheinteractiontermofLMandhighlinguisticisolation,however,arestrikingLMstudentsin
environmentsofhighlinguisticisolationarepredictedtoscoreanaverageofover3pointshigherinmath
andalittleover1.5pointsinreadingthantheirobservationallyequivalentLMpeersinenvironmentsof
averageisolation.Togetasenseofmagnitudeofthiseffect,theseestimatesrepresent31%and68%ofthe
readingandmathlanguageminoritytestscoregaps(Figure1).
Viewlargerversion:
Inthispage
Inanewwindow
DownloadtoPowerPointSlide
Figure1.
Selectedachievementgapsfromordinaryleastsquaresestimatormodelingdiscreteresponserelationship,
bydependentvariable
SelectionBias
TheseresultssuggestthatobservationallyequivalentLMstudentsinenvironmentsofhighlinguistic
isolationoutperformtheirLMpeers.Thediscreteresponsemodelsprovidefurtherevidencethatthe
relationshipisespeciallynoticeablewhencomparingLMstudentsinenvironmentsofhighlinguistic
isolationtothoseinaverageorlowenvironments.Thisseemstosuggestthattheseenvironmentsallowfor
greaterlanguageparticipationandmoreabundantformsofculturalwealthassetsLMstudentscan
leveragetoimprovetheirachievement.Inarrivingatthisconclusion,however,itisimportanttothink
aboutthetypesofLMstudentsthatmayberesidinginenvironmentsofhighlinguisticisolation.Put
simply,arethesestudentssystematicallydifferentthantheirLMpeers?Ifso,how?
Inconsideringthispossibility,itisimportanttofirstaccountforthediversityoftheLMstudentpopulation.
Withtherecentadventofrichdistrict,state,andnationaldatabasesandcomputertechnology,empirical
workhasbeenabletoadequatelyincorporatethisdiversityintopredictivemodelsofstudentachievement.
ThetwomostcommonwaysinwhichtheheterogeneityoftheLMpopulationhasbeenaddressedisby
incorporatinginformationaboutstudentgenerationalstatusandrace/ethnicity.PortesandMacLeod(1999),
forinstance,useddatafromNELS:88andmultivariateregressiontopredicttheacademicachievementof
LMstudentsbyethnicityandgenerationalstatus.Theirfindingsshowlargegapsinachievementby
race/ethnicityand,toalesserextent,generationalstatus.ReardonandGalindo(2009)furtherdisaggregate
thedataonLatinostudentsbyfourconditioningvariables,includingnationalorigin,generationalstatus,
languagespokenathome,andsocioeconomicstatus.Usingavarietyofestimationstrategies,theauthors
estimatetheLatino/aWhiteachievementgapinkindergartenthroughfifthgradebyeachofthesevariables
andfindthattheresultsareconditionaluponthesemoderators.FloresandPark(2013)furtherdemonstrate
withpropensityscorematchingthatstudentsvaryinglevelsofexposuretoELLservicesisalsoassociated
withvariationsintheirachievement.
Thus,inconsideringthediversityoftheLMstudentpopulationinthisanalysis,eachmodelholdsconstant
race/ethnicityandgenerationalstatus;however,itispossiblethatthepositiveeffectoflinguisticisolation
onstudentachievementisdrivenbycertainLMsubpopulations.Forinstance,maybethisislessastory
aboutthebroaderLMstudentpopulationandmoreoneaboutLatinos/as,thelargestLMstudentgroup.Or
maybethisisreallyastoryaboutassimilationandacculturation,whereresearchisstartingtoestablishan
immigrantparadoxwithkeyoutcomes,likehealthandachievement,becomingworseovertime(Coll
&Marks,2012).Perhapsthesecommunitieswithhighlinguisticisolationarepredominatelyfirst
generationenclaves,placeswherefamiliesandcommunitiescankeepmanyofthecharacteristicsofthe
nativecountryandcultureandmaintainbettersocialandacademicoutcomes.
Toexaminethesepossibilities,Tables4and5reporttheOLSparameterestimatesforLMstatus,linguistic
isolation,andtheinteractionbetweentheseforthethreedependentvariablesofinterest,withseparate
regressionsreportedbyrace/ethnicity,generationalstatus,andlanguageusedathome.Thesemodelsshow
apositiverelationshipontheinteractiontermforbothmathandreadingoutcomes,althoughthereduced
samplesizeandsmallcoefficientsmakemanyoftheestimatesnodifferentfromzero.Ofparticularnoteis
thepositiveimpactforWhiteandfirstgenerationLMstudentsacrossallofthemodels.BlackLMstudents
aretheonlyonesthatareadverselyaffectedbytheseenvironments,andtheestimateisstatistically
significant.Themodelsthatemployadiscreteresponserelationshiphavesignificantlylargerstandard
errorsduetothereducedsamplesizebutshowasimilarpattern. 6Noneofthemodelsestimatedby
differencesinstudentlanguageyieldsstatisticallysignificantresults.Fromthesemodels,itwouldappear
thattheoriginalresultisrobusttothediversityoftheLMstudentpopulation.Exceptforthecaseof
AfricanAmericanLMstudents,eachofthesesubgroupsexhibitsanestimatedcoefficientinthepositive
direction.
Viewthistable:
Inthiswindow
Inanewwindow
DownloadtoPowerPointSlide
Table4
SelectedCoefficientsFromSeparateOrdinaryLeastSquaresRegressions(DependentVariable=Math),by
Race,GenerationalStatus,andLanguage
Viewthistable:
Inthiswindow
Inanewwindow
DownloadtoPowerPointSlide
Table5
SelectedCoefficientsFromSeparateOrdinaryLeastSquaresRegressions(DependentVariable=Reading),
byRace,GenerationalStatus,andLanguage
Asecondconsiderationmightbetoconsidertheroleofstudentinnateabilityorintelligence.Often,models
concernedaboutthisincludepriorstudenttestscoresasaproxyforthisgeneticendowmentofintelligence
(Todd&Wolpin,2003).Inthisstudy,itseemsunlikelythatstudentswho,onaverage,havehigherinnate
intelligencewouldchoosetosystematicallyanddisproportionatelysettleinenvironmentswithhigh
linguisticisolation.
Athirdandfinalconsiderationthatcouldcontributetothesedifferencesrelatestotheschoolcontext.
Specifically,ifschoolslocatedinneighborhoodswithhigherlinguisticisolationarebetterabletoserve
theirLMstudentpopulations(i.e.,throughbetterlanguageprogramsorsupports),thentheinteraction
effectsoflinguisticisolationandLMcouldbepickinguptheeffectoftheschoolsandnotthecommunity
context.Morespecifically,itseemsplausiblethatschoolsmightrespondtohighlevelsofLMstudentsby
improvinglanguageservicesandsupports.Whatismore,itisimportanttorememberthatthesestudents
areinhighschoolandmayhavebeenexposedtobetterservicesthroughouttheireducationalcareers.
Importantly,ifschoolsarerespondingtoneighborhoodconditions,thentheeffectoflinguisticisolation
wouldbeindirectthatis,theneighborhoodcontextaffectstheschoolsprovisionofservices,whichthen
affectsstudentachievement.
Totheextentpossible,Itriedtocontrolforthispossibilityinmodelsthatincludedcontrolsforthe
proportionofstudentsreceivingbilingualandEnglishlanguageservices.7Theresultsarerobusttothese
inclusions.Similarly,whereassomestates,likeTexasandCalifornia,haveahistoryofworkingwithlarge
LMstudentpopulations,others,likeKentuckyorWyoming,havehadrelativelysmallLMpopulations.
Modelsincludingstatefixedeffectswouldcontrolfortheseinterstatedifferences,butIfindthattheyalso
yieldsimilarresults.
PreviousSectionNextSection
Discussion
Studiesoninschoolstudentcompositionoftenfocusonthetensionbetweenschoolsegregationand
integration,belongingnessanddiversity,andthebenefitsofstudentrecognitionamonglowerperforming
peers(i.e.,frogpondmodels)andnormativemodelsofstudentunderperformance(Anderman,2002;
Brenner&Crosnoe,2011;Goldsmith,2010).Thebroaderresearchliteratureonschoollevelsegregation
demonstratesthatschoolswithhigherlevelsofminoritysegregationtendtoachieveandattainlessthan
schoolswithhighconcentrationsofWhitestudents(Goldsmith,2009;Hanushek,Kain,&Rivkin,2002).
Theracial/ethnicachievementgapsfoundinthisdatasetareatestamenttothisfindingmanyminority
groups,includingAfricanAmerican,Latino,NativeAmerican,andmultiracial(nonHispanic)students,are
predictedtoscoreanywherefromtwotosevententhsofastandarddeviationlowerthantheirWhitepeers
inbothmathandreading,evenaftercontrollingforimportantindividual,family,school,andcommunity
characteristicsknowntobeassociatedwithvariationinstudenttestscores.Likewise,theLMtestscoregap
suggeststhatthesestudentsareamongthelowestperformingstudentsubgroups,holdingtheseother
factorsconstant.
Inthisstudy,however,Ifocusontheneighborhoodlinguisticcontextsinwhichschoolsresideandtheir
relationshiptoLMstudentachievement,nettheeffectoftheseinschoolfactors.Theresultssuggestthat
observationallyequivalentLMstudentsinenvironmentsofhighlinguisticisolationoutperformtheirLM
peers.Thediscreteresponsemodelsprovidefurtherevidencethattherelationshipisespeciallynoticeable
whencomparingLMstudentsinenvironmentsofhighlinguisticisolationtothoseinaverageorlow
environments.ThatcommunitylevellinguisticisolationattenuatesthisperformancegapinLMstudent
achievementsuggeststhatweneedtomorecarefullyconsidercommunitylevelprocessesaffectingLM
studentperformance,especiallyinlieuoftheotherworkreviewedabovethathighlightsthedetrimental
effectsofK12minoritystudentsegregationandisolation.Whatisitaboutthesecommunitiesthat
generatespositivecrosssectionalachievementforLMstudents?
Tobegin,itisimportanttoconsiderlinguisticisolationasatheoreticalconstructthatrepresentsmorethan
thepercentageoflinguisticallyisolatedadultswithinagivengeography;rather,linguisticisolationcanbe
operationalizedasaproxyformanyotherimportantculturalandlinguisticattributesofacommunity(Vega
etal.,2011).InhypothesizingabouttheseattributesandtheirpotentialaffectonLMstudentachievement,I
developedaframeworkrootedinsocioculturaltheoriesoflanguageacquisitionandcommunitycultural
wealthimportanttheoreticaltraditionsbasedinassetbasedmodelsofminoritystudentachievementand
LMlanguagedevelopment(Villenas&Deyhle,1999).Thesetheoriessuggestthatenvironmentswithhigh
linguisticisolationallowforgreaterlanguageparticipationwithothernovicespeakersandmoreabundant
opportunitiestoactastranslators,aprocessthatstrengthenslinguisticskillsanddevelopsothercore
competencies,likeaudienceawareness,crossculturalawareness,metalinguisticawareness,andsocial
maturity,amongothers.Assuch,theselinguisticexperiencescanbeviewedasaformoflinguisticcultural
capital.Otherformsofcommunityculturalcapital,includingfamilial,aspirational,navigational,andsocial
(Yosso,2005),havealsobeenfoundtobeassociatedwithLMstudentachievement.8
Inemployingthisframework,Ialsohighlighttheneedtoconsiderthemechanismsthatmightoperatefor
thebenefitofLMstudents.Thatis,simplyadvocatingforgreaterlinguisticisolationinLMstudent
communitiesisnottheappropriate(orethical)policyresponse.Asmentionedpreviously,increasing
minoritystudentsegregationoftenresultsinlowerstudentachievement.Instead,thisworkhighlightsthe
needtomoredeeplyexplorethemechanismsatplayinthemostlinguisticallyisolatedU.S.communities
communitieswhereLMtestscoregapsarereducedby31%and68%inreadingandmath,respectively.
Therefore,futureworkcouldbegintoexaminethequalitativedifferencesinschoolandneighborhood
contextsbyleveloflinguisticisolation.Thoughthisnationallyrepresentativedatasetprovideasnapshotof
thisrelationshipatonepointintimeandforonespecificgradelevel,qualitativeworkwouldbeableto
morecloselyinvestigatethisrelationshipinspecificcontextsandlocalesamongLMstudentsandtheir
familiesthatcanbeinterviewed,surveyed,andobserved.Suchworkwouldbeanimportantnextstepin
uncoveringtheinteractionbetweenhome,community,andschoolenvironmentsandtheirrelationshipto
LMstudentachievement.
PreviousSectionNextSection
Conclusion
ThedemographicexpansionoftheLMstudentpopulationhasprofoundimplicationsforK12education
policyandpracticeintheUnitedStates.Toaconsiderableextent,thesechangesarelinkedtonew
accountabilitystructuresthatplaceemphasisonstudentachievement.Thedataexaminedinthispaperoffer
anumberofinsightsabouttheLMstudentpopulationandtheirperformance.Ofprimaryimportanceisthe
findingthatLMstudentslagwellbehindtheirnonLMpeersinmeasuresofstudentperformanceinmath
andreading.Consistentthroughoutalltheanalysesarethelargeandstatisticallysignificantdifferences
betweenthesetwogroupsofstudents.Theresultsarealsofairlyconsistentacrossrace/ethnicity,
generationalstatus,andlanguagespokenathome.Thattheseresultsarerobusttotheinclusionofimportant
individual,school,family,andcommunitybackgroundcharacteristicsisatestamenttothesizeofthisgap.
Withtheinclusionofthe2000Censusdata,Ialsoestimatedmodelsthatincorporatedimportant
communitylevelcovariatesthathavebeenpreviouslyexcludedfrompriorresearchpredictingLMstudent
highschoolperformance,includingtheleveloflinguisticisolationfoundinschoolneighborhoods.The
resultsareconsistentwithpreviousworkthatfoundanattenuatingeffectofcommunitylinguisticisolation
onLMearlychildhoodcognitiveoutcomes(Glicketal.,2013).Theseresultsarerobusttoavarietyof
specificationsandseemtobeparticularlyrelevantforLMstudentsresidingincommunitieswithhigh
levelsoflinguisticisolation.
Theframeworkdevelopedabove,highlightingworkonthesocioculturalframeworkofsecondlanguage
acquisitionandlinguisticcapital,describesavarietyofmechanismsthatmightcontributetothisdifference.
Centraltoallofthesemechanismsistheassumptionthatcommunitylinguisticisolationrepresents
communitytraitsthatgobeyondlinguisticcharacteristics.Totheextentthattheyarecorrelatedwith
specificindividual,school,andcommunitycharacteristicshasbeenaccountedforintheempirical
estimationstrategy.Includedamongthesearestudentgenerationalstatus,language,andtimeintheUnited
States;familysocioeconomicstatus;andvariousschoolandcommunitycharacteristics,includingmobility,
medianhouseholdincome,andpercentageofresidentswithabachelorsdegreeorhigher.Nonetheless,
muchoftheprocessesdescribedinthetheoreticalframeworkisunobservable.Forexample,studenttraits,
likemotivation,aspiration,orresiliencetraitsoftenassociatedwithachievementcouldbe
disproportionatelyfoundamongLMstudentsresidinginsettingsofhighlinguisticisolation.Whichoneof
these(orother)characteristicsthatmightcontributetothisdifferencecannotbeempiricallydetermined
fromthisstudy.
Despitethislimitation,thisstudyseemstoprovideempiricalsupportfortheassetbasedframework
championedinagrowingbodyofcriticaltheoryandethnographicresearchonLMandotherminority
studentpopulations(Villenas&Deyhle,1999;Yosso,2005).Morespecifically,theresultssuggestthat
thereissomethingaboutneighborhoodswithhighlinguisticisolation(i.e.,environmentsthatmightbe
viewedasdetrimentaltostudentlanguagedevelopmentandgrowth)thatisgenerativefortheLM
populationandnotforthegeneralpopulationofstudents.Examiningwhetherornotthesedifferencescan
beaccountedforbycommunityenvironmentswithrichformsofcommunityculturalwealthandrobust
sociallanguageacquisitionprocessesorbyschoolsrespondingtospecificneighborhoodlinguisticisolation
throughtargetedinterventions,policies,andpracticescanbeanimportantnextstepintheresearchonLM
studentoutcomes.
PreviousSectionNextSection
Appendix
Viewthistable:
Inthiswindow
Inanewwindow
DownloadtoPowerPointSlide
TableA1
MissingValues
PreviousSectionNextSection
Notes
1Othercommunitycharacteristics,suchaslevelofeducation,willalsobiasthecoefficienton
linguisticisolationupwardassumingthatareaswithhigherlevelsoflinguisticisolationwillbe
negativelyassociatedwitheducationlevel.Thedatausedinthisanalysesseemtosupportthese
assumptions.
2TheU.S.CensusBureaucreatedazipcodefilein1999asabyproductofinternaloperations.
Givenpublicdemandforzipcodeinformation,theCensusBureaucreatedzipcodetabulation
areas(ZCTAs)forthe2000Census.TheZCTAwasdesignedtoovercomethegeographicand
logisticaldifficultiesinusingU.S.postalzipcodeinformation,sincezipcodesoftencross
geographicallines.ZCTAsarecreatedbygatheringallofthezipcodeinformationbycensus
blockandreportingthezipcodethatismostfrequentlyoccurring.Giventhismethodology,school
zipcodesmightberepresentedbyazipcodethatisnotthemostfrequentlyoccurringand
thereforewouldnotbeapartofthecensusdata.Inthesampleusedforthisanalysis,216
observations,representingstudentsfrom11schoolsor1.5%ofthesample,fellintothiscategory
andcouldnotbematched.MoreinformationonZCTAscanbefoundat
http://www.census.gov/geo/ZCTA/zcta.html.
3Allofthemeasuresinthisstudyareweightedwiththebaseyearstudentweight.These
weightsaredesignedtoaccountfortheunequalprobabilityofselectionintothestudy.Sincethe
EducationalLongitudinalStudyoversampledLatino/aandAsianstudents(onesubgroupof
particularconcerntothisstudy),theweightprovidesanimportantadjustmenttothebiases
inherentinthesampleselectionprocessthatwouldotherwiseunderestimatevarianceestimates
andstandarderrors,producinglargertvaluesthatincreasethechanceforType1errorduring
hypothesistesting.Thesecalculationsalsoaccountfortheclusteredandstratifiednatureofthe
surveydesignbyusingfirstorderTaylorserieslinearapproximationsofthestandarderror.
4LargergapsformatharealsofoundinlongitudinaldatafromLatinos/as(Reardon&Galindo,
2009).StudiesonmathassessmentandEnglishlanguagelearneroutcomesiswelldocumentedin
theliterature(e.g.,Abedi&Lord,2001;Menken,2008).
5Theseparameterestimatescomefromestimationsofthefullmodelwithallcontrols.Please
contacttheauthorforthecompletetables.
6Resultsavailableuponrequest.
7Pleasecontactauthorfortablesincludingthesecontrols.
8Formoreontherelationshipbetweentheseotherformsofcapitalandlanguageminority
studentachievement,seethefollowing:aspirational(Gndara,1982,1994;Raleigh&Kao,2010),
navigational(Gonzales&Padilla,1997),familial(DelgadoGaitan,2001;Moll,Amanti,Neff,&
Gonzalez,1992;Morris,1999),andsocial(Portes,2000;Zhou,1992).
ReceivedAugust14,2013.
RevisionreceivedJanuary29,2014.
RevisionreceivedJuly20,2014.
AcceptedJuly23,2014.
Author
TIMOTHYARTHURDRAKE,MEd,isaPhDstudentatPeabodyCollege,VanderbiltUniversity,PMB
#414,230AppletonPlace,Nashville,TN372035721;tim.drake@vanderbilt.edu.Hisresearchfocuseson
Latinoeducationpolicy,schoolleadership,anddatadrivendecisionmaking.
PreviousSection
References
1.
1. AbediJ.,
2. DietelR.(2004).ChallengesintheNoChildLeftBehindActforEnglishlanguage
learners.CRESSTPolicyBrief,7.Retrievedfrom
http://faculty.weber.edu/mtungmala/ed4250/Articles/Chal%20in%20the%20NCLB
%20Act.pdf
2.
1. AbediJ.,
2. LordC.(2001).Thelanguagefactorinmathematicstests.AppliedMeasurementin
Education,14(3),219234.
CrossRefWebofScience
GoogleScholar
3.
1. AndermanE.M.(2002).Schooleffectsonpsychologicaloutcomesduringadolescence.
JournalofEducationalPsychology,94(4),795809.
CrossRefWebofScience
GoogleScholar
4.
1. AriasM.B.,
2. MorilloCampbellM.(2008).PromotingELLparentalinvolvement:Challengesin
contestedtimes.EastLansing,MI:GreatLakesCenterforEducationResearchand
Practice.Retrievedfromhttp://epsl.asu.edu/epru/documents/EPSL0801250EPRU.pdf
5.
1. BaliV.A.,
2. AlvarezR.M.(2004).Theracegapinstudentachievementscores:Longitudinalevidence
fromaraciallydiverseschooldistrict.PolicyStudiesJournal,32(3),393415.
CrossRefWebofScience
GoogleScholar
6.
1. BattistichV.,
2. SolomonD.,
3. KimD.,
4. WatsonM.,
5. SchapsE.(1995).Schoolsascommunities,povertylevelsofstudentpopulations,and
studentsattitudes,motives,andperformance:Amultilevelanalysis.American
EducationalResearchJournal,32(3),627658.
Abstract/FREEFullText
7.
1. BennerA.D.,
2. CrosnoeR.(2011).Theracial/ethniccompositionofelementaryschoolsandyoung
childrensacademicandsocioemotionalfunctioning.AmericanEducationalResearch
Journal,48(3),621646.
Abstract/FREEFullText
8.
1. CallahanR.,
2. WilkinsonL.,
3. MullerC.(2010).Academicachievementandcoursetakingamonglanguageminority
youthinUSschools:EffectsofESLplacement.EducationalEvaluationandPolicy
Analysis,32(1),84117.
Abstract/FREEFullText
9.
1. CamarotaS.A.(2011).Arecordsettingdecadeofimmigration,2000to2010.
Washington,DC:CenterforImmigrationStudies.
GoogleScholar
10.
1. CollC.G.E.,
2. MarksA.K.E.(2012).Theimmigrantparadoxinchildrenandadolescents:Isbecoming
Americanadevelopmentalrisk?Washington,DC:AmericanPsychologicalAssociation.
GoogleScholar
11.
1. CongerD.(2010).Immigrantpeersinschoolandhumancapitaldevelopment.
Unpublishedmanuscript.Retrievedfrom
http://umdcipe.org/conferences/Maastricht/conf_papers/Papers/Immigrant_Peers_in_Sc
hool_and_Human_Capital_Development.pdf
12.
1. DelgadoGaitanC.(2001).Thepowerofcommunity:mobilizingforfamilyandschooling.
Boulder,CO:RowmanandLittlefield.
GoogleScholar
13.
1. DonatoR.(1994).Collectivescaffoldinginsecondlanguagelearning.InLantolfJ.P.,
AppelG.(Eds.),Vygotskianapproachestosecondlanguageresearch(pp.3356).
Westport,CT:Ablex.
GoogleScholar
14.
1. EfronB.(1981).Nonparametricestimatesofstandarderror:thejackknife,thebootstrap
andothermethods.Biometrika,68(3),589599.
Abstract/FREEFullText
15. ELS:2002baseyeardatafileusersmanual.(2004).Retrievedfrom
http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2004405
16.
1. FloresS.M.,
2. ParkT.J.(2013).Race,Ethnicity,andcollegesuccessexaminingthecontinued
significanceoftheminorityservinginstitution.EducationalResearcher,42(3),115128.
Abstract/FREEFullText
17.
1. FongA.B.,
2. BaeS.,
3. HuangM.(2010).PatternsofstudentmobilityamongEnglishlanguagelearnerstudents
inArizonapublicschools(IssuesandAnswersReport,REL2010No.093).Washington,
DC:U.SDepartmentofEducation,InstituteofEducationSciences,NationalCenterfor
EducationEvaluationandRegionalAssistance,RegionalEducationalLaboratoryWest.
GoogleScholar
18.
1. FryR.(2007).HowfarbehindinmathandreadingareEnglishlanguagelearners?
Washington,DC:PewHispanicCenter.
GoogleScholar
19.
1. FryR.,&PewHispanicCenter.(2005).Thehigherdropoutrateofforeignbornteens:
Theroleofschoolingabroad.Washington,DC:PewHispanicCenter.
GoogleScholar
20.
1. GndaraP.(1982).Passingthroughtheeyeoftheneedle:HighachievingChicanas.
HispanicJournalofBehavioralSciences,4(2),167179.
Abstract
21.
1. GndaraP.(1994).Choosinghighereducation:EducationallyambitiousChicanosand
thepathtosocialmobility.EducationPolicyAnalysisArchives,2(8),143.
GoogleScholar
22.
1. GndaraP.,
2. OrfieldG.(2012).SegregatingArizonasEnglishlearners:AreturntotheMexican
room?TeachersCollegeRecord,114(9).
GoogleScholar
23.
1. GassS.M.,
2. MackeyA.,
3. PicaT.(1998).TheRoleofinputandinteractioninsecondlanguageacquisition:
Introductiontothespecialissue.ModernLanguageJournal,82(3),299307.
CrossRefWebofScience
GoogleScholar
24.
1. GiffordB.R.,
2. ValdsG.(2006).ThelinguisticisolationofHispanicstudentsinCaliforniaspublic
schools:Thechallengeofreintegration.YearbookoftheNationalSocietyfortheStudyof
Education,105(2),125154.
GoogleScholar
25.
1. GlickJ.E.,
2. WalkerL.,
3. LuzL.(2013).Linguisticisolationinthehomeandcommunity:Protectionorriskfor
youngchildren?SocialScienceResearch,42,140154.
CrossRefMedline
OrderarticleviaInfotrieve
GoogleScholar
26.
1. GoldsmithP.R.(2009).Schoolsorneighborhoodsorboth?Raceandethnicsegregation
andeducationalattainment.SocialForces,87(4),19131941.
Abstract/FREEFullText
27.
1. GoldsmithP.R.(2011).Colemanrevisited:Schoolsegregation,peers,andfrogponds.
AmericanEducationaResearchJournal,48(3),508535.
Abstract/FREEFullText
28.
1. GonzalezR.,
2. PadillaA.M.(1997).TheacademicresilienceofMexicanAmericanhighschoolstudents.
HispanicJournalofBehavioralSciences,19(3),301317.
Abstract
29.
1. HagyA.P.,
2. StaniecJ.(2002).Immigrantstatus,race,andinstitutionalchoiceinhighereducation.
EconomicsofEducationReview,21(4),381392.
CrossRefWebofScience
GoogleScholar
30.
1. HanushekE.A.(1995).Educationproductionfunctions.InInternationalencyclopediaof
economicsofeducation(pp.277281).Oxford,UK:Elsevier.
GoogleScholar
31.
1. HanushekE.A.,
2. KainJ.F.,
3. RivkinS.G.(2002).NewevidenceaboutBrownv.BoardofEducation:Thecomplex
effectsofschoolracialcompositiononachievement(No.w8741).Cambridge,MA:
NationalBureauofEconomicResearch.
GoogleScholar
32.
1. IngelsS.J.,
2. PrattD.J.,
3. RogersJ.E.,
4. SiegelP.H.,
5. StuttsE.S.(2004).EducationLongitudinalStudyof2002:Baseyeardatafileusers
manual(NCES2004405).Washington,DC:NationalCenterforEducationStatistics.
Retrievedfromhttp://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2004405
33.
1. JencksC.,
2. MayerS.(1990).Thesocialconsequencesofgrowingupinapoorneighborhood.In
LynnL.E.,McGearyM.G.(Eds.),InnercitypovertyintheUnitedStates(pp.111186).
Washington,DC:NationalAcademyPress.
GoogleScholar
34.
1. JohnsonM.(2003).Aphilosophyofsecondlanguageacquisition.NewHaven,CT:Yale
UniversityPress.
GoogleScholar
35.
1. LaddH.F.(2012).Educationandpoverty:Confrontingtheevidence.JournalofPolicy
AnalysisandManagement,31(2),203227.
CrossRefGoogleScholar
36.
1. LillieK.E.,
2. MarkosA.,
3. EstrellaA.,
4. NguyenT.,
5. PeerK.,
6. PerezK.,
7. ...WileyT.G.(2010).Policyinpractice:TheimplementationofstructuredEnglish
immersioninArizona.LosAngeles:CivilRightsProject/ProyectoDerechosCiviles,
UniversityofCalifornia.
GoogleScholar
37.
1. MenkenK.(2008).Englishlearnersleftbehind:Standardizedtestingaslanguagepolicy.
Tonawanda,NY:MultilingualMatters.
GoogleScholar
38.
1. MollL.C.,
2. AmantiC.,
3. NeffD.,
4. GonzalezN.(1992).Fundsofknowledgeforteaching:Usingaqualitativeapproachto
connecthomesandclassrooms.TheoryIntoPractice,31(2),132141.
CrossRefGoogleScholar
39.
1. MoranR.F.(2005).Undonebylaw:TheuncertainlegacyofLauv.Nichols.BerkeleyLa
RazaLawJournal,16,110.
GoogleScholar
40.
1. MorrisJ.(1999).Apillarofstrength:AnAfricanAmericanschoolscommunalbonds
withfamiliesandcommunitysinceBrown.UrbanEducation,33(5),584605.
Abstract/FREEFullText
41.
1. OrellanaM.F.,
2. ReynoldsJ.,
3. DornerL.,
4. MezaM.(2003).Inotherwords:Translatingorparaphrasingasafamilyliteracy
practiceinimmigranthouseholds.ReadingResearchQuarterly,38(1),1234.
CrossRefWebofScience
GoogleScholar
42.
1. PavlenkoA.,
2. LantolfJ.P.(2000).Secondlanguagelearningasparticipationandthe(re)construction
ofselves.InLantolfJ.P.(Ed.),Socioculturaltheoryandsecondlanguagelearning(pp.
155177).Oxford,UK:OxfordUniversityPress.
GoogleScholar
43.
1. PoehnerM.E.,
2. LantolfJ.P.(2005).Dynamicassessmentinthelanguageclassroom.LanguageTeaching
Research,9(3),233265.
Abstract/FREEFullText
44.
1. PortesA.(2000).Socialcapital:Itsoriginsandapplicationsinmodernsociology.In
LesserE.L.(Ed.),Knowledgeandsocialcapital(pp.4367).Boston,MA:Butterworth
Heinemann.
GoogleScholar
45.
1. PortesA.,
2. MacLeodD.(1999).Educatingthesecondgeneration:Determinantsofacademic
achievementamongchildrenofimmigrantsintheUnitedStates.JournalofEthnicand
MigrationStudies,25(3),373396.
CrossRefGoogleScholar
46.
1. RaleighE.,
2. KaoG.(2010).Doimmigrantminorityparentshavemoreconsistentcollegeaspirations
fortheirchildren?SocialScienceQuarterly,91(4),10831102.
CrossRefGoogleScholar
47.
1. RampeyB.D.,
2. DionG.S.,
3. DonahueP.L.(2009).NAEP2008:Trendsinacademicprogress(NCES2009479).
Washington,DC:NationalCenterforEducationStatistics.
GoogleScholar
48.
1. ReardonS.F.,
2. GalindoC.(2009).TheHispanicWhiteachievementgapinmathandreadinginthe
elementarygrades.AmericanEducationalResearchJournal,46(3),853891.
Abstract/FREEFullText
49.
1. RolstadK.,
2. MahoneyK.,
3. GlassG.V.(2005).Thebigpicture:Ametaanalysisofprogrameffectivenessresearch
onEnglishlanguagelearners.EducationalPolicy,19(4),572594.
Abstract
50. RumbautR.G.(2008).Thecomingofthesecondgeneration:Immigrationandethnicmobility
inSouthernCalifornia.AnnalsoftheAmericanAcademyofPoliticalandSocialScience,620(1),
196236.
Abstract/FREEFullText
51.
1. SchaferJ.L.(1999).Multipleimputation:Aprimer.StatisticalMethodsinMedical
Research,8(1),315.
Abstract/FREEFullText
52.
1. SfardA.(1998).Ontwometaphorsforlearningandthedangersofchoosingjustone.
EducationalResearcher,27(2),413.
Abstract/FREEFullText
53.
1. SiegelP.,
2. MartinE.,
3. BrunoR.(2001).Languageuseandlinguisticisolation:Historicaldataand
methodologicalissues.Washington,DC:U.S.CensusBureau.Retrievedfrom
https://www.census.gov.edgekey.net/hhes/socdemo/language/data/census/lifinal.pdf
54.
1. ShoenL.T.
(2011).Conceptualandmethodologicalissuesinsocioculturalresearchandtheorydevelopment
ineducation.InMacInerneyD.M.,WalkerR.A.,LiemG.A.D.(Eds.),Socioculturaltheoriesof
learningandmotivation:Lookingback,lookingforward(pp.1142).NewYork,NY:Information
Age.
GoogleScholar
55.
1. SurezOrozcoC.,
2. SurezOrozcoM.M.(2009).EducatingLatinoimmigrantstudentsinthetwentyfirst
century:PrinciplesfortheObamaadministration.HarvardEducationalReview,79(2),
327340.
GoogleScholar
56.
1. ThomasW.,
2. CollierV.(2002).Anationalstudyofschooleffectivenessforlanguageminoritystudents
longtermacademicachievement.Washington,DC:CenterforResearchonEducation,
Diversity,andExcellence.
GoogleScholar
57.
1. ToddP.E.,
2. WolpinK.I.(2003).Onthespecificationandestimationoftheproductionfunctionfor
cognitiveachievement*.EconomicJournal,113(485),F3F33.
CrossRefWebofScience
GoogleScholar
58. U.S.CensusBureau.(2001).Census2000summaryfile.Washington,DC:Author.Retrievedfrom
http://www.census.gov/prod/cen2000/doc/sf1.pdf
59.
1. VegaW.A.,
2. AngA.,
3. RodriguezM.A.,
4. FinchB.K.(2011).NeighborhoodprotectiveeffectsondepressioninLatinos.American
JournalofCommunityPsychology,47(1),114126.
CrossRefMedline
OrderarticleviaInfotrieve
Webof
ScienceGoogleScholar
60.
1. VillenasS.,
2. DeyhleD.(1999)Criticalracetheoryandethnographieschallengingthestereotypes:
Latinofamilies,schooling,resilienceandresistance.CurriculumInquiry,29(4),413
445.
CrossRefWebofScience
GoogleScholar
61.
1. WangJ.,
2. GoldschmidtP.(1999).Opportunitytolearn,languageproficiency,andimmigrantstatus
effectsonmathematicsachievement.JournalofEducationalResearch,93(2),101111.
CrossRefWebofScience
GoogleScholar
62.
1. WilsonW.J.(1987).Thetrulydisadvantaged:Theinnercity,theunderclass,andpublic
policy.Chicago,IL:UniversityofChicagoPress.
GoogleScholar
63.
1. WooldridgeJ.M.(2009).Introductoryeconometrics:Amodernapproach.Cincinnati,
OH:SouthWestern.
GoogleScholar
64.
1. YossoT.J.(2005).Whoseculturehascapital?Acriticalracetheorydiscussionof
communityculturalwealth.RaceEthnicityandEducation,8(1),6991.
CrossRefGoogleScholar
65.
1. ZhouM
.(1992).NewYorksChinatown:Thesocioeconomicpotentialofanurbanenclave.Philadelphia,
PA:TempleUniversityPress.
GoogleScholar