Non-Hispanics with Latin American ancestry: Assimilation, race, and

17 Pages·2011·1.15 MB·English

by  Amon Emeka

SocialScienceResearch40(2011)1547–1563
ContentslistsavailableatScienceDirect
Social Science Research
journal homepage: www.elsevier.com/locate/ssresearch
Non-Hispanics with Latin American ancestry: Assimilation, race,
and identity among Latin American descendants in the US
⇑
Amon Emeka , Jody Agius Vallejo
UniversityofSouthernCalifornia,DepartmentofSociology,3620SouthVermontAvenue,KaprielianHallRoom352,LosAngeles,CA90089-2539,UnitedStates
a r t i c l e i n f o a b s t r a c t
Articlehistory: Inthe2006AmericanCommunitySurvey(ACS),6%ofrespondentswithLatinAmerican
Received28July2009 ancestry answered ‘no’ when asked whether they were Hispanic themselves.
Revised10June2011 Conventional definitions of the Hispanic population exclude such respondents as ‘not
Accepted24June2011
Spanish/Hispanic/Latino’ even though they are self-identified Latin American descen-
Availableonline1July2011
dants. Since their exclusion may bias our assessments of Hispanic social mobility, it is
important to know more about them. Non-Hispanic identification is most common
Keywords:
among Latin American descendants who (1) list both Latin American and non-Latin
Latinos
American ancestries, (2) speak only English, and (3) identify as White, Black, or Asian
Hispanics
whenaskedabouttheir‘race.’Ancestryandracialidentityareconsiderablymoreinfluen-
Racialidentity
Ethnicidentity tialthanrespondents’education,income,placeofbirth,orplaceofresidence.Thesefind-
Assimilation ings support both traditional straight-line assimilation and a more recent ‘‘racialized
Immigrants assimilation’’ theory in explaining discrepant responses to the ethnicity and ancestry
questionsamongLatinAmericandescendants.
(cid:2)2011ElsevierInc.Allrightsreserved.
1.Introduction
Intheyear2006,therewereanestimated44.1millionUSresidentswhoreportedSpanishorLatinAmericanancestry
(e.g.,Mexican,PuertoRican,Cuban,Dominican,Guatemalan,Spanish,Salvadoran,Colombian,etc.).Ofthatnumber,2.5mil-
lion(6%)answeredinthenegativewhenaskedwhethertheywere‘‘Spanish/Hispanic/Latino.’’Thisfactmayreflectapattern
of‘ethnicattrition’thathasimportantmethodologicalandtheoreticalimplicationsforthestudyofracial/ethnicidentifica-
tionandimmigrantincorporation.SincemoststudiesofHispanic1experienceandadvancementdefinetheHispanicpopula-
tion(s)onthebasisofHispanicidentityandnotLatinAmericanorSpanishancestry,those2.5millionnon-Hispanic(identifying)
LatinAmericandescendantsareroutinelyexcludedfromassessmentsofintergenerationalmobility(DuncanandTrejo,2007a)
aswellasfromprojectionsofHispanicpopulationgrowth(Golash-BozaandDarity,2008).
Strongcaseshavebeenmadethatassimilationwillnotproceedatthepacenortotheextentforrecentimmigrantsasit
didfortheEuropeanimmigrantsoftheearly20thcentury(Gans,1992;Levitt,2003;PortesandZhou,1993).Non-Hispanic
identitiesheldbyLatinAmericanorSpanishdescendants,however,mayspeaktothecentripetalforceassimilationcontin-
uestoexertonAmericanlifeinthe21stcentury.Recentstudieshavedemonstratedtheemergenceoftheunhyphenated
‘American’identitiesamongLatinos(Golash-Boza,2006;TellesandOrtiz,2008),butinthosestudiesthe‘American’label
is not necessarily to the exclusion of Hispanic or Latino identities. The question here is not why some Latin American
⇑
Correspondingauthor.
E-mailaddress:[email protected](A.Emeka).
1 Whileweunderstandthattheterm‘Latino’ispreferableissomeregards,weusetheterm‘Hispanic’throughoutthepapertominimizeconfusionaswe
discusstheimperfectrelationshipbetweenLatinAmericanancestryandHispanicidentity.ThosewhoidentifyasMexican,PuertoRican,Cuban,orsome
‘otherSpanish/Hispanic/Latino’inresponsetotheAmericanCommunitySurveyHispanicityquestionarereferredtoas‘Hispanic;’thosewhorespond‘no,
notSpanish/Hispanic/Latino’arereferredtoas‘non-Hispanic.’
0049-089X/$-seefrontmatter(cid:2)2011ElsevierInc.Allrightsreserved.
doi:10.1016/j.ssresearch.2011.06.002
1548 A.Emeka,J.A.Vallejo/SocialScienceResearch40(2011)1547–1563
descendantschoosetoidentifyas‘American’butratherwhytheychoosetoidentifyethnicallyas‘notSpanish/Hispanic/La-
tino.’Fewstudiesfocussquarelyontheissueofnon-HispanicresponsesamongLatinAmericandescendants,andthosethat
dofocusentirelyonMexicandescendants(AlbaandIslam,2009;DuncanandTrejo,2007a,2007b).However,theMexican–
AmericanexperienceisuniqueandnotnecessarilygeneralizabletootherHispanicgroupswithrespecttoidentity(Jimenez,
2008;AgiusVallejo,2009).ThisstudyisnovelinthatwefocusonpatternsofracialandethnicidentificationamongLatin
Americandescendantsofallancestries.
ThetheoreticalandmethodologicalcontributionsofthispaperwillinformdebatesontheextentandprocessesofLatin
Americanidentificationalassimilationbyusing2006AmericanCommunitySurvey(ACS)datatoanswerthequestion,what
explainstheoccurrenceofnon-HispanicidentificationamongLatinAmericandescendants?
2.Measuringhispanicidentity:problemsandpotentialities
Itshouldbenotedfromtheoutsetthatwecanneverknowthemeaningsandmotivationsrespondentsdrawonwhen
answeringsurveyquestionsregardingtheirracialandethnicidentitieswhichtendtobefluidandsituational(Harrisand
Sim, 2002; Nagel, 1994; Omi, 2001). This is no less true among Hispanic people (Eschbach and Gomez, 1998; Oboler,
1995;Rodriguez,2000)whoareoftenconfusedbyraceandethnicityquestions(Hirschmanetal.,2000;Rumbaut,2006)that
treat their Mexican, Puerto Rican, Cuban, etc., identities as ethnic but not racial (Grieco and Cassidy, 2001). The complex
natureofracial/ethnicidentityisrevealedintheUSCensusandACSquestionsregardingrace,ethnicity,andancestrywhich
oftenyieldinconsistentanswers–especiallyamongLatinAmericandescendants.
Fig.1. ExactwordingandorderingofHispanicity,race,andancestryquestionsonthe2006AmericanCommunitySurvey.
A.Emeka,J.A.Vallejo/SocialScienceResearch40(2011)1547–1563 1549
Fig. 1 displays the Hispanicity,2 race, and ancestry questions exactly as they appear on the 2006 American Community
Surveyenumerationform.Aswepointoutabove,anestimated44.1millionUSresidentshadsomeoranotherLatinAmerican
orSpanishancestrylistedinresponsetothequestion,‘‘Whatisthisperson’sancestryorethnicorigin?’’3WecallthemLatin
Americandescendantsor‘LADs’forthepurposesofthispaper.LongbeforetheyaregivenachancetoacknowledgetheirLatin
Americanand/orSpanishroots,however,theyencounterthequestion,‘‘IsthispersonSpanish/Hispanic/Latino?’’andsome2.5
millionLADssaytheyare‘‘notSpanish/Hispanic/Latino’’inresponse.Wecallthesepeople‘non-Hispanic’forthepurposesofthis
study.Itispossiblethatnon-HispanicidentitiesamongLADsregisteredintheAmericanCommunitySurveyreflect‘ethnicattri-
tion’asAlbaandIslam(2009)suggest,butitcouldbearguedthatdiscrepantanswerstotheethnicityandancestryquestions
reflectlittlemorethanaflawedsurveyinstrument.
First,thereisthematteroftheproblematicethniclabels—Spanish/Hispanic/Latino—employedonthesurvey.Discrepant
responsestotheethnicityandancestryquestionscouldstemfromalackoffamiliarityoraversionthatsomeLADshaveto
‘onesizefitsall’panethniclabelslike‘Hispanic’or‘Latino’(Rumbaut,2006;TellesandOrtiz,2008).Somemaymark‘No,not
Spanish/Hispanic/Latino’in protestofthoselabels.Butitiscrucialtonotehere thattheSpanish/Hispanic/Latinoidentity
question is a close-ended question with the response options: No, not Spanish/Hispanic/Latino; Yes, Mexican, Mexican
American, Chicano; Yes, Puerto Rican; Yes, Cuban; Yes, other Spanish/Hispanic/Latino. For Mexican descendants who are
aversetothepanethniclabelsmentionedabove,theseresponseoptionsmaypresentaquandary.Forthem,marking‘No,
notSpanish/Hispanic/Latino’isalsotosay‘no’tothe‘Yes,Mexican,MexicanAmerican,Chicano’option.Thesameistrue
ofPuertoRicansandCubans.ItseemsunlikelythateventhoseMexican,PuertoRican,andCubandescendantswithdeep-
seatedmisgivingsaboutpanethniclabelswouldchoosethe‘notSpanish/Hispanic/Latino’optionoverthe‘Mexican’or‘Puer-
toRican’or‘Cuban’options.
Wemightexpectthatifunfamiliaritywithoraversiontotheterms‘Hispanic’and‘Latino’werebehindnon-Hispanicre-
sponsesamongLADs,‘notSpanish/Hispanic/Latino’responseswouldbemorecommonamongthosewhosespecificnational
originsdonotappearasresponseoptionstotheHispanicityquestion.GuatemalansorSalvadorans,forinstance,maysay‘No,
notSpanish/Hispanic/Latino’becausetheyhavenoconnectiontoanyofthosepanethnictermsANDnoconnectiontoMex-
ican,PuertoRican,orCubanidentities.However,Table1showsthat,despitetheabsenceoftheirnationalitiesintheresponse
options,morethan98%ofGuatemalansandSalvadoransidentifyas‘someotherSpanish/Hispanic/Latino.’4
Second, there is that matter of question ordering on the survey. As Fig. 1 illustrates, the Hispanicity and race ques-
tions appear fifth and sixth on the questionnaire and ancestry is addressed after 30 intervening questions about the
quality and location of their household as well as the socioeconomic characteristics of individuals and families residing
there. This question spacing is fortuitous in that it reduces the chances that respondents identify ancestries based on
their expressed ethnic and/or racial identities rather than on the basis of their known family histories. Were all
three—Hispanicity, race and ancestry—questions asked in succession, respondents might more often answer them in
ways that were mutually corroborating but not entirely accurate. For instance, a respondent of German, Irish, and
Mexican ancestry who had chosen ‘not Spanish/Hispanic/Latino’ in response to the Hispanicity question and ‘White’
in response to the race question might be more likely to write ‘German and Irish’ in response the ancestry question
rather than choosing a combination suggestive of any Spanish or Latin American heritage. Spacing the Hispanicity, race
and ancestry questions this way may distract respondents from their own personal identities before asking them about
whotheir ancestorswere,and thisisimportantif wewanttomeasuretheassociation(or disassociation) betweeniden-
tity and ancestry as we attempt to do here.
Third,andperhapsthemostseriouspotentialproblem,isthefactthat,inmostcases,onerespondentanswerstheHispa-
nicity,race,andancestryquestionsforallhouseholdmembers.Accurateself-identification,therefore,mayonlybeavailable
forthe(typically)onepersoninthehouseholdwhocompletesthesurvey.Thereisreasontothinkthatthiswouldaffectthe
resultsofthisstudy,butitisnotclearhow.Inanycase,weavoidthisproblembyincludingonlyLADswhoarelikelytohave
completedthesurveyforthemselves.5
2 Question5inFig.1capturesethnicity,butonlyforthosewhoconsiderthemselvesMexican,PuertoRican,Cuban,or‘someotherSpanish/Hispanic/Latino.’
Forthisreason,weusetheterm‘Hispanicity’ratherthanethnicity.
3 TherearetwoadditionalgroupsofLatinAmericandescendantswhoare,bynecessity,excludedfromthisstudy.First,therewereanestimated2.7million
USresidentsidentifyingasMexican,PuertoRican,Cuban,or‘‘someotherSpanish/Hispanic/Latino’’inresponsetotheHispanicityquestionbutnotlisting
SpanishorLatinAmericanancestriesinresponsetotheancestryquestionwhobroughttheLatinAmericandescendantpopulationto46.8million.Second,itis
likelythattherewereLatinAmericandescendantswhoidentifiedassuchonneithertheHispanicityquestionnorontheancestryquestion.Theirinclusion
mightpushtheLatinAmericandescendantpopulationtoward50millionormore.Theformerofthesegroupsisexcludedfromthisstudysincenoneofthem
identifyasnon-Hispanicrenderingtheiridentitychoicesimpossibletomodel;thelatterisexcludedbecausewecannotidentifytheminthedata.Without
knowingthesizeofthelattergroupitisimpossibletosaywhattheeffectoftheirinclusionwouldbe,buttheinclusionofthefirstandsecondofthesegroups
wouldpartlyoffsetoneanothersincethe0%oftheformeridentifyasnon-Hispanicwhile100%ofthelatteridentifyassuch.
4 WemightfurtherexpectforeignbornLatinAmericanstohavelessfamiliarityandlessaffinityforpanethnictermsand,thusly,optfornon-Hispanic
identitiesmoreoftenthanUSbornLatinAmericandescendants,butTable2showsusthatthisexpectationdoesnotholdupeither.Itcouldbeargued,
conversely,thatasLatinAmericandescendantsbecomepoliticizedintheUStheytakeonreactiveethnicities(PortesandRumbaut,2001)thatdefypanethnic
labelinginfavorofmoreprecisenationality-basedidentities.Totheextentthatthishappens,itwouldprobablynotleadpeopletochoose‘notSpanish/
Hispanic/Latino’over‘Mexican’or‘PuertoRican’or‘Cuban.’
5 Weincludeonlyrespondentswhoarelistedas‘Person1’ontheASCsurvey.‘Person1’is‘‘theperson...inwhosenamethe[dwelling]isowned,being
boughtorrented.’’Heorsheisoften,butnotalways,theonewhocompletesthesurveysincerespondentsareaskedtolistinformationfor‘Person1’
immediatelyafterthequestion,‘‘Whatisyourname?’’Inanycase,ourmainconclusionsarethesamewhetherornotweapplythisselection.
1550 A.Emeka,J.A.Vallejo/SocialScienceResearch40(2011)1547–1563
Table1
Percent identifying as ‘not Spanish/Hispanic/Latino’ by ancestry and ancestral mix among Latin American descendants. Data Source: 2006 American
CommunitySurvey(Rugglesetal.,2008)
%WhoIdentifyas‘‘notSpanish/Hispanic/Latino’’ Distribution
Foreignborn(%) USborn(%) Total(%) n %
ANCESTRYbasedon1stresponseonly
Mexican 0.7 3.9 1.9 44,747 46.8
MexicanAmerican 1.9 2.7 2.6 5559 5.8
Hispanic 1.8 6.2 4.6 6632 6.9
PuertoRican 1.5 4.0 2.7 7771 8.1
Spanish 7.0 29.2 21.4 6010 6.3
Cuban 0.9 12.3 3.7 4114 4.3
Salvadoran 0.4 3.8 0.5 2366 2.5
Dominican 2.9 6.2 3.3 2447 2.6
Colombian 1.0 8.7 2.1 1769 1.9
Guatemalan 0.9 5.6 1.2 1536 1.6
Other 6.1 40.8 21.5 12,663 13.2
ANCESTRYbasedon1stand2ndresponses
SingleLatinAmericanAncestry 1.3 5.0 2.7 85,063 89.0
MixedLatinAmericanAncestry 0.5 2.9 2.0 1112 1.2
MixedLatinAm/non-LatinAmAncestry-Latino1st 16.1 27.8 25.9 5462 5.7
MixedLatinAm/non-LatinAmAncestry-Latino2nd 31.9 56.9 53.1 3977 4.2
SAMPLETOTAL 1.8 11.8 6.1 95,614 100.0
(cid:2)LatinAmericandescendentsarethoserespondentswhoacknowledgeoneormoreLatinAmerican,Spanish,orHispanicancestryinresponsetothe
ancestryquestion.RespondentswhoseresponsestoraceorethnicityquestionshavebeenimputedoralteredbytheCensusbureauareexcluded.Only
householdersareincluded.
TheparticularsoftheAmericanCommunitySurveyitselfposenoinsurmountableproblemsforthepurposesofthisstudy.
Tothecontrary,thesurveydesignexposesdiscrepantresponsestoethnicityandancestryquestionsthathavetothispoint
gonelargelyunexamined.Wecannotpossiblyknowthemotivationsofindividualsrespondinginthesediscrepantways,but
wecanuncoveragreatdealaboutwhattypesofLatinAmericandescendantsarelikelytoidentifyasnon-Hispanic.Evenif
theirresponsesareerroneousissomeway,therealquestioniswhethertheyarepredictable;socialsciencetheoryandre-
searchliteratureprovidesusreasontobelievethattheyare.
3.Literaturereview
Scholarsquestionwhetherrecentimmigrantsandtheirchildren,thevastmajorityofwhomarenon-European,areincor-
poratinginthe‘straight-line’fashionoftheirEuropeanpredecessors.MiltonGordon’s(1964)canonicaldefinitionofassim-
ilationpositsthatimmigrantsmayacculturatebyadoptingthelanguageandotherculturaltrappingsofthe‘host’society.
Acculturationisoftenfollowedbyintegrationintothecoreeducational,occupationalandeconomicstructuresleading,ulti-
mately,tointimatesocialandevenfamilialrelationsbetweennewcomersandmoreestablishedmembersofthehostsoci-
ety.Alogicaloutcomeofallofthisiswidespreadintermarriageandthedissolutionofold-countryethnicattachments.Inthis
definition,completeassimilationmeanstheerosionofethnicantipathiespartlypredicatedonthedeclineofethnicidentities
andtheriseofunhyphenated‘American’identities.
ThisexplanationholdsupinmostrespectsfortheUSimmigrantsofthenineteenthandearlytwentiethcenturies.Though
nineoutoftenhailedfromEuropeancountriestheywereinitiallyviewedasforeignanddistinctracial/ethnicgroupssuchas
Catholic,SlavicorJewish.Eventually,theracial/ethnicboundariesofthehostsocietyshiftedandblurredandtheseoncedis-
tinct‘‘racial’’groupsbecame‘‘White’’(Ignatiev,1996;Lieberson,1980;Roediger,2005).AlbaandNee(2003)haveargued
thatthemainstreamofUSsocietyhasbecomeessentiallymulticulturalandabletointegratediverseimmigrantgroupsinto
its fabric making assimilation a real possibility for 21st century immigrants. Others, however, suggest that the unprece-
dentedracialdiversityofrecentimmigrants,thelessfavorableeconomiccircumstancesthatgreetthem,andthemainte-
nance of close ties to countries of origin (by way of inexpensive and rapid communication and travel) may hinder the
assimilation of recent immigrants (Gans, 1992; Portes and Zhou, 1993; Portes and Rumbaut, 2001; Levitt, 2003). As the
US-born children of recent immigrants begin to age into adulthood in significant number, their assimilation, generally,
andindentificationalassimilation,morespecifically,remainsanopenempiricalquestion.
3.1.IdentificationalassimilationamongLatinAmericandescendantsintheUS
There are two possibilities traditional assimilation theory offers to explain non-Hispanic identities among LADs—one
views identificational assimilation as an unconscious process and the other views it as conscious and/or strategic. First,
A.Emeka,J.A.Vallejo/SocialScienceResearch40(2011)1547–1563 1551
‘‘Latino-ness’’ may come to hold no salience in the lives of some LADs. They may declare ‘‘No, not Spanish/Hispanic/
Latino’’ because they lack the knowledge, opportunities, and/or desire to live archetypal ‘Hispanic’ lives. The structural
and cultural trappings typically associated with Hispanic experience, such as residence in areas of high Latino popula-
tion concentration and/or speaking Spanish, may be foreign to them. This foreign-ness may reflect the strivings and
socioeconomic incorporation of immigrants (i.e., educational attainment and intermarriage) rather than any contempt
for tradition. Alba and Nee (2003) said it best: ‘‘Assimilation, one could say, is something that frequently enough hap-
pens to people while they are making other plans’’ (p. 282). In any case, theory would predict that LADs who evince
traditional indicators of assimilation such as mixed (Hispanic/non-Hispanic) ancestry, residence in non-Hispanic areas,
English language exclusivity, and/or high educational attainments will be more likely to identify as non-Hispanic than
others.
Some LADs, however, may make conscious efforts to distance themselves from identities and cultures associated with
their(parents’)homecountriestoavoidstigmatizationoroutofasenseofUSnationalismorpatriotism.Theymayperceive
socialandeconomiccostsassociatedwithHispanicity,broadlydefined(Bendick,1992;DarityandMason,1998),and,on
thosegrounds,attempttoevadeHispaniclabelsnomatterhowcentral(specific)Hispanicfolkwaysareintheirpersonallives
(Oboler,1995).
BothofthesepossibilitiesreflectassimilationonthepartofLatinAmericandescendedpersonswhoidentifyasnon-His-
panic.Inthefirstexplanation,LatinAmericanorSpanishancestrylosesitssalienceinthelivesofgrandchildren,children,
and,inrarecases,immigrantsthemselves.Ethnicattachmentstakeonamoreperipheralandvoluntarycharacter,especially
asimmigrantsandtheirchildrenachieveupwardmobilityoverthegenerations(Alba,1990;Waters,1990).Immigrantsfrom
LatinAmericaand,moreoften,theirUS-bornprogenymaycometoseethemselvesas‘‘Americanfirst’’andinsomecases
‘‘Americanonly.’’SuchidentitiesmayreplaceratherthanaugmentethnicattachmentsbasedonLatinAmericanancestry.
Thisistheessenceofidentificationalassimilation(Gordon,1964).
Inthesecondexplanationabove(consciousdistancing),LADswithsignificanttiestoHispanicculturesandcommunities
maybegintoperceiveprevailinganti-Hispanicsentimentsandattempttodistancethemselvesfromthegroup.Totheextent
thatanti-Hispanicaffectispartofthe‘‘host’’culture,thisdistancingmayreflectculturalassimilation—internalizingUSper-
ceptionsofHispanicsasalien,unassimalable,andundesirable.AsLopezandStanton-Salazar(2001)assert,‘‘whenchildren
learnwhatitmeanstobeMexicaninCalifornia[forinstance]theyareundergoingpreciselywhatGordon(1964)meantby
‘acculturation’’’(p.73).
Theseadaptationsmaytakeplacebothwithinandacrossimmigrantgenerations.AlbaandIslam(2009)trackintra-gen-
erationalchangesincohortsofMexican–Americansacrossthe1980,1990,and2000USCensusesnotingsubstantialpopu-
lationlossesthatcannotbeexplainedbymortalityormigration.Theymakeacompellingcasethattheselossesarelargelya
functionof‘‘identityswitching’’betweenCensuses.Thatis,substantialnumbersofrespondentswhoidentifiedthemselves
asMexican–Americanonthe1980censusfailedtodosoin1990and/or2000.Previousscholarshiphascontended,however,
that‘‘thedecisiveturningpointforchangeinethnicandnationalself-identitiescanbeexpectedtotakeplaceinthesecond,
andnotthefirst,generation’’(PortesandRumbaut,2001:150).Accordingly,Ono(2002)uncoversevenmoredramaticiden-
tificational shifts—away from Mexican or Mexican American and toward ‘American’—among third and fourth generation
Mexican–Americans. Theory tells us that identification is a matter of time; the longer a person or a group resides in the
US,themorelikelytheyaretoidentifyas‘‘American.’’
Onthesegrounds,itisreasonabletopredictthat:
Hypothesis 1. Latin American descendants who have experienced more acculturation and/or integration into the
United States’ core social and economic structures are more likely to identify as non-Hispanic than those who have
experiencedless.
Hypothesis1a. LatinAmericandescendantsofmixed(Latinoandnon-Hispanic)ancestrieswillbemorelikelytoidentifyas
non-HispanicthanthosewithonlyLatinAmericanorSpanishancestry.
Hypothesis1b. LatinAmericandescendantsbornintheUSwillbemorelikelytoidentifyasnon-Hispanicthanthoseborn
abroad.
Hypothesis1c. LatinAmericandescendantswhospeakEnglishexclusivelywillbemorelikelytoidentifyasnon-Hispanic
thanthosewhocontinuetospeakSpanish.
Hypothesis1d. LatinAmericandescendantswithhigherlevelsofeducationandincomewillbemorelikelytoidentifyas
non-Hispanicthanthosewithlowerlevels.
Hypothesis1e. LatinAmericandescendantsresidinginareasoflowerHispanicpopulationconcentrationwillbemorelikely
toidentifyasnon-HispanicthanthoseresidinginareasofhigherHispanicpopulationconcentration.
1552 A.Emeka,J.A.Vallejo/SocialScienceResearch40(2011)1547–1563
3.2.ThepossibilityofaracializedassimilationamongLatinAmericandescendantsintheUS
Contemporaryproponentsofassimilationtheoryarguethatitremainsacentripetalforcethatpullsnewcomersintoa
‘newAmericanmainstream’whosecultureandidentityaremoremulticulturalorhybridizedthan‘Anglo’incharacter(Alba
andNee,2003).Segmentedassimilationaccountsrecognizethatsomeimmigrantsandtheirdescendantsmaynotfindtheir
wayintothemainstreamsincedifferentnationaloriginsgroupsaredifferentlysituatedwithrespecttothe‘Americanmain-
stream’(PortesandRumbaut,2001).Othersargue,however,thatraceandracialidentityfigurecentrallyintosocialtrajectories
ofimmigrantsgenerally(BashiandMcDaniel,1997),andLADs,morespecifically(Golash-Boza,2006).
SomeLADsmaychoosetoidentifyasnon-Hispanicbecauseanotherracialorethnicidentityhasbecomemoresalientin
theirdailylivesthanthefactoftheirLatinAmericanorSpanishdescent.TheymaycometoidentifysimplyasWhiteorBlack
orevenAsian(seeRopp,2000);theymaycometounderstandWhite,Black,Asian,andHispanicasmutuallyexclusivecat-
egories;andthismaybearonthelikelihoodthatLatinAmericandescendantsexpressnon-Hispanicidentities.Theseideas
arecapturedinFig.2,aschematicdiagramofracializedassimilationborrowedfromGolash-BozaandDarity(2008:5).
Fig.2suggeststhattheHispanicpopulationwasraciallyheterogeneousin2000butmaybecomemoreraciallyhomog-
enousas‘HispanicWhites’gain entreintothe(non-Hispanic)‘White’populationandbegintoidentifyas such,‘Hispanic
Blacks’begintoidentifywiththe(non-Hispanic)‘Black’population,leavingmainly‘Hispanicothers’inthe‘Hispanic’group
bytheyear2050.Thispaperisuniquelypositionedtoillustratewhatmaybetheearlystagesofthisprocessbyexamining
theracialidentitiesofLADswhoidentify,explicitly,asnon-Hispanic.
Golash-Boza(2006;Golash-BozaandDarity,2008)providesevidencethattheprocessdepictedinFig.2maybeafoot.
QualitativestudiesbyLopez(2003),NewbyandDowling(2007),andothersarealsosuggestiveofaracializedassimilation
processamongLADs.InherrecentworkontheMexican-originmiddleclassAgiusVallejo(2010)recountsaconversation
aboutidentitywithVincent,acollegeeducatedsecond-generationMexicanprofessionalwhoreveals:
‘‘IgrewupinaWhiteneighborhoodandschooldistrictwithWhitefriendsandpeopleatworkhavealwaysassumedIwas
White.AndIdon’thaveanaccentandIdon’twearclothesthatwouldclassifymeasMexicanandeveryonejustthoughtI
wasItalianbecauseofthat.AndIdon’tspeakSpanishthatgoodsopeoplealwayssayI’mWhite.’’
Vincent’sMexicanancestrycouldultimatelyberelegatedtoalaundrylistoftrivialfactsinhisfamilyhistory—anethnic
option, if you will—and may be even less salient in the lives of his children. Because US residents tend to use the words
‘‘White’’and ‘‘American’’as synonyms(Feagin,2000),Vincentandothers likehimmaysometimesoptfor aWhiteracial
identityasanassertionoftheirownunhyphenatedAmericanidentity.Suchanassociationbetween(White)racialidentity
and(non-Hispanic)ethnicidentitywouldbeconsistentwithtraditionalstraightlineassimilationtheoryaswellasthe‘‘idea
thatracialstatusplaysakeyroleinimmigrantadaptation’’(Golash-Boza,2006:35).
Racializedassimilation,maymeansomethingverydifferentforLatinAmericanimmigrantsanddescendantswhoareof
primarilyAfricanandAsianancestries. Fromthemomenttheyarrive,theymayhavea racialminorityidentitiesimposed
on them. Dominicans of primarily African ancestry may often be seen simply as ‘Black’ and thereby be subjected to tra-
ditional patterns of Black exclusion (Denton and Massey, 1989; Bailey, 2001; Candelario, 2001; Lopez, 2003; Rodriguez,
2000). Much as ‘Asian’ stereotypes will often be applied to Chinese- and Japanese-descended immigrants from Mexico,
Centraland SouthAmerica,‘Black’stereotypesmaybeappliedtodark-skinnedAfro-Latinosirrespectiveoftheirnational
origins(seeOjito,2001;LeeandBean,2010).ThismayleadtosomeLADstomorereadilyidentifyasnon-Hispanicracial
minorities.
To this point we have discussed racial identities that may be associated with dis-identification from the Hispanic
groups—that is, ‘Hispanic Whites’ and ‘Hispanic Blacks’ in Fig. 2. But who are the ‘Hispanic others?’ The answer to this
questionliespartlywhotheyarenot;theyare thosewhoseemthemselvesasneitherWhitenorBlack.Thelong history
of contact between Indigenous, African, European,and Asian peoples in the Americas has left an indeliblemark on Latin
Fig.2. Adepictionofracializedassimilation.FromGolash-BozaandDarity(2008).
A.Emeka,J.A.Vallejo/SocialScienceResearch40(2011)1547–1563 1553
Americanidentities.Mestizaje—theracialandculturalhybridizationofEuropeanandindigenouspeoplesintheAmericas—
hasbeenacentralaspectofLatinAmericanidentities(Lugones,1994;Rodriguez,2007).Inatleastonerespect,then,the
Hispanic/Latino identity can be understood as fundamentally ‘‘multiracial,’’ and it is reasonable to expect that self-
identifiedHispanicsmayoftenexercisetheoptionofmarking‘oneormore’boxesonstandardsurveyquestionsregarding
their race. That ‘one or more’ may often include ‘American Indian’ or any of its variants since indigenous peoples often
figurecentrally(ifonlysymbolically)innarrativesofpeoplehoodinmuchofLatinAmerica.Therefore,itisalsoreasonable
to expect that some Hispanics identify as ‘American Indian’ (Menchaca, 1993) as an acknowledgment that some part of
their lineage is indigenous.6 In short, we might expect that non-Hispanic identities would be less prevalent among those
LADs who identify racially as multiracial and/or as ‘American Indian’ since ‘mixed’ and/or indigenous heritages are part
and parcel of many Hispanic identities.7
Still,multiracialandAmericanIndianresponseoptionsmaynotbesatisfyingforthoseLADswhounderstandthemselves
as the products of contact between European, African, and indigenous people. Many of them view their race (raza) as
‘Mexican’or‘PuertoRican’or‘Latino’—nothingmoreandnothingless.BecauseLatinAmericangroupsarenotrecognized
as‘socialraces’bythecensusbureaulikeWhites,Blacks,AmericanIndians,andAsians,the‘someotherrace’response8is
the only option for those insistent on an explicitly Hispanic or Latin American racial identity (see Landale and Oropesa,
2002). Prior research suggests that the ‘‘some other race’’ category has become the de facto Hispanic racial identity (Jones-
CorreaandLeal,1996;Rodriguez,2000).Totheextentthatthisistrue,weshouldfindthatthosewhoidentifyraciallyas‘some
otherrace’ontheracequestionarelesslikelytoidentifyas‘non-Hispanic’thanLADsofotherracialdesignations.
OurdatanotallowustospeaktothequestionofwhyLADschoosetheracialidentitiestheydo,butwecanassessthe
strengthofassociationbetweenracialidentityandHispanicidentityamongLatinAmericandescendants.Forreasonsdis-
cussedabove,wehypothesize:
Hypothesis 2. Racial identity is a statistically significant predictor of non-Hispanic identity among Latin American
descendantssuchthat;
Hypothesis2a. LatinAmericandescendantswhoidentifyraciallyasWhite,Black,orAsianwillbethemostlikelytoidentify
asnon-Hispanic;
Hypothesis2b. LatinAmericandescendantswhoidentifyraciallyasAmericanIndian,Multiracial,or‘someotherrace’will
betheleastlikelytoidentifyasnon-Hispanic.
4.Dataandmethods
Weemploydatafromthe2006AmericanCommunitySurvey9totestthehypothesesdiscussedabove.Thisdatasetwas
chosenfor(1)itssizewhichallowsfortheanalysisofsmallnationaloriginsgroups(i.e.,Cubans,Dominicans,Guatemalans,
Colombians),(2)itsrecentdateofcollection,(3)itscoverageofallgeographicareasoftheUSandinclusionoflocalareaindi-
catorswhichallowsforanalysesoftheinfluenceofgeography/communityonidentity,and(4)itsplaceinanongoingdatacol-
lectioneffortthatallowsfortemporalcomparisonswithpastandfuturesamples.Thereisnootherdatasetthatcanmatchthe
ACSonthesecriteria.
Since we are interested in how respondents see themselves we include only Latin American descendants presumed to
have responded to the Hispanicity, race, and ancestry questions for themselves and whose responses to the Hispanicity
andracequestionshavenotbeenalteredbytheCensusbureau.10Thisleaves95,614respondentsofLatinAmericanand/or
Spanishancestryofwhom5806(6.1%)declarethattheyarenotthemselves‘‘Spanish/Hispanic/Latino.’’
Thedependentvariableis‘‘Non-Hispanicidentity.’’Thesurveyasks,‘‘IsthispersonSpanish/Hispanic/Latino?’’Respon-
dentswhoanswer:Yes,Mexican,MexicanAm.,Chicano;Yes,PuertoRican;Yes,CubanorYes,otherSpanish/Hispanic/Latino
receivea‘‘0’’onthismeasure.Thosewhorespond‘‘No,notSpanish/Hispanic/Latino’’receivea‘‘1.’’Intheanalysestofollow,
thegoalwillbetofigureoutwhichcovariatespredisposerespondentstoreceivinga‘‘1’’onthismeasure.11
6 ThismaybetrueintheUSevenifLadinosandIndiosarethoughtofasseparateanddistinctgroupsinLatinAmericancontexts(Rodriguez,2000).
7 OntheACSand2000Censusenumerationformsthe‘AmericanIndian’responseisfollowedbytheinstruction‘‘Printnameofenrolledorprincipletribe’’
andablankspace.ThelargestshareofLatinAmericandescendantswhoidentifyraciallyas‘AmericanIndian’writeintribesthatarenotcodedbytheCensus
bureau—inyearspast,manysuchresponseswerecodedinto‘LatinAmericanIndian.’
8 The‘otherrace’responseisfollowedbyablanklineandtheinstruction,‘‘Printracebelow.’’
9 TheCensusBureauboastsa97.5%responseratewhichisarrivedatbydividingthe‘weightedestimateofinterviews’bythe‘weightedestimateofcases
eligibletobeinterviewed.’However,theresponserateislower(68.2%)whenyousimplydividethenumberofhouseholdsurveyscompletedbythenumberof
addressesinitiallyselected(USCensusBureau,2007).
10 6.5% of Latinos in the 2006 ACS had their responses to the race item altered by the Census Bureau. We are unable to ascertain their racial self-
identifications,andtheyarethereforeexcludedfromtheanalysestofollow.Incidentally,only1.1%ofnon-Hispanicrespondentshadtheirresponsestotherace
questionaltered.
11 Thismayseemcounterintuitive,butsincewewanttoknowwhysomeLatinAmericandescendantsidentifyasnon-Hispanic,thisoperationalizationmakes
themostsense.
1554 A.Emeka,J.A.Vallejo/SocialScienceResearch40(2011)1547–1563
4.1.Independentvariables
Theinfluenceofancestrywillbeassessedbycomparingtheprevalenceofnon-Hispanicidentityamongdescendantsof
thetenmostprevalentLatinAmericanancestries,andbycomparingthosewithonlyLatinAmericanancestriestothosewho
listsomenon-HispanicancestryinadditiontotheirLatinAmericanancestries.12Thesetwomeasuresareemployedtocapture
theindependenteffectsofmembershipinspecificLatinAmericanancestrygroupsandtheeffectsofhaving‘mixed’ancestry.
Nativityiscapturedinathree-categoryvariablewhichcomparesratesofnon-Hispanicidentificationbetweenrespon-
dentswhowerebornintheUSbutnotintheUS–MexicoborderregiontothoseborninArizona,California,NewMexico
orTexastotheforeign-born.13Englishlanguageexclusivityisadichotomousmeasureonwhichrespondentswhosaythey
speak‘Englishonly’receivea‘‘1.’’Thistreatmentoflanguageprovidesthebestchanceatidentifyingthecausalpriorityoflan-
guagesincespeaking‘Englishonly’likelyreflectstheinabilitytospeakSpanishratherthananunwillingnessthatmaybeendog-
enoustoidentity.
Theeffectsofsocioeconomicstatusareassessedherebyincludingmeasuresofeducationalattainmentandfamilyincome
relativetothepovertyline.Educationalattainmentiscombinedwithadichotomousmeasureofage(child/adult)toyielda
fivecategoryvariablethatcomparesaverysmallnumberofchildren(18-under)householderstoadultnon-graduates,high
schoolgraduates,collegedrop-outs,andcollegegraduates.Theeconomicstandingofeachrespondent’sfamilywasincluded
asacovariateinpreliminaryanalysescomparingthosewhosefamilyincomefallbelowthefederallyestablishedpovertyto
thoseinfamilieswithincomes1–2,2–3,3–4,4–5,or5timesthepovertythresholdbasedontheirfamilysizeandcompo-
sition.14Theinfluenceofgeographyonidentitywillbeexaminedbyincludingdichotomousmeasuresindicatingwhether
oneresidesintheUS–MexicoborderregionandwhetheroneresidesinaPublicUseMicrodataArea(PUMA)inwhichno
morethan5%ofresidentsidentifyas‘Spanish/Hispanic/Latino.’
Finally, Racial identity is treated here as a six-category variable to observe differences between LADs who identify as
White,Black,AmericanIndian,Asian/PacificIslander,Multiracial,orOtherinresponsetotheACSracequestion.Controls
forrespondents’ageandsexareincludedaswell.
4.2.Analyticalstrategy
Simplemeanscomparisonswillbeemployedtoobtainaroughideaofhoweachofthecovariateslistedaboveisrelatedto
identityamongLADs.Menandwomenarenotdifferentiatedintheseanalysessincepreliminaryanalysesindicatethatthey
arepracticallyidenticalintheirpatternsofidentityacrossallcategoriesoftheindependentvariableslistedinthepreceding
paragraphs.Instead,wewilldisplayratesofnon-HispanicidentificationacrossUS-andforeign-bornsamplesforallcatego-
riesofthecovariates.Next,logisticregressionwillbeusedtoascertaintheeffectsofeachoftheindependentvariablesnetof
theothers.Themultivariateanalysiswillalsoyieldamodel-chi-squarestatisticthatwillserveasabaselineforcomparingthe
explanatorypowerofeachcovariaterelativetoalltheothers.Themostnovelstepintheanalysiswillbetheexaminationof
block-chi-squarestatisticswheneachcovariateisenteredintotheequationlast.15Thiswilltellushowmuchmoreexplan-
atorypowerisgainedwhenweaddeachcovariateintoamodelwithalloftheothercovariatesalreadyinit.Thenetexplanatory
powerofeachcovariatewillbeascertainedthisway.Bycomparingtheexplanatorypowerofeachcovariatenetoftheotherswe
cancometosomeempiricallyfoundedanswersastowhicharethemostpivotalpredictorsofnon-Hispanicidentityamong
LatinAmericandescendants.
Finally,sinceMexicanimmigrantsanddescendantsmakeupmorethanhalfoftheLatinAmericandescendedpopulation
wewillrunthemultivariateanalysesseparatelyforMexicanandnon-Mexicandescendantstoseewhetheranyassociations
wemayuncoveraredrivenbytheMexicanexperienceintheUSasopposedtoasharedHispanicexperience.
5.Results
Ancestriesthatrespondentslistwhenasked‘‘Whatis[your]ancestryorethnicorigin?’’appeartobepredictiveofhow
theyanswerwhenasked‘‘[Areyou]Spanish/Hispanic/Latino?’’Table1providestheancestralcompositionoftheLatinAmer-
icandescendantpopulationinthe2006AmericanCommunitySurveysampleaswellasabivariatelookattherelationship
betweenancestryandnon-Hispanicidentity.86.8%writeinoneofthetenancestrieslistedintheupperpanelofthetable
leaving13.2%identifyingwithsmallerHispanicandnon-Hispanicancestrygroups.Forallancestriesandancestralcombi-
nations,theUS-bornarefarmorelikelytoidentifyasnon-Hispanicthanaretheforeign-born.
12 ACSrespondentsmaywriteinasmanyancestriesastheycanfitinthespaceprovidedinresponsetotheancestryquestion,butonlythefirsttwoarecoded
bytheCensusbureau.InordertobeincludedinthissampleatleastoneoffirsttwoancestryresponsesmustbeaLatinAmericanorSpanishancestry.
Therefore,allrespondentsinthesamplewhohavenon-HispanicancestryaremixedLatino/non-Hispanicdescendantswhohavelistedtheirnon-Hispanic
ancestryeitherfirstorsecondinresponsetotheancestryquestion.
13 Preliminaryanalysesdemonstratedconsistentlythat‘ageatarrival’hasnostatisticallysignificantbearingonthisoutcome.
14 Seehttp://www.census.gov/hhes/www/poverty/threshld/thresh06.htmlfor2006povertythresholdvalues.
15 AsMenard(1995)explains,commonPseudo-R2measurestreattheModelv2asanalogoustotheRegressionSumofSquaredDeviations(SSR)inordinary
leastsquaresregression.Itconstitutesthenumerator,forinstance,inMcFadden’sR2.Blockv2statisticstellushowmuchimprovementthereisinModelv2
withtheintroductionofanewblockofcovariatestothemodeland,therefore,hasadirectandeasilyinterpretableeffectontheoverallexplanatorypowerof
themodelaswedemonstratelaterinthepaper.
A.Emeka,J.A.Vallejo/SocialScienceResearch40(2011)1547–1563 1555
Thoseidentifying‘Mexican’or‘Mexican–American’ancestryconstitute,byfar,thelargestgroup(53%)inthesample.A
distantsecondare thoseidentifying‘Puerto Rican’ancestry(8%).Muchhas beenmadeofthe differences betweenimmi-
grants(andUS-bornchildrenofimmigrants)whomaintainexclusiveattachmentstoforeignnationalitiesandthosewho
hybridizetheirancestry/identityby adding‘‘-American’’(seePortesandRumbaut,2001;Waters,1999). However,inthis
sample,thoseidentifying‘Mexican–American’ancestryareonlyslightlymorelikelythanthoseidentifying‘Mexican’toiden-
tifyasnon-Hispanicwhenasked.‘Mexican–American’istheonlyhyphenatedancestryamongthetopten,butthoseUSborn
respondentswhoclaimitarelesslikelythanallbuttwogroupstoidentifyasnon-Hispanic.
Wemightexpectthatifunfamiliaritywithoraversiontothepanethnicterms‘Hispanic’and‘Latino’werebehindnon-His-
panicresponsesamongLADs,aswassuggestedearlier,suchresponseswouldbemorecommonamongthosewhosespecific
nationaloriginsdonotappearasresponseoptionstotheethnicityquestion.GuatemalansorSalvadorans,forinstance,may
say‘No,notSpanish/Hispanic/Latino’becausetheyhavenoconnectiontoanyofthosepanethnictermsANDnoconnectionto
Mexican,PuertoRican,orCubanidentities.Table1tellsus,however,thatdespitetheabsenceoftheirnationalitiesinthelistof
responseoptions,morethan98%ofGuatemalansandSalvadoransidentifyas(someother)Spanish/Hispanic/Latino.
Therearetwopanethnicresponsestotheancestryquestionthatarecommon.6.9%list‘‘Hispanic’’ancestrybeforeany-
thingelse,andanother6.3%list‘‘Spanish’’astheirprimaryancestry.Theformergroupexhibitsslightlygreaterpropensity
(4.6%)toidentifyasnon-HispanicthanthosewholistspecificLatinAmericannationalitieswhile‘‘Spanish’’descendantsare
far more likely to do so (21.4%). Preliminary analyses revealed that foreign-born LADs identifying ‘Spanish’ or ‘Hispanic’
ancestryweremostoftenbornintheAmericas—82%and94%respectively—suggestingthatitisnotimmigrantsfromSpain
drivingthesepatterns.
Table2
Percentidentifyingas‘notSpanish/Hispanic/Latino’byselectedbackgroundcharacteristicsamongLatinAmericanDescendants.DataSource:2006American
CommunitySurvey(Rugglesetal.,2008)
%WhoIdentifyas‘‘notSpanish/Hispanic/Latino’’ Distribution
Foreignborn(%) USborn(%) All(%) N %
PLACEOFBIRTH/AGEATARRIVAL
USnon-borderregions 20.4 20.4 14,749 15.4
US–Mexicoborder-region(AZ,CA,NM,IX) 6.9 6.9 25,817 27.0
Foreign-born–arrivedpriortoage6 3.9 3.9 4054 4.2
Foreign-born–arrivedatage6–12 2.1 2.1 5048 5.3
Foreign-born–arrivedatage13–18 1.4 1.4 11,390 11.9
Foreign-born–arrivedatage19+ 1.7 1.7 34,556 36.1
ENGLISHLANGUAGEEXCLUSIVITY
SpeakssomeSpanish 1.3 3.2 1.9 75,378 78.8
SpeaksEnglishonly 13.2 22.9 21.7 20,236 21.2
EDUCATION/AGE
Child(lessthan18yrsofage) 3.0 7.5 5.8 86 0.1
Adult–non-graduate 0.9 6.4 2.1 33,441 35.0
Adult–highschoolgraduate 1.8 10.5 5.9 24,763 25.9
Adult–attendedsomecollege 3.0 13.0 8.9 22,558 23.6
Adult–collegegraduate 4.0 17.0 10.9 14,766 15.4
ECONOMICSTANDING
Familyincomebelowthepovertyline 1.2 8.3 3.7 18,029 18.9
Familyincome1to2Xthepovertyline 1.4 9.5 4.0 25,441 26.6
Familyincome2to3Xthepovertyline 1.5 11.2 5.4 18,214 19.0
Familyincome3to4Xthepovertyline 2.1 11.8 6.8 11,935 12.5
Familyincome4to5Xthepovertyline 2.5 12.6 8.1 7809 8.2
Familyincomemorethan5Xthepovertyline 4.4 16.6 11.9 14,186 14.8
PLACEOFRESIDENCE
USnon-borderregions 2.3 19.6 8.4 44,147 46.2
US–Mexicoborder-region(AZ,CA,NM,IX) 1.4 6.9 4.1 51,467 53.8
Localareasw/5%ormoreLatinoresidents 1.7 9.5 4.9 87,854 91.9
Localareasw/lessthan5%Latinoresidents 4.1 31.6 18.9 7760 8.1
RACE
White 1.5 17.7 8.7 52,979 55.4
Black 20.8 38.5 28.3 1374 (cid:3)1.4
AmericanIndian/Alaskanative 2.8 8.2 6.2 872 0.9
Asian/Pacificislander 46.2 9.2 28.0 464 0.5
Multiracial 4.2 8.5 6.8 2799 2.9
Someotherrace 1.0 1.3 1.1 37,126 38.8
SEX
Men 1.7 12.6 5.8 54,521 57.0
Women 2.1 11.0 6.4 41,093 43.0
ALL 1.8 11.8 6.1 95,614 100.0
1556 A.Emeka,J.A.Vallejo/SocialScienceResearch40(2011)1547–1563
% Identifying as ‘notSpanish/Hispanic/Latino’
Educational Attainment Family Income Relative to the Poverty Line
Fig.3. PercentofLatinAmericandescendantsidentifyingas‘notSpanish/Hispanic/Latino’byeducationalattainmentandeconomicclassstanding.
ThebottompanelofTable1illustratestheimpactofmixedancestryonnon-Hispanicidentity.Aswehypothesized,those
withmixedancestriesaremorelikelytoidentifyasnon-HispanicthanthosewithSpanishorLatinAmericanancestryonly.
While less than 3% of LADs with only Latin American ancestry identify as non-Hispanic, more than a quarter (25.9%) of
mixed-ancestryLADswholistanon-HispanicancestryaftertheirLatinAmericanancestryidentifyasnon-Hispanic.More
than half (53.1%) of mixed-ancestry LADs who list a non-Hispanic ancestry before their Latin American ancestry identify
asnon-Hispanic—suggestingthatintermarriagemayexplainthepatternofnon-HispanicidentificationamongLADs.16
TherearemyriadotherfactorsthatinfluenceinternalizedandexpressedidentitiesamongLADsintheUSTable2displays
non-Hispanicidentityfigureswhenthesampleisdisaggregatedbyseveralsalientdemographicandsocioeconomiccharac-
teristics. The Place of Birth/Age at Arrival panel suggests that the US-born are more likely to identify as non-Hispanic—
especially those born outside of the US–Mexico border region (20.4%). Foreign-born LADs are substantially less likely to
identify as non-Hispanic regardless of how old they were when they immigrated (1.4–3.9%). Those born in Arizona,
California,NewMexicoorTexasconstituteanintermediatecategoryinwhich7%identifyasnon-Hispanic.Thosewhospeak
‘Englishonly’arealsofarmorelikelytoidentifyasnon-Hispanic(21.7%)thanthosewhospeakanyamountofSpanish(1.9%).
Ofcourse, ancestry,birthplace,and languageare correlatedwithoneanotherand mayinfluenceidentityby wayofeach
other.Themultivariateanalysiswillsorttheseeffectsout.
Educationandfamilyincomeappearbeassociatedwithpatternsofidentityinwaysthatareconsistentwithtraditional
assimilationtheory.Therelationshipsarenotdramaticbutareeasilydiscernable.Fig.3makesclearthataswemovefrom
lowertohigherstandingontheeducationalattainmentandfamilyincomegradients,non-Hispanicidentitiesbecomemore
prevalentamongLADs,andthisisespeciallytrueamongtheUS-born.
Geographyalsoseemstomatter.ThoseresidingintheUS–Mexicoborderregionarehalfaslikelytoidentifyasnon-His-
panic(4.1%)asthoselivingtothenorthandeast(8.4%).Amuchlargerdifference,however,isobservedatthelevelofthe
PublicUseMicrodataArea(PUMA).About8.1%ofLatinAmericandescendantsresideinPUMA’swherelessthan5%ofthe
residentsidentifyasnon-Hispanic.Theyareconsiderablymorelikelytoidentifyasnon-Hispanic(18.9%)thanthoseliving
inlocalareaswithmoreHispanicpresence(4.9%).
Finally,thoseLADswhoidentifyraciallyasBlackorAsianaremostlikelyofalltoidentifyasnon-Hispanic(28%,forboth
groups)whileWhite,AmericanIndian,andMultiracialrespondentswithLatinAmericanancestryidentifyasnon-Hispanic
lessthan10%ofthetime.Thosewhoraciallyidentifyas‘‘Other’’—arefarlesslikelytoidentifyasnon-Hispanic(1.1%)than
thosewhoidentifywiththemoreconventional/recognizableracialcategories.OntheCensusandACS,Hispanicityistreated
asethnicandnotracial.The‘‘someotherrace’’categorymaybetheonlyacceptableracialoptionforrespondentswhocon-
siderthemselvesraciallyLatinoorMexicanorPuertoRican,etc.17
16 Ofcourse,itisequallylikelythatnon-HispanicidentificationamongLADsexplainspatternsofintermarriage.
17 Respondentswhocheck‘someotherrace’inresponsetotheracequestionareprovidedspacetowriteinaspecificracialidentity,buttheseidentitiesare
notcodedandmadepublicbytheCensusBureau.ManyLatinoscheck‘‘Other’’ontheracequestionandwritewhattheyfeeltobeamoreaccurateracialor
ethnicdescriptor,buttheyremainracially‘‘Other’’inCensustabulations.