Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
*Direct all correspondenceto: Xiaoru Liu, Departmentof Sociology, College of Arts and Letters,San Diego State
University,5500 Campanile Drive, San Diego, CA 92182-4423;e-mail:rliu@mail.sdsu.edu.
Sage
is collaborating with JSTOR to digitize, preserve, and extend access to
Sociological Perspectives ®
www.jstor.org
500 SOCIOLOGICALPERSPECTIVES Volume43,Number3,2000
Age
Race
Gender
ActualLabeling ,
FamilySES \0
FamilyStructure
UrbanResidence
PriorDelinquency
PerceivedLabeling YouthDelinquency
DelinquentPeer |
Parent/Child
Relationship PeerInfluences
Attachnmnt
to Peer
Figure 1
Hypothesized
BaselineModel,Conditional ofPeerGroupsand Common
Effects
AntecedentVariables
1). This occurs especially when those who do the labeling (e.g., parents) never
communicatetheirreactions(or labels) to theyouth.Instead,theysimplychange
theirbehavior toward the persons (e.g., the youth) who are being labeled (Mat-
sueda 1992; Triplettand Jarjoura1994). Their behavior changes, such as with-
drawing theirlove or acceptance,may in turnalienate the youth,leading to sub-
sequent delinquencyinvolvement.
Following previous research(Matsueda 1992; Triplettand Jarjoura1994), the
presentanalysis focuses on parentevaluations of the youth as sources of actual
labeling. For consistency,youth perceptionof informallabeling reflectsthose of
parents only (see Triplettand Jarjourafor a similar focus). Few will deny that
parentalopinions remainone of the more importantsources of negativeinformal
reactionsforadolescents.
Empiricalstudies on theeffectsofinformallabelingon youthdelinquencyhave
yielded supportiveevidence indicatingthatinformalreactionsof parentsand/or
others are related to youth delinquency (Matsueda 1992; Ward and Tittle1993;
Zhang 1997). In addition, actual appraisals of parentshave been shown to exert
independenteffectson youth delinquencynet of theirindirecteffectsvia youth
perceivedor reflectedappraisals (Matsueda 1992).
METHOD
Sample
The hypothesized relationshipswere tested using data fromthe initial three
waves of the National Youth Survey (NYS) (Elliott,Huizinga, and Ageton 1985;
Elliott,Huizinga, and Menard 1989). The NYS is a nationalprobabilitysample of
adolescentswho were between the ages ofeleven and seventeenand who resided
in households in the United States in 1976. The original sample of 1,725 youths
was obtainedthroughmultistageclustersamplingdesign. The firstthreewaves of
the interviewswere conducted in 1977 (Wave 1), 1978 (Wave 2), and 1979 (Wave
Attritionover thesewaves was minimaland did not substantially
3), respectively.
affectthe representativenessof the sample (Elliott,Knowles, and Canter 1981). In
addition to youth response,the firstwave (1977) of the NYS containedinforma-
tionprovided by one parentof each youthwho agreed to participatein the study.
The threewaves ofyouthinterviewstogetherwithparentreportsare used to esti-
mate thehypothesizedmodels.
Among the youths who participatedin all threewaves of the study and who
provided completeinformation on thestudyvariables (N = 1,261),approximately
49 percentare femalesand 51 percentare males. Regardingrace/ethnicstatus,12
percentof the sample are AfricanAmericans,4 percentare Mexican Americans,
and the restare non-Hispanicwhitesby majority(81 percentof the sample). The
average age of the youths was approximatelyfourteenat the time of the initial
interview.At that time, approximately26 percentof the youths lived in urban
areas and 18 percentlived with parents who were divorced, separated, or had
neverbeen married.
Measures
A totalof fourteenmeasures (not countinginteractionterms)were included in
the study to test the hypothesized relationships.For ease of interpretation,
the
variables (scales) were standardizedbeforetheywere enteredinto the regression
analysis (fordetails,see Jaccard,Turrisi,and Wan 1990).
Delinquency
The dependentvariable,delinquency,is derived fromWave 3, while an identi-
cal measure fromWave 1 serves as a controlvariable.FollowingHeimer and Mat-
sueda (1994),thestudyfocuseson a twenty-eight-item scale (alpha = .88) ofgeneral
delinquencyacts,1includingparticipationin vandalism (e.g., destroyingproperty),
theft(e.g., auto theft),use of violence (e.g., attackingothers),and drug offenses
(e.g., selling drugs). To minimizepotentialbias introducedby differential distri-
bution (i.e., prevalenceand frequencyrates) of individual items,factorweightsof
componentitemswere used to constructscales priorto modelingthemas depen-
dent (or control)variables. Similar to additive indices, the weighted composite
scales have higher scores reflectinggreaterrates of offending(individual item
response ranges from1 for "never" to 9 for "2-3 times a day"). As originally
ParentalLabelingandYouthDelinquency 505
ParentalLabeling
Theindependent
variables,actualand perceivedparentallabeling,arederived
fromthefirst
twowavesoftheNYS. Actuallabeling(fromWave1) is represented
by a four-itemscale (alpha = .79) drawn fromquestions asking parents how
delinquenttheythinktheirchildis. The scale itemsincludewhethertheythink
theirchildis "a bad kid,"a kid"whogetsintotrouble,""breaksrules,"and "does
thingsagainstthelaw."Perceivedparentallabelingis takenfromWave2 and con-
sistsof identicalitems(alpha = .81) thatask youthrespondents whetherthey
thinktheirparentssee themas "a bad kid," a kid "who getsintotrouble,""breaks
rules,"and "does thingsagainstthelaw." The itemresponsesrangefrom1 to 5,
with1 indicating"stronglydisagree"and 5 indicating"strongly
agree."Higher
scoresoftheadditiveindicesindicatemorenegativelabelingbyparentsorgreater
youthperception ofparentallabeling.
InteractionTerms
Four interactiontermswere constructedby multiplyingthe standardized mea-
sures of parental labeling, actual or perceived,by the standardized measures of
peer attitudes or delinquency,respectively.Thus the interactiontermsinclude
(1) Actual Labeling x Peer Attitudes,(2) Actual Labeling x Peer Delinquency,
(3) PerceivedLabeling x Peer Attitudes,and (4) PerceivedLabeling x Peer Delin-
quency.The interaction
termsthemselvesare not standardized.Thisprocedure
followsthemethodrecommended byAikenand West(1991)foranalyzinginter-
actioneffects.Ifhypothesesare valid, these interactiontermsshould have signifi-
cant and positiveeffectson youthdelinquency.
506 SOCIOLOGICALPERSPECTIVES Volume43,Number3,2000
ControlVariables
A numberof variables relevantto labeling hypothesesare included in the cur-
rentmodel since observed relationsmay be the spurious outcome of common
antecedents.These include sociodemographicvariables such as age, race,gender,
familystructure,socioeconomic status,and urban residence (see Fig. 1). Other
variablesthatmay be commonantecedentsofparentallabelingprocessand youth
delinquencyinclude priorlevel of delinquency,priorassociationwith delinquent
peers, parent-childrelationship,and youthattachmentto peers (see Fig. 1). Prior
level of delinquency is considered since youths who offendearlier are likelyto
continueoffendingand to evoke negative labeling by parents (Matsueda 1992).
Similarly,prior association with delinquent peers should influencelater delin-
quency and evoke negative parentallabeling (Kaplan 1984). Youths who do not
maintaingood relationshipswith parentsand who are more susceptibleto peer
influencesmay also participatein delinquency and evoke negative labeling by
parents(Triplettand Jarjoura1994).
Measures of common antecedentvariables are all derived fromWave 1. Prior
delinquency (alpha = .85) and prior association with delinquentpeers (alpha =
.83) consistof identicalitemsas describedabove foryouthdelinquencyand peer
delinquency,respectively.Parent-childrelationshipmeasured by a four-itemscale
(alpha = .67) draws fromquestions asking youthrespondentshow well theyare
doing with regard to "having parents they can talk to," "having parents com-
fortthem when they are unhappy," "getting along with their parents," and
"doing things togetherwith the family."The responses range from1 indicat-
ing "not well" to 5 indicating "very well." Attachmentto peers is reflectedin
an index of fouritems (alpha = .63) that measure the importanceof "having a
special boyfriend/girlfriend,""being included in friend'sactivities,"and hav-
ing friends "ask them to spend time and do things with them" and having
"lots of dates." The responses to these items range from1 indicating "not at all
important" to 5 indicating "very important." Higher scores of the additive
indices indicate closer relationships with parents or greater attachment to
friends,respectively.
Sociodemographicvariables are measured as follows:Age is a continuousvari-
able rangingfrom11 to 17. Race is dichotomouslycoded withblack = 1 and 0 oth-
erwise. Gender is coded with male respondents assigned the value of 1 and
females the value of 0. Family SES is measured by the Hollingshead composite
scale forthe head of the family(Miller 1991). It ranges from11 to 77, with higher
scoresindicatinglower SES. Familystructureis measured by parents'maritalsta-
tus at the initialinterview.Adolescents whose parentsare not marriedand thus
live with one adult parent are assigned with the value of 1 and 0 otherwise.
Finally,urban residenceis coded with 1 forrespondentsliving in an urban area
and 0 forsuburbanor ruralsettings.
RESULTS
The resultsof the OLS regressionmodels2are presentedin accordance with the
hypothesizedconditionalvariables,thatis, (1) peer attitudestoward delinquency,
ParentalLabelingandYouthDelinquency 507
(2) peer delinquency participation,and (3) both peer attitudesand peer delin-
quency simultaneously.
TABLE 1
Mainand Interaction
Effects
ofParentalLabelingand PeerAttitudes
towardDelinquency
on YouthDelinquencyControllingforCommonAntecedent Variables
Interaction Interaction Interaction
Variables Baseline I II III
Actualparentallabeling 0.10*** 0.07** 0.10*** 0.08**
Perceivedparentallabeling 0.10*** 0.11*** 0.08** 0.10***
Peerattitudes
towarddelinquency 0.12*** 0.09*** 0.08*** 0.08***
ActualLabelingX PeerAttitudes 0.19*** 0.14***
PerceivedLabelingX PeerAttitudes 0.15*** 0.10***
Age -0.01 -0.02 -0.01 -0.02
Race(black) -0.06* -0.06* -0.05+ -0.05*
Gender(boys) 0.06** 0.08*** 0.08*** 0.09***
FamilySES 0.01 0.02 0.01 0.01
Familystructure 0.04 0.05* 0.04+ 0.05*
Urbanresidence 0.04+ 0.04+ 0.04+ 0.03
Priordelinquency 0.41*** 0.36*** 0.37*** 0.34***
Delinquentpeerassociation -0.02 -0.00 -0.00 -0.00
Parent/child
relationship 0.03 0.03 0.03 0.02
Peerattachment 0.03 0.03 0.03 0.03
R2(adj.) 0.34 0.40 0.39 0.42
Note: Standardized effectsare shown.
+p < .10; * p < .05; **p < .01; ***p < .001 (two-tailed).
508 SOCIOLOGICALPERSPECTIVES Volume43,Number3,2000
in Delinquency
PeerParticipation
as moderating
Theresultsshowingpeerdelinquency between
therelationships
parental labeling (actual and perceived) and youth delinquency are shown in
Table 2.
Again, in the baseline model, actual and perceived parentallabeling and peer
delinquencyare all positivelyrelatedto subsequent delinquencyinvolvementnet
of the control variables. The positive effectsindicate that parental labeling
observed at the earlierpoints in timeincreases subsequent delinquencyinvolve-
ment and thathaving friendsengaging in delinquency also predictsyouth self-
reportedinvolvementin delinquencyover time.The effectsof common anteced-
entvariablesare generallysimilarto thebaselinemodel reportedearlier(in Table 1).
To examinewhetherpeer delinquencymodifiesthe relationshipbetween actual
ParentalLabelingandYouthDelinquency 509
TABLE 2
Main and InteractionEffectsof ParentalLabeling and Peer Delinquency on Youth
Delinquency ControllingforCommon AntecedentVariables
Interaction Interaction Interaction
Variables Baseline I II III
TABLE 3
Mainand Interaction
Effects
ofParentalLabelingand PeerAttitudes/Peer on
Delinquency
YouthDelinquency forCommonAntecedent
Controlling Variables
Interaction Interaction Interaction
Variables Baseline I II III
Actualparentallabeling 0.10*** 0.06* 0.09** 0.07**
Perceivedparentallabeling 0.09** 0.10** 0.07* 0.08**
Peerattitudestowarddelinquency 0.07** 0.07** 0.06* 0.06*
Peerdelinquency 0.18*** 0.13*** 0.09** 0.08**
ActualLabelingX PeerAttitudes 0.11*** 0.07*
PerceivedLabelingX PeerAttitudes 0.08*** 0.07**
ActualLabelingX PeerDelinquency 0.10*** 0.08***
PerceivedLabelingX PeerDelinquency 0.10*** 0.06*
Age -0.02 -0.03 -0.01 -0.03
Race(black) -0.06* -0.06* -0.05* -0.05*
Gender(boys) 0.05* 0.07** 0.08*** 0.08***
FamilySES 0.01 0.02 0.01 0.01
Familystructure 0.04 0.05* 0.04+ 0.05*
Urbanresidence 0.04+ 0.04+ 0.04+ 0.03
Priordelinquency 0.41*** 0.34*** 0.33*** 0.30***
Delinquentpeerassociation -0.04+ -0.02 -0.03 -0.02
Parent/childrelationship 0.03 0.03 0.03 0.03
Peerattachment 0.03 0.03 0.03 0.03
R2(adj.) 0.37 0.43 0.42 0.45
Note: Standardizedeffectsare shown.
+p < .10; * p < .05; **p < .01; ***p < .001 (two-tailed).
CONCLUSION
In summary,this study sets out to address two importantresearchquestions: (1)
whetherinformallabeling by significantothersmay predictyouth delinquency
involvementand (2) whethersocial contexts,specifically, peer group context,may
modifythe outcomes of informallabeling on youth delinquency.The analyses
yielded a number of noteworthyfindingsthat are generallyconsistentwith the
earlierhypotheses.First,parentalevaluations of the youth and youthperception
of parental evaluations representtwo distinctdimensions of parental labeling.
Each predictsyouth delinquencyindependentlyof the other.The observed posi-
tive associations between parental labeling (actual and perceived) and youth
delinquencyare consistentwiththelabeling argumentthatstigmatizingreactions
of significantothersmay have a detrimentalimpacton subsequentyouthinvolve-
ment in delinquency (Braithwaite1989; Matsueda 1992; Zhang 1997). The direct
effectof actual parentallabeling on youthdelinquencynet of youthperceptionof
parental reactions is consistentwith the earlier speculation that parents may
not always communicatetheirevaluations to the youths and thus theirevalua-
tion and concurrentbehavior change are likely to influence youth behavior
directly.Meanwhile, the inability of youth perception of parental labeling to
completely mediate the effectof actual parent labeling on youth delinquency
may also indicate that youths choose to selectivelyperceive the cues of social
responses, including those reflectingparental evaluations of themselves (Mat-
sueda 1992). The selectiveobservationthus partiallycontributesto the observed
independent effectsof parents' actual and perceived labeling on youth delin-
quency respectively.
Second, peer group influences,reflectedin peer attitudestoward delinquency
and peer participationin delinquency,are shown to have modifiedthe effectsof
parentallabeling on youth delinquency.As observed,when peers are more sup-
portive of delinquency and when youths are associated with more friendswho
participatein delinquency,parentallabeling,whetherit is the objectivereactions
or the subjectiveinterpretation, is associated with greaterincreasesin subsequent
youth delinquency. These observations are consistentwith the earliercontention
(Braithwaite1989; Tittle1975) that a social a peer contextper-
context,specifically,
missive of delinquency,mightexacerbatethe negative consequences of informal
labeling and increasethe likelihoodof criminaloffendingby the individuals who
are the objectof such a labelingprocess.
It is interestingto note, in addition, that peer attitudesand peer delinquency
512 SOCIOLOGICAL
PERSPECTIVESVolume
43,Number
3,2000
NOTES
1. The robustnessofthefindingswas checkedusingthreedifferent measuresofdelin-
quency-violence(againstpersons/property), theft,and drugoffenses. Correspond-
ingly,peerattitudesand peerdelinquency weremeasuredwithitemsreflecting attitudes
towardor involvement in violence,stealing,
and drugoffenses.Overall,theresultsare
verysimilarto thosereported here,and no substantive are observedwhen
differences
alternative
measuresofdelinquency and peerinfluencesareused.
2. The resultswerealso confirmed usingcensored(Tobit)regression modelswithmaxi-
mumlikelihoodestimation. Censoredregression modelstakeaccountofpotential floor
and ceilingeffectson thedependentvariables(delinquency involvement),whichmay
causeartifactualinteraction
effects
(see Mareand Chen1986).
3. Main effectsin theinteractionmodelsareinterpreted in accordancewithJaccard, Tur-
risi,andWan(1990),whohavestatedthatthemaineffects canbe understood as average
effectsoftheindependent variableon thedependent variableacrossvaluesofthemod-
eratingvariable.Interactioneffects
indicatethattheamountofchangein theeffects of
parentallabelingon youthdelinquency (measuredin standardizedscores)is associated
witheverystandardizedunitincreasein peer attitudestowarddelinquency(or peer
involvement in delinquency).
REFERENCES
Aiken,Leona S., and StephenG. West.1991.MultipleRegression: TestingandInterpreting
Interactions.
NewburyPark,CA: Sage.
Akers, RonaldL. 1985.Deviant A SocialLearning
Behavior: Approach.Belmont, CA: Wadsworth.
Aseltine,RobertH., Jr.1995."A Reconsideration ofParentaland PeerInfluences on Ado-
lescentDeviance."JournalofHealthandSocialBehavior
36:103-21.
Bames,GraceM.,and MichaelP.Farrell.1992."ParentalSupportand Controlas Predictors
of AdolescentDrinking, Delinquency, and RelatedProblemBehaviors." Journalof
Marriage andtheFamily54:763-76.
Becker, HowardS. 1963.Outsiders:StudiesintheSociology
ofDeviance.New York:FreePress.
Braithwaite,John.1989.Crime, ShameandReintegration.Cambridge: CambridgeUniversity
Press.
Burkett, StevenR.,and CarolA. Hickman.1982."An Examination oftheImpactofLegal
Sanctionson AdolescentMarijuanaUse: A Panel Analysis."Journal ofDrugIssues
12:73-87.
Cloward,RichardA., and LloydE. Ohlin.1960.Delinquency andOpportunity.Glencoe,IL:
FreePress.
Cohen,LawrenceE., and MarcusFelson.1979."SocialChangeand CrimeRateTrends:A
RoutineActivityApproach."American Sociological
Review 44:588-608.
Elliott,
DelbertS., David Huizinga,and SuzanneAgeton.1985.Explaining Delinquencyand
DrugUse.BeverlyHills,CA: Sage.
Elliott,
DelbertS., David Huizinga,and ScottMenard.1989.Multiple ProblemYouth:Delin-
quency,SubstanceUse,andMentalHealthProblems. New York:Springer-Verlag.
DelbertS.,BrianA. Knowles,and RachelleJ.Canter.1981.TheEpidemiology
Elliott, ofDelin-
quentBehavior and Drug Use amongAmerican Adolescents.ProjectReportno. 14,A
NationalYouthSurvey.Boulder,CO: BehavioralResearchInstitute.
Foster,JackD., SimonDinitz,and WalterC. Reckless.1972."Perceptions ofStigmaFollow-
ingPublicInterventionforDelinquentBehavior." SocialProblems20:202-9.
Grasmick, Harold,and RobertBursik,Jr.1990."Conscience,SignificantOthers,and Ratio-
nal Choice:ExtendingtheDeterrence Model."LawandSociety Review 24:83741.
514 SOCIOLOGICAL PERSPECTIVES Volume 43,Number 3,2000