Bowling Green State University
ScholarWorks@BGSU
Sociology Faculty Publications
Sociology
3-1986
Friendships and Delinquency
Peggy C. Giordano
Bowling Green State University - Main Campus, [email protected]
Stephen A. Cernkovich
M. D. Pugh
Follow this and additional works at: http://scholarworks.bgsu.edu/soc_pub
Part of the Sociology Commons
Repository Citation
Giordano, Peggy C.; Cernkovich, Stephen A.; and Pugh, M. D., "Friendships and Delinquency" (1986). Sociology Faculty Publications.
Paper 2.
http://scholarworks.bgsu.edu/soc_pub/2
This Article is brought to you for free and open access by the Sociology at ScholarWorks@BGSU. It has been accepted for inclusion in Sociology
Faculty Publications by an authorized administrator of ScholarWorks@BGSU.
Friendships and Delinquencyl
PeggyC. Giordano,StephenA. Cernkovich,and M. D. Pugh
BowlingGreenState University
in thewaysin which
Major delinquencytheoriesdiffersignificantly
theyhave portrayedthe friendship
patternsof male and, morerecently,femaledelinquents.Psychologicalstudiesand controltheory
have depicteddelinquents'peer relationshipsas inadequateor exploitiveand cold, whereassubculturaltheoriesgenerallyemphasize
the intimacyand solidarityof the delinquentgang. In spite of its
pivotaltheoreticalrole, few studieshave actuallyexamineddelinand theways in whichtheydifferfromthoseof
quents'friendships
moreconforming
adolescents.Multipledimensionsoffriendship
are
identified
thatallow examinationof adolescents'perceptionsof the
rewardsand vicissitudesof theirrelationshipsand the patternsof
interactionand influencethat characterizethem.The data reveal
many similaritiesin the friendshippatternsof adolescentswith
in theirlevels of self-reported
significantdifferences
delinquency
involvementand challengethe conceptionof femaledelinquentsas
sociallydisabled.These data suggestthatboththe"cold and brittle"
the naand "intimatefraternity"
images may have oversimplified
and differences
tureofdelinquents'friendship
relations.Similarities
in the friendshipstylesof black and white respondentsare also
examined.
Peer relationshipshave been centralto the logic of most delinquency
theories.There is generalagreementthatdelinquencyoccursmostoften
withina groupcontext,butthereis muchless consensusaboutthenature
and qualityoftherelationships
delinquentshave withtheirfriends.Early
subculturaland differential
associationtheoristsemphasizedthe delinquent group'sprimarycharacterand reliedon such termsas "solidarity"
and "espritde corps"to describeit. This viewpointis perhapsmostfirmly
anchoredin theworkofThrasher,who describedtheemotionalcloseness
1 This research was supported by PHS Research grant no. MH 29095-03, National
Instituteof Mental Health, Center forStudies of Antisocialand Violent Behavior. We
wish to thank the anonymous reviewers of AJS for their helpful comments on an
earlier version of this manuscript. Requests for reprintsshould be sent to Peggy
Giordano, Departmentof Sociology, Bowling Green State University,Bowling Green,
Ohio 43403.
? 1986 by The Universityof Chicago. All rightsreserved.
0002-9602/86/9105-0005$01
.50
1170
AJS Volume 91 Number 5 (March 1986): 1170-1202
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
and intimacyoftheChicago gangshe studied.He saw a rapportthatwas
sometimesso completethatone "receivestheimpressionof interpenetrationof personalities,if such a mysticalconceptionis possible"(1963, p.
210). In Shaw's case study,The Jackroller,Stanleyechoes this view in
describinghis own relationshipwith fellowjackrollers:"We were like
brothersand would stick by each otherthroughthick and thin. We
cheered each other in our troublesand loaned each other dough. A
mutual understandingdeveloped, and nothingcould break our confidence in each other"(Shaw 1966, p. 66).
Recentinvestigators
have criticizedthisimageofthedelinquentgroup
as reflecting
and their
"the vicariousgratifications
of adult investigators
own childhoodfantasiesto a greaterextentthantheydo theperspectives
ofgangmembers"(Shortand Strodtbeck1965,p. 231). Perhapspartlyin
reactionagainsttheearlierrosyimagery,Shortand Strodtbeckoffereda
morecomplexportrait.While agreeingthatthe delinquentgroup"likely
offeredtheseyoungstersa largermeasureof . . . play and interpersonal
gratification
than any alternativeformof associationof whichtheyare
aware and whichis available to themby virtueof preparationand other
realityconsiderations"(p. 233), theystressedthe lack of interpersonal
skillsof many gang boys as part of theirmore generaltheoryof social
disability:"Even withinthegang,uponwhichtheboycomesto be dependentfora large share of interpersonal
in many
interaction
gratification,
respects is not rewardingand lacks characteristicsessential to the
fulfillment
of these[interpersonal]
needs" (Short1963,p. xlii). Klein and
Crawford(1967) also depicted the gang's cohesion as fragileand as
generatedmoreby externalforces(e.g., threatsfromrivalgangs)thanby
personalregard.
But it is controltheorythat departsmore completelyfromthe early
subculturalview; indeed,muchofitsuniquenessas a theoretical
position
is derived fromthe way in which it differsfromearlierexplanations
preciselyon the peer issue. Hirschiin particularbelievesthatthe causal
significance
of friendships
has been overstated,arguingthat"sincedelinquents are less stronglyattached to conventionaladults than non-
delinquents they are less likely to be attached to each other....
The idea
thatdelinquentshave comparatively
warm,intimatesocial relationswith
each other(or withanyone)is a romanticmyth"(1969, p. 159). Instead,
theserelationships
are describedas "cold and brittle"(p. 141). Psychological treatments
ofdelinquency,althoughdiffering
in etiologicalemphasis,
oftendescribedelinquents'peer relationsin a similarfashion.Hartup,in
a recentreview of researchon peer relations,declared unequivocally:
"Delinquency among adolescentsand young adults can be predicted
mainlyfromone dimensionof earlypeer relations. . . notgettingalong
withothers"(1983, p. 165).
1171
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
The issueofthenatureand qualityofdelinquents'peerrelationsis thus
largelyunresolved,but it keeps surfacingbecause of its theoreticaland
applied implications.To the degreethat delinquents'relationshipsare
foundnotto be in anyrealsenseprimary,cohesive,or solidary,thecausal
of groupprocessesin the etiologyof delinquentbehavioris
significance
findingin thesocial psychobelievedto be minimized.A well-entrenched
literatureis that highlycohesive groupsare more
logical/experimental
fromtheirmembersthanare moreephemeralor
able to exactconformity
ones.2
looselystructured
Thrashernotonlydescribedtheintimacyhe believedto be characteristo the
tic of the Chicago gang boys but also linkedthesecharacteristics
gang'sabilityto exertinfluenceon its members:"The individualmember
of a gang is almostwhollycontrolledby theforceof groupopinion.The
justification
way everybodyin thegangdoes or thinksis usuallysufficient
or dissuasion for the gang boy. In such cases he is reallyfeelingthe
pressureof public opinionin that part of his own social worldwhichis
mostvital to him and in whichhe wishesto maintainstatus"(1963, p.
204). Using the same logic, Hirschi reaches opposite conclusions:"It
seemsreasonableto concludethatpersonswhose social relationsare cold
and brittle,whose social skillsare severelylimited,are incapable of influencingeach otherin the mannersuggestedby thosewho see the peer
groupas the decisivefactorin delinquency"(Hirschi1969, p. 141).
In additionto its pivotalrolein theoriesof delinquencycausation,the
image of the delinquentas sociallydeficientand lackingclose relationships also has implicationsfortreatment.The emphasis on the social
maladjustmentof delinquents,in at least implicitcomparisonwithother
betweenthe two
adolescents,sustainsthe view of intrinsicdifferences
groups(see Matza 1964). While this conceptionfitswell with current
juvenilejustice treatmentstrategiesthatcontinueto emphasizeindivid-
2 In the real world, it is probable that there is somethinga littleless orderlythan a
linear relation between intimacy and influence.Glaser (1956) emphasizes situations
where individuals may be most influencedby groups of which they are not even
members.One's location in the group (whetheras a centralor marginalactor) can also
affectthe way in which the group seeks to exert control(Giordano 1983). Ridgeway
points out that, under some conditions, cohesiveness, while increasing conformity
pressures on members, "at the same time can potentiallyincrease their freedomto
rebel against those pressures" (1983, p. 105). Nevertheless, we agree with the basic
premisethat some amount of intimacyin general enhances the group's abilityto exert
influencebecause (1) the more cohesive the group, the higherthe level of interaction
and communicationwithinit-this maximizes the opportunitiesforgroupmembersto
expresstheirviews of things-and (2) the more attractiveindividuals findtheirmembershipin a group, the more theymay be willingto accede to such influenceattempts
in order to maintain or enhance theirstanding.
1172
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
ual counseling,psychologicalclassification
schemes,and so on, it has yet
to be demonstratedadequatelyas an empiricalreality.
We agree with Empey, who argued in 1967 and again in 1983 that
"definitive
researchon the precisecharacterof delinquent,as contrasted
withconventional,groupsis desperatelyneeded"(1982,p. 274). Previous
empiricalattemptsto examinethe qualitiesof friendships
of delinquents
are less than definitiveforseveral reasons:(1) therehas been a kind of
cavalier and interchangeable
use of potentiallyverydifferent
friendship
processes. Psychological studies often equate "peer relations" with
sociometricrank (i.e., popularity),usuallywithinthe classroomsetting.
Hirschi(1969) developedthe pivotalconcept"attachmentto peers,"but
thisis variouslyreferred
to as "sensitivity
to others,""loyalty,""warmth
and intimacy,""cohesion,""solidarity,"and "dependenceon peers." (2)
Many oftheearlierstudiesreliedprimarily
on directobservationofdelinquent gangs. This methodology
allowed a detailedand dynamicanalysis
ofgangbehavior.However,in theabsenceofmeaningful
controlgroups,
it is impossibleto determinewhetherthe relationalqualities observed
(positiveor negative)would not also be foundwithinotheradolescent
friendship
networks.(3) More recentsurveys,whileprovidinga comparison across different
levels of self-reported
delinquency,have reliedon
verynarrowmeasurementstrategies,usuallytappingonlyone aspectof
peerinvolvement(e.g., timespentwithfriends)or skirting
thefriendship
relationsentirely.3
Anothertypical measurementstrategyconfoundspeer relationships
withdelinquencyinvolvementitself,focusingon theextentto whichthe
youthis involved with delinquentpeers (see, e.g., Short 1957; Jensen
1972; Elliott,Huizinga, and Ageton1982). A relatedmethodhas been to
comparethesubjects'levelsofself-reported
delinquencydirectlywiththe
reportedlevels of their friends'involvements(e.g., Poole and Regoli
1979). These studiesare importantin thattheyusuallydemonstrate
high
levels of behavioralconcordancewithinfriendship
networks.We know
thatdelinquentsspend considerabletimewithotherdelinquents,but the
qualitiesof theirrelationsremainunclear.
Whileresearchon the relationshipbetweenadolescentmale friendship
processesand delinquencyis, at theleast,inadequate,thereare almostno
to
lack ofattachment
aboutdelinquents'
3 Hirschi(1969),e.g., based his conclusions
on responsesto two items(theamountof respectforbestfriends'
friendsprimarily
wouldliketo be thekindofpersonthattheirbest
respondents
opinionsand whether
friendsare). Wiatrowski,Griswold,and Roberts(1981),in anothertestof control
of
offriendsand theimportance
theory,
reliedon two itemsaboutthe"importance"
ofdelinquency,
spendingtimewithfriends.Because theseitemswerenotpredictive
todelinquent
youth"(p.
offriends
"theunimportance
theyconcludedthatthisreflected
535).
1173
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
data about females.The few descriptivestudiesof gangs that include
femalesas membersare old (e.g., Bowker, Gross, and Klein's [1980]
analysisof unpublished1964 data) or verynarrow,usuallycenteringon
females'marginalparticipationin male gangs or groups(Miller 1973).
Nevertheless,the femaledelinquentis even morelikelythan her male
counterpartto be placed on the sociallydisabled list: "The need to 'belong' is as greatas hungerand thirst.Yet to the youngsterwho already
has problems-either because of belongingto a discriminatedracial
groupor because of innerproblemsor because of a difficult
familyrelationship-the way to friendshipgroupsis almost totallyclosed or the
hurdlesso greatshe cannottake them"(Konopka 1966, p. 88).
Due to her basic "incapacityforfriendshipwith contemporaries"
(p.
123), the delinquentgirlwill cope withherincrediblelonelinessby temin thecrowd"(p. 123)or creatinga love relationporarily"losing[herself]
ship to fillan interpersonal
void. Rittenhouse(1963) providedsome empiricalsupportforthisview in herfindingthat"relationalstrivings"
were
morecharacteristic
ofthedelinquentgirlsshe studied,while"statusstriving"was moretypicalofdelinquentboys.Wattenberg
(1956),in an examinationof case filesof boy and girl"repeaters,"foundthatin theirrelationswithpeers,boys "were moreoftenreportedto be activein games,
membersof gangs and gettingin troublewiththeirgangs. More of the
girlsquarreledwiththeirpeersand were 'lone wolves"' (p. 143). Campbell, in a cogentreview of existingfemaledelinquencyliterature,concludesthat"thereis a completeabsenceofanytheoretical
formulations
to
explaingroupdelinquencyamonggirls.... Most writerson thesubject
have proceededfromtheassumptionthatdelinquentgirlsare isolatesand
misfits"(1980, p. 380).
CONCEPTUALIZING FRIENDSHIP
Our goal hereis to examinecharacteristics
ofthefriendships
ofmale and
femaleadolescentswho varyin the extentof theirinvolvementin delinquentbehavior.We beginwithtwo assumptions:(1) Friendshiprelations
are complexsocial bondsthatwilllikelyalwaysbe describedincompletely
withreference
to a singledimensionor constructsuch as "attachment"
or
"importanceof friends,"and (2) attempting
to derivea setofcomprehensive and meaningfulfriendshipcomponentsexclusivelyfromthe delinquencyliteraturewould probablybe misguidedintellectualloyalty.This
is so because mostof our knowledgeof how previousdelinquencytheoristsviewed theserelationshipsmustbe gleanedfrombits and pieces of
ratherevocative imagery(Konopka's [1966] delinquentgirlslost in the
crowd, the fraternity
atmosphereof Thrasher's[1963] gang boys, the
excessivedependenceof Cohen's [1955] frustrated
lower-classboys) in1174
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
stead of froma comprehensiveexaminationof what these friendships
were reallylike.
The developmentaland social-psychological
literatureoffermorefully
is actuallyexperiencedby adolesdevelopedanalysesof how friendship
in the friendship
centsand of similaritiesand differences
stylesof males
and females.Our own interviewswitha wide varietyofadolescentswere
also helpfulin pointingus away from"our childhoodfantasies"or, worse
are
yet,our own middle-agedperspectiveson what youthfulfriendships
or shouldbe like. In additionto theseinterviewsand thedevelopmental
literature,exchangetheoryprovideda usefulorientingframework.As
Burgessand Hustonnote,"an explicitlook at exchangeprocessessetsthe
stage forconsideringthe relationshipitself-ratherthanthe individuals
or thelargersystemas a unitofanalysis"(1979, p. 9). A basic characteristic of adolescentfriendships
is thattheyare relativelyvoluntary.4
Thus,
we need firstto considerwhattheparticipantsgetout oftherelationship,
what theyenjoy about it-in short,what are its rewards?
Exchangetheoryalso underscorestheimportanceofconsidering
reciprocal effects
in social interaction.Instead of viewingpeer influenceas a
one-directional
process(i.e., the monolithicgrouppressuringthe adolescent to dress in a certainway, use drugs,have intercourse,etc.), this
theoryconsidersthe actoras bothrecipientand producerof influencein
the group(Burgessand Huston 1979).
Finally,exchangetheoryhas also noteda tendencytowardimbalance
in relationships(Chadwick-Jones1976). Intimacyshouldnot be equated
with the absence of problemsor conflictin a relationship,nor should
forfriendship.5
As Rubinsugperfectreciprocity
be a necessarycriterion
Begests,"Children'sfriendships
rarelycontainonlypositivesentiments.
cause they involve such extensivecontactand interdependence,
close
friendshipsinvariably give rise to negative feelings as well.
. .
. Among
older children,althoughfriendsmay be expectedto be constantlyloyal
and supportive,to respecteach others'rightsand needs,and to agreeon
just about everything,
such expectationscannotbe fulfulledin reality"
(1980, pp. 73-74). It is importantforus to includesomeofthesepresumably negativerelationalqualities because of the assumptionthat they
4 Kimmel (1974) suggeststhat reward-costconsiderationsmay be most salient in adolescence, where friendshipalliances oftenshift.In middle and old age, commitment
and investmentin relationshipsbecome more criticaland may outweigh reward-cost
calculations. But it is incorrectto regardpeer affiliationsas entirelyvoluntary.Many
adolescentswhom we interviewedexpresseda desire to change or at least expand their
friendships,but for a variety of reasons (especially those concerningthe feelingsof
existingfriends)theyfeltunable to do so.
A
AlthoughCicero definedfriendshipas "a completeaccord on all subjects human and
divine," few relationshipswe know would qualify.
1175
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
would be moretypicalof delinquentthanof moreconforming
friendship
groups.
Next we describethe specificelementsof friendshipthatwe derived
fromour generalinterestin capturingtherewardstructure
ofthegroups,
theircharacteristic
patternsof interactionand influence,and the vicissitudesor the moreproblematicaspectsof theserelationships.
Rewards
McCall and Simmons(1978) outlinethreekinds of rewardspresentin
intimaterelationships:
intrinsicrewards(thoserewardingin themselves),
extrinsicrewards(thosehavingsomevalue independentofthegiver),and
a thirdcategory,the supportof role identities.
Intrinsicrewards.-One of themostbasic intrinsicrewardsof friendit providesforconversationand the sharingof
ship is the opportunity
or what has been referred
to as self-disclosure.
Chaikan and
confidences,
Derlega (1976) contendthat as interpersonal
exchangeprogressesfrom
superficialto intimatelevels,the information
exchangedis a "barometer
of the state of the relationship"(p. 184). Accordingto Jourard,"the
amountof personalinformation
thatone is willingto discloseto another
appears to be an index of the closenessof the 'relationship,'and of the
affection,
love and trustthatprevailsbetweentwo people"(1971, p. 33).
In orderto fitthe image of having cold and brittlerelationshipsdelinquentsshouldbe less likelythanotheradolescentsto shareprivacieswith
one another.Are they?
In addition,researchsuggeststhatfemalesdisclosemoreinformation
and moreintimateinformation
than males (Pedersonand Higbee 1969;
Jourard1971; Marks and Giordano 1978; but see Davidson, Balswick,
and Halverson 1980). But do delinquentgirls(also consistentwithexistingliterature)
evidencea departurefromthegenerallyhighself-disclosure
ratescharacteristic
of otheradolescentfemales?
Basic feelingsof caringand trustoffera furtherindex of intimacy.
Chadwick-Jones
(1976) suggeststhattrustis an essentialelementin difsocial fromeconomicexchange. FollowingBlau's lead, he
ferentiating
also notes how "trusttends to build up graduallythroughcumulative
commitmentto a relationship."Bell (1981) found supportfor this in
interviewsin whichrespondents
indicatedthattrustwas themostimporIt is thusimportantto gauge theextent
tantelementin theirfriendships.
to whichdelinquentsmaydifferfromotheradolescentson whatmightbe
of intimatefriendships.
called a basic underpinning
Extrinsicrewards.-McCall and Simmonsdefineextrinsicrewardsas
to theindividualsimplybecause
"eventsand objects[that]are gratifying
theyare usefulto himin pursuinghis variousendeavors.Money,labor,
1176
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
information,
materialgoods, privileges,favors,social status, all these
elementsand more may be helpfulto him in carryingout his various
enterprises"(1978, p. 147). Our interviewswith adolescentssuggesta
numberof ways in which friends"use" one another,even while they
protestthatthisis not a primarybasis of theirrelationships.We include
questionsabouttheextrinsicbenefits
youthsmayderivefromtheirfriendshipsin orderto emphasizethat,in additionto theirqualitativecharacteristics,these friendshipsalso have an agenda or content.Sentiments
like caringand trustaside, a friendis also someoneto sit with in the
cafeteria,to copya mathproblem"offof,"or to getthecar whenyoucannot. These typesof benefitsderivedfroma relationshipafforda somewhat less loftyview of adolescents'connectedness,or interrelatedness.
in notingthat
Cohen alludes to theintrinsic/extrinsic
rewardsdistinction
"the workingclass child [and by inferencethe delinquentboy] is more
dependentemotionallyand forthe satisfactionof manypracticalneeds
upon hisrelationship
to hispeergroups"(1955,p. 101). But do delinquent
kindsofextrinsicor practicalbenefits
youthderivemoreand/ordifferent
fromtheirfriendships
than theirless delinquentcounterparts?
Identitysupport.-McCall and Simmonssuggestthat a crucialfunctionin any intimaterelationshipis thatof providingidentitysupport.As
symbolicbeings,we fancyourselvesin a varietyof roles,but it is importantthatothersprovideus withenoughsupportto keep up thesevisions
of the roles that we rank highly.This additionto the usual intrinsicextrinsicrewardsdistinctionis attractive,given the importantidentity
workthatoccursduringadolescence:"The adolescentis about to crystallize an identity
and forthisneedsothersofhisgeneration
to act as models,
mirrors,helpers,testers,foils"(Douvan and Adelson 1966, p. 179; see
also Foot, Chapman, and Smith 1980; Seltzer 1982; or even Thrasher,
who noteshow "the boy sees himselfthroughthe gang's eyes"[1963, p.
207]). Thus a group will be rewardingto the extentthat it providesa
comfortablearena in which to exploreidentityconcerns.One woman
interviewedby Bell nicely summarizesthe self-confirmation
role of
friendships:
"My friendsgive me a sense of who I am. They do thisby
lettingme expressmyselfand sortof reflectoffof them.What I mean is
thatI can be whoeverI want to be and theylet me be me" (1981, p. 15).
Are delinquentyouthsless likelyto believe theyreceivethis kind of
identity
supportfromfriendsthanadolescentsinvolvedin moreconforming peer relationships?
Patternsof Interactionand Influence
It is important
to examinemorecloselytheparticularnatureand extentof
theinteractions
withinfriendship
networksbecause thesecan be takenas
1177
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
additionalindexesof interdependence
and ofthedegreeto whichwe can
legitimately
assign the primarygrouplevel to them.They also have an
advantage(in termsofmeasurement)
in beinggroundedmorein behavior
than in sentiment.A basic featureof interactionis the sheeramountof
timefriendsspend in each others'company.Althoughtimeis in itselfa
ratherqualitativelyneutralindexof friendship,
it is unlikelythatyouths
who have extremelylow rates of interactionhave the same kinds of
opportunities
to experiencethe rewardsof friendship,
bothtangibleand
intangible,as thosewho interactmorefrequently.
Do delinquentsspend
less timewithfriendsthan do otheradolescents?6
A second characteristic
thatmay differentiate
friendship
stylesis how
longtherelationships
last-their stability.This appearsto relatedirectly
to the "cold and brittlerelationships"argument(presumably,what is
brittleis morelikelyto break). Shortand Strodtbeckand even Thrasher
have commentedon the fluidityand instabilityof many of the gang
be characaffiliations
theyobserved.But shoulddelinquents'friendships
terizedas any moreephemeralthan thoseof less delinquentyouth?We
wouldcautionhereagainstnecessarilyregardinggreaterstabilityas being
inherentlysomehow more positive, especially during adolescence.
Bigelow and LaGaipa (1980) note, forexample,that children'sfriendshipsare nottypicallycharacterizedby stabilityuntilabouttheage of 16,
and theycriticizetheuse of such wordsas "breakdown"and "instability"
in describingchildren'sfriendshipsas implyingan unnecessaryvalue
judgment.I
A thirdqualitative characteristicof friends'interactionpatternsis
definedas thedegreeto whichindividualsare affectedby theirfriendsin
makingchoices about theirown behavior-what is usuallycalled peer
pressureor influence.We have alreadydiscussedthe presumedlink between intimacyand the possibilitiesfor influence,what Thrasherhas
called "controlthroughrapport"(1963, p. 209). Are delinquentsmore
thereciprocal
likelyto feelunderpressurefromfriends,and, considering
6
Cohen(1955)suggeststhatlower-class
youths(and,byinference,
delinquents)
spend
moretimeineachother'scompanythanmiddle-class
youthswhoareofteninvolvedin
a moreformaland variedagenda (such as clubs or athletics).On thisdimension,
Hirschiappearsto be in agreement.
In thecourseofsuggesting
thatifcohesiveness
is
definedas mutualattraction
or respect,delinquentgangswouldbe characterized
by
lowcohesiveness,
he notesthat"ifcohesiveness
is defined
bythefrequency
orduration
ofinteraction
amonggroupmembers,
the'cohesiveness'
ofthedelinquent
gangis no
doubtoftenimpressive"
(1969,p. 160).
7 Sex differences
in stability
havealsobeennoted.Bell believesthatboys'relationships
tendto be morestableand lasting,whichhe hypothesizes
maystemfroma lower
degreeofintensity
and henceless chanceforconflict.
Viewedthisway,then,stability
is producedby a relativelack ofintimacy.
1178
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
natureof group influence,are theymore,or less, likelyto believe that
theyexertsocial power in thesefriendship
networksthemselves?
The Vicissitudesof Friendship8
Roll and Millen(1979) have notedthetendencywithintheexistingliterature on friendshipto see the world of friendshipthroughrose-colored
glasses and "were struckby the absence of any attentionpaid to the
everydaytravailsof friendship"(1979, p. 259). Otherwritershave alluded to the fragility
and vulnerability
of children'sfriendships
(Hartup
1975) or characterizedthemas "tempestuousand changeable"(Douvan
and Adelson 1966), but empiricalworkhas been restricted
to theirmore
positiveaspects. We need to includeattentionto the strainsand imbalances fosteredbyintimacyas well as to itsmoreobvioussatisfactions.
We
have includedthreesuchproblemareas. The firstis conflict,
or theextent
to whichfriendshave argumentsor disagreements
witheach other.This
dimensionalso figuresin the cold and brittlerelationshipsargument.
Empey (1967), forexample,pointsto the practiceof sounding(trading
insults)in delinquentgroupsand suggeststhat"primarygroups,ideally
are supposedto providewarmthand support.Withthe constantsounding thatgoes on it is questionablewhetherlower-classgangsare conducive to close friendships."
It is not clear thatconflictis necessarilyantitheticalto intimacy.Indeed, in our own earlierstudyof friendships
of
collegestudents,we foundthatbothdegreeof contactand comfortwith
friendswere associated with higherlevels of disagreement(Marks and
Giordano 1978; see also Rubin 1980, p. 74). Moreover,it has not been
establishedempiricallythat the levels of conflictand discordin delinquent groupsare any higherthan those foundwithinotheradolescent
friendship
networks.
A secondproblematicaspectoffriendship
is imbalance,whichencompasses a lack ofequalityor reciprocity
in thefriendship,
as wellas feelings
ofjealousy or competition.As withconflict,
thereis no a priorireasonto
believethatimbalancedrelationships
cannotbe rewardingor thatparticipantsin such relationships
will notinfluenceeach other.Reisman(1979),
forexample,makesthedistinction
betweenfriendships
ofreciprocity
and
friendships
of receptivity.
He contendsthatthelatterare no less important(especially,ofcourse,to themoredependentparticipant)as a friendshipformthanare moreperfectly
reciprocalrelations.Thus, thenotionof
imbalancesimplyoffersanothermechanismthroughwhichwe can better
definethe contoursof thesefriendships.
8 We borrowed this termfromRoll and Millen (1979).
1179
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
A thirdproblematicaspectof friendships
we termloyaltyin theface of
trouble.Thrasher(1963) notesthat"loyaltyis a universalrequirement
in
the gang,and squealingis probablytheworstinfraction
ofthecode....
Most boys preferto take a beatingratherthan stoolon theirassociates"
(p. 202). Douvan and Adelson(1966) also foundthatloyaltywas a particularly importantcomponentof friendship;boys in particularoften
defineda friendas "one who will supportyou when troublecomes"
(p. 196). (In theirview thisusuallymeanttroublewithadult authority,
parents, teachers, police, etc.) Are there significantdifferencesbetween delinquentsand otheradolescentsin the salience of this kind of
loyalty?
THE NEED FOR COMPARATIVEDATA
in the
Our researchstrategy
hereinvolvesan assessmentofthedifferences
friendshipexperiencesof youthswho representa broad rangeof delinit is parquencyinvolvement.However,withinthisgeneralframework,
ticularlyimportantto examinethe extentand natureof sex differences.
has been neglectedat boththe theoretHistorically,the femaleoffender
ical and empiricallevels, but, nevertheless,highlystereotypical
images
in theliterature.These data will allow us to confront
flourish
directlythe
image of the femaledelinquentas a lonelyand asocial misfit.
We also see a real need to comparethefriendship
experiencesofblack
withthoseof whiteadolescents,in generaland in relationto delinquency
involvement.The delinquencyliterature
has notoftentackledrace differences (as evidenced by the sheer volume of studies that exclude all
minoritiesfromthe analysis). Studies that do include attentionto race
in thedifferences
have primarilybeen outcomeoriented(moreinterested
or similarities
in ratesofdelinquencyinvolvement
byracialcategorythan
in the processesleading to thoseoutcomes).The friendship
literatureis
also surprisingly
weak in thisarea. Most oftheresearchstudieson adolescent friendships
have used samples of whitesor, if theyhave included
race, have emphasizedthe natureand dynamicsof cross-versussamein the school setting(Sinrace interactionsand friendships,
particularly
gletonand Asher1979). Whilethisinterestno doubtstemsfromconcerns
about schooldesegregation
and interracialacceptance,thefactthatmost
friendships
tendto be intraracialsuggeststheneed formoreworkon the
everydaysocial networksof black (as well as white)youthsand on the
ways in which thesefriendshippatternslink up to otherbehaviors(includingbut not limitedto thatof delinquencyinvolvement).(See Kleinman and Lukoff[1978],who also see a need forthiskind of analysis.)
1180
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
The Sample
The data forthisstudywerederivedfrompersonalinterviewsconducted
in 1982 witha sample(N = 942) of all youth12-19 yearsofage livingin
privatehouseholdsin a largenorthcentralSMSA. A multistagemodified
probabilitysamplingprocedurewas employed,in whicharea segments
were selectedwithknownprobability.Census data wereused to stratify
byracial compositionand averagehousingvalue. Withinsegments,eligible householdsand respondentswereselectedto fillspecifiedsex and race
quotas. The respondentswereequallydividedamongmales and females,
blacks and whites,and lower/middle
socioeconomicstatusrespondents.
This samplingstrategyreflectsour interestin simultaneously
examining
in friendship
and genderas well as
differences
patternsbased on ethnicity
in delinquencyinvolvement.9
The Measures
Our data come fromrespondents'own perceptionsof the nature and
characteristics
oftheirrelationships
insteadoffromdirectobservationsof
friendship
networks(themethodologicalstrategy
thatwas moretypically
employedin much of the early gang research).Eithermethodologyinvolves "lettinggo" of potentiallyimportantdata. Directobservationcan
mosteffectively
capturethedynamicelementofinteractions
and/orrelationships,and it can provide a richercontextualbase fromwhich to
understandgrouplife.Observationalstrategiesalso allow theresearcher
to checkon whetherthe verbal descriptionsyouthofferin an interview
settingcorrespondto therealityoftheirongoingfriendship
relations.This
discrepancyis perhaps most beautifullyillustratedin Liebow's (1967)
discussionof friendshipson Tally's corner.Althoughthe neighborhood
men accorded great importanceto theirfriendships,
Liebow observed
and documentedhow certainharsh realities,economicand otherwise,
oftenintrudedon theirmoreidealisticperceptionsof intimacy.
But the interviewapproach also has some importantadvantages. In
this research,we are most interestedin how youthsexperiencethe rewardsand vicissitudesoffriendship.
Some aspectsofthismustbe viewed
as subjectivelyexperiencedor determined(and to removeall the subjective elementsfromanalysiswould representanotherkindof distortion).
Thus, theadolescentsthemselves(and notoutsideobservers)are particularlywellplaced to determinewhethertheyhave thetrustoftheirfriends,
9 For furtherdetails regardingsampling procedures,see Cernkovich, Giordano, and
Pugh (1983).
1181
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
can be themselvesin theirpresence,or sometimesfeelleftout or marginal.
Some of our otherdimensionsdo have a moreobjectivebase. Here we
have triedto avoid theglobaland sociallydesirablein favorofthespecific
and, whereverpossible, the behavioral(e.g., a questionsuch as "How
oftendo you speak withyourfriendsaboutthefollowingsubjects. .. ?" is
preferableto "Do you feelthatyou and yourfriendscan trulycommunicate with one another?").Anothergeneralstrategyhas been to include
specificattentionto the more problematicaspects of relationshipsconflicts,
jealousy, and thelike. The assumptionthatsuch problemsare
commonto all relationships
shouldalso have theeffectofmovingrespondentsoffthesociallydesirable"myfriendships
are perfect"responseset.
Anotheradvantage of thisprocedureis thatit facilitatesthe kindsof
comparisonsthatare muchneededin theorydevelopmentin criminology.
Liebow, forexample,observeda small groupof lower-classblack men
livingin a particularsectionofBoston.We do notknowverymuchabout
thefriendships
of the womenof Tally's corner,exceptin relationto each
of the men. Neitherdo we know whetherthe kind of intimacythat
Liebow observed,the levels of conflict,or even the discrepancies(betweentheirideal and real friendship
patterns)would notalso be foundin
moremiddle-classsettings.Our sample size and characteristics
allow us
to specifythenatureofsimilarities
and differences
bylevelofdelinquency
ofsex and/orraceinvolvement
withoutbracketingoffthepossibleeffects
ethnicity.
The specificfriendshipitemsincludedin our finalinterviewschedule
interviewsover a 14-yearperiodwith
weredevelopedfromunstructured
a varietyof adolescentmales and females.Interviewswere conducted
withyouthswho rangedin age fromsixthgradersthroughhighschool
locationin the
students,who variedin termsofdelinquencyinvolvement,
schoolprestigehierarchy,
race, socioeconomicstatus,and so forth.Interviewswerealso conductedwithinstitutionalized
femalesand withyouths
on probation.In additionto theseindividualinterviews,
discussionswere
heldwithsmallgroupsofadolescentfriendnetworks.These smallgroups
the periodof developingthe more
servedalso as consultantsthroughout
structured
interviewschedule.
Anothersourceofitemswas a pool ofessaysby highschooland college
studentswho wroteabout thingsthattheyliked and dislikedabout their
friends.An earlierstudyby Marks and Giordano(1978) and thedevelopmentaland delinquencyliteraturewere also sourcesforitem development. A factoranalysis of all the pooled itemsresultedin 13 distinct
dimensionsof friendship
(see App. A fora completelistoftheitemsthat
composeeach scale). In additionto thesequalitativedimensions,youths
of theirfriends.At the
were asked about the backgroundcharacteristics
1182
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
witha
beginningofeach interview,therespondentprovidedinterviewers
listof names(firstnames,nicknames,or initials,to assureanonymity)
of
the groupof friendstheyusuallyhang aroundwith.'0 Respondentsindicatedwhethereach personlistedwas male or femaleand whethermostof
theirfriendswere eitherolder, about the same age, or youngerthan
themselves.These backgroundvariables were includedprimarilyas a
in the composicheckto determinewhethertherewere basic differences
tionof theirfriendship
thequalitative
groupsthatmighthave influenced
characteristics
we have outlined.
Delinquencyinvolvement
instruwas measuredbya 27-itemself-report
mentthat is a revised versionof that used by Elliott et al. (1982). It
containsa broad range of items that include minoras well as major
offenses.However, we avoided the use of summatedscale scoresas the
measure of delinquencyinvolvementbecause of the wide range of
seriousnesscapturedby the items(withsuch a methodyouthswho score
in thehigh-frequency
rangeon statusor othernonseriousoffenseswould
have been equated with youthswho have committedseriousfelonies).
Instead, we developed a categoricaloffendertypologythat takes into
accountthedimensionsof bothfrequencyand seriousness.Five offender
categoriesweredefinedas follows:(1) Nonoffenders
are youthswho have
committed
no morethanone or twominoroffenses
and no majoroffenses
duringthe past year. Minor offenseswere definedas behaviorsthat
would usually be treatedas a misdemeanor(e.g., runningaway, petty
theft,disorderlyconduct),whereasmajor offenseswere definedas those
behaviorsthatwould be treatedordinarilyas felonies(e.g., grandtheft,
aggravatedassault, breakingand entering).(2) Low-frequency
minoroffendersare youthwho scored relativelylow (underthe median) on the
minoroffensescale and indicatedtheyhad committedno major offenses
duringthe precedingyear. (3) High-frequency
minoroffenders
have a
scoreabove themedianon theminoroffensescale but admitno involvementin major offenses.(4) Low-frequency
major offenders
had a score
below the median on the major offensescale. (5) High-frequency
major
offenders
weredefinedas thosewho had a scoreabove themedianon the
major offensescale.
10 We adoptedthisstrategy
insteadoffocusing
on othersocialunits(suchas thebest
friends'
dyador thegang),becausewe are interested
in thecontextin whichmostof
therespondent's
socialactivityactuallytakesplace. Althoughrestricting
attention
to
thedyadhas somemethodological
advantages(see Kandel1980),itminimizes
therole
ofthelargergroupand one'splacein it.Also,thegangmaybe highly
visible,buteven
gangmembers
spendmuchoftheirtimein smallernetworks
(see Shortand Strodtbeck
1965).Differences
betweenourfindings
andthoseofearlierresearchers
couldreflect
at
leastin parttheuse ofthisunitofanalysis,whosesizeis essentially
determined
bythe
respondent.
1183
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
For our purposes here, it is appropriateto consider 1 as the most
conforming
group, with a gradual increasein level of delinquencyinvolvementto 5, whichis the highestlevel of involvement.For example,
althoughcategory5 is definedsolely on the basis of participationin
seriousoffenses,we know thatthemajorityofyouthsin thisgrouphave
also participatedin a wide rangeof nonseriousor statusoffensesas well.
(App. B presentsthedistribution
ofoffender
categoriesby race and sex.)
FINDINGS
Tables 1-3 presentthree-wayanalysesofvarianceforeach ofthecomponentsof friendship
by sex, race, and level ofdelinquencyinvolvement."
Because of the potentialfor age differences
even withinthe period of
adolescence(see, e.g., Honess 1979; Bigelow and LaGaipa 1980; Mannarino1980; Sharabany,Gershoni,and Hofman 1981),age is controlled
as a covariate.12
The Rewards of Friendship
Table 1 presentsmean scoresas well as F-values forthemajor comparison groupsas theyrelateto the intrinsic,extrinsic,and identitysupport
rewardsof friendship.
The two intrinsicrewardsscales-self-disclosureand, more particularly,caringand trust-are mostrelevantto thecold and brittlerelationships argument.Contraryto what controltheorywould predict,there
were no significantdifferencesacross increasinglevels of delinquency
in theextentto whichrespondents
believetheirrelationships
involvement
containtheseelementsof caringand trust.
in self-disclosure
rates,it
Althoughtherewas a significant
difference
reflectsa slightlylower self-disclosure
rate among the least delinquent
on thesescales are moreproSex differences
category,the nonoffenders.
in intimateways
nounced.Females are muchmorelikelyto self-disclose
and to characterizetheirfriendships
as consistingofcaringand trustthan
between
are males.13 It is importantto underscorethelack ofinteraction
" Separate analyses of variance revealed no significantdifferencesacross levels of
delinquencyinvolvementin youths' descriptionsof theirgroups' basic structuralfeatures,such as size, age, and sex composition.Thus, any qualitative differencesfound
should representmore than a simple artifactof the structuralmakeup of these groups.
12 For a discussion of age differences
in friendshipprocessesusing the presentdata, see
Kinney (1984).
13 There is also a race by sex interactionforthe self-disclosure
scale, produced primarily by the particularlyhigh scores of white females.
1184
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
0
00
0
ed
t
*
*
06~~~~~~~~~
<
4 2
(-.
0v
D0
>)
4)2
00
00 0
Q
UD
' 't00
~~ ~ ~
0
Z.
)
u
C)
O
4)
00 0
\00
*
0
&
>
-,
.
.-
S
(4~
N
.
.
-.
~~>
00 I
Z
0
("4~~~0'
4)
0
t
'C00
et
(-e
-'
.
-
.
0'-1
0
~ ~ ~~~~~4O~
*
42
4)
00
c4~~~~~~~~~~~~
0
4
4
.
Cd
'0'-
COr
00
0
0~~~
~
~~~
~~
4-~~
CT)
42
*
44
0
4-24'-)
(4
*
)
a~
*
*
a.0
0
~
4-0C'4
4.)
(
0
0m3
Z
;KX
=
0~4)
4")
(-.0
Or-
A
4.
4
4
4-4000 4-400
0
00'0
d
C
i
4.)
~ ~ *~~~~~~~1
~ ~ ~ ~ ~ ~
~
~
44
(-
4)0
~~
2
gY<Z
|
4-)
~ ~
Z
z
OO
44,
Ct'40
C
-7
00
|~~~~~~~%
"44
0
0
44
C/)*
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
o.
(44
9*
0
C00
t r <
?
~~~~~~~'C
? [A
0
-X
VZ)
Z
0)
>
A
cO
Z
000
Lo0
t
*
* 0
40
m0 0
-~~~~C
o
z
o D0
0
0.
z
0-0
>
0-0
0-00 0-1
00
00
0-
-
-
'
'
'
'
0C
Z
0.0
~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~0
0.0
0
0
000
00
af0-
0-000~~~~~~~
a'~~~~~~0
0
~00'
H
~~~a,4~~~~~~5 *
0.0tc>S
m3>_N
>1800
0*
it
00~~~~~~~~~
U
k~~~~~0
004)4-
00
0
-
0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~c
(cc
C/0
o
-
~
~0.)
c~o
000
000-0
0c-c11860-
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
0
S
Z
4=
%t
4
04~~~~~~~~~~
~~~4
0
0.
i
s*
~
**
4
?
?4
~~~~~~~~~~~~~4
~ ~ ~ ~ ~ ~ ~ ~ ~ ~
0~~~~~~~~~~~~~~~~~~~.
0
-o~~
X
z
4
~~~~~~~~~~~
4)
e.
1 187
z;=-Nto
-
t
*0
0
X m44t
4)
0.
z0
~~~4
0
CA~~~~~~~~~~~~~-
Cd
~~~~4~-4
a
O0*41-
z4
_f~
<
4?
C
4
)
J
b
0
U
0.
.
.0
z
4
44
o)
~00
4,)-..
-l-
--'___
m
c,0
o
.d0
0
cf)
z
0
414
z
0~~~~~~~~~~~~
~~~
Q
~
~
~~40444
4
a..~~~~z
18
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
b
AmericanJournalof Sociology
sex and delinquencyin these respects:There were no significant
differences amongfemalesin ratesof caringand trustacrossthefivelevelsof
delinquency,and femalesat almost everylevel of delinquency(except
black femalesin the low-frequency
minoroffendercategory)are more
intimatethanare malesat a similarlevelofdelinquencyinvolvement
(See
fig. 1).
In additionto thesesex differences,
a significant
main effectforrace
was obtained: Whites in our sample scored significantly
higheron the
caringand trustscale than did blacks.
The analysisof covarianceof extrinsicrewardsrevealed,as expected,
thatyouthin themorehighlydelinquentcategoriesare muchmorelikely
to say thattheyreap certaintangiblesfromtheirrelationship
thanare less
delinquentyouth.(This is one dimensionthatobviouslydid not escape
our generaldesireto be independentof delinquencyinvolvement.)What
is perhapsmorerevealingis thattherewas no significant
difference
across
delinquencylevels in the extentto whichyouthsdiscussedor relatedto
each otheron school matters(althoughthe mean scoresare somewhat
lowerforthemajor offender
categories,thisdifference
is notsignificant).
This suggeststhat,althoughsuchyouthsmaynotbe as schoolorientedor
successfulas nondelinquents,
youthswho do engagein delinquencyhave
notabandonedthe concernsof schoolaltogether.Interestingly,
moredelinquentyouthalso scoredhigheron the status-striving
scale, whichindexed the degree to which theyreceivedcertainratheradult-oriented
benefitsfromtheirfriendships
(e.g., discussionsof job plans forthe future,help in meetingmembersof the oppositesex, etc.).14
the onlymain effectforrace
Withregardto sex and race differences,
occurredin the area of tangiblerewards,wherewhiteswere morelikely
in the analysisof
to agreethattheirfriendshelpedthem.Sex differences
extrinsicrewards consistedof a much higherscore for femalesin the
extentto which theydiscussed and reliedon friendsforschool-related
mattersand a significantly
higherscore by males on the status-striving
scale. The lack of any significant
difference
by genderon the tangibles
scale shouldnot be overlooked:femalesindicatedthattheyare about as
likelyas males to get drugs or alcohol fromfriends,to hang out at a
friend'shouse when parentsare gone, and so on. Althoughfemalesas a
groupare not as delinquentas males, theseresponsesdo pointto a relativelygreaterdegreeofparticipation
by girlsand theirfriendsin whatare
cultureactivities"thanis generally
usuallyconsidered"hedonistic/youth
assumed.
In the extentto whichrespondentsfelta lack of identitysupport(felt
14 There is a three-wayinteractionon the status-striving
scale, the resultprimarilyof
the higherscores of black male delinquents.
1188
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
4.3/
01.
v
4.1
o
3.9
,,
O
--s-White
Females
Females
*
~~~~~~~~~~~~Black
? -White Males
Black Males
-0~~~~~~~
,
3.3
72
Non offender
3
4
Low Freq.
Low Freq.
High Freq.
minoroffender minoroffender majoroffender
5
High Freq.
majoroffender
OffenderGroup
FIG. 1.--Levels ofcaringand trustbydelinquencyinvolvement,
race, and sex
that theycouldn't be themselveswithintheirfriendshipgroups)there
wereno significant
acrosslevels of delinquencyinvolvement.
differences
That is, moredelinquentyouthsbelieve thattheyenjoy the rewardsof
self-confirmation
withinfriendshipsabout as much as less delinquent
adolescentsdo. Similarly,therewere no main effectsforrace; however,
as a subgroup,males were significantly
morelikelyto believe thatthey
W
couldn'tbe themselveswhile withtheirgroup.Patternsof Interactionand Influence
across delinAs table 2 suggests,the re no significant
differences
quencylevels in the frequencyof reportedinteractionwithfriends.Fe15
is producedby thesomewhathigher
interaction
delinquency-by-race
A significant
score.
ofblackswitha higherdelinquency
agreement
1189
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
males do reporthigherlevels of contact;however,this difference
is in
largepartattributable
to theirmuchhigherratesof"talkingon thephone
withfriends."''6
Althoughtherelationships
ofdelinquentshave oftenbeen
describedas tenuous and unstable, we findno significantdifferences
acrossdelinquencylevelsin thelengthoftimerespondents
reportedbeing
friends.On the otherhand, main effectswere foundforboth race and
sex. ConsistentwithBell's (1981) research,males reportedthattheyhad
been friendsforsomewhatlongeraveragetimes.In addition,thefriendships of blacks were foundto be significantly
morestablethan thoseof
whiteyouths.
In general,the findingsreportedabove suggesta lack of significant
differences
across manyqualitativeindexesin theways in whichyouths
in delinquencyinvolvementdescribethelevels of
who varysignificantly
intimacyand interaction
withintheirfriendship
groups.However,delinto peer influence
quent youthdid reporthigherlevels of susceptibility
thandid theirless delinquentcounterparts
(See table 2). A higherlevel of
susceptibility
was also reportedby whitesand males in our sample.
In additionto reporting
a significantly
greaterlevel of susceptibility
to
main effectforthe recippeer influence,we also obtainedan interesting
morelikely
rocal peer influencemeasure.Delinquentswere significantly
to believe that they"oftenpressuredtheirfriendsto behave in certain
more
ways"thanweretheless delinquentyouths.Males werealso slightly
likelyto believethattheyexertedsocial power,as wereblacks,but these
differences
were not significant.
The Vicissitudesof Friendship
Regardingthe extentto which disagreementor conflictswere foundin
main effectfor
these friendships(see table 3), therewas a significant
delinquency:delinquentyouthsreportedhigherlevels of disagreement
than did less delinquentadolescents.Males also reportedsignificantly
in conflict
higherlevels of conflict,whereas therewere no differences
levels reportedby black in contrastto those reportedby whiteadolescents.
In addition,therewas a significant
difference
by delinquencyin the
scale indexingfeelingsof imbalance(as well as jealousy and competition
in the group),but the means do not forma simplelinearpattern.There
were no main effectsforrace, but males were morelikelyas a groupto
experiencethesefeelings.
16 Thereis also a significant
sex-by-delinquency
interaction,
whichis producedbythe
relatively
low contactscoresofnonoffender
males.
1190
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
believedthat
Finally,the more delinquentyouths,not unexpectedly,
theywould be moreloyal to theirfriendsin the face of trouble.(This is
anotherscale in whichthe findingsshould not be altogethersurprising,
giventhelikelihoodthatdelinquentsmayhave had morepreviousexperiencewithtrouble,lyingto police,etc.). On theotherhand,blacks,as well
as females,scoredsomewhatloweron theseitems,indicatingtheywould
be less likelyeitherto stickby friendsheadingfortroubleor lie to protect
them.17
DISCUSSION AND CONCLUSION
In this article,we have developed several dimensionsalong which to
make basic comparisonsof the qualitiesand behaviorsthatcharacterize
adolescents'friendships.The factoranalysisof all the pooled friendship
itemscorroboratedour suspicionthatit is impossibleto capturethe essence of theserelationshipsusing a singleattached-unattached,
or positive-negative,dimension. Nevertheless,these friendshipcomponents
have allowed us to examinethe validityof many of the images about
thathave emergedfromourmajordelinquency
delinquents'relationships
theories.
The data presenta picturemorecomplexthanthatprovidedbycontrol
ofdelinquentsas "exploitive
theorists,
who have depictedthefriendships
ratherthanwarmand supportive"(Empey 1982,p. 273) or,alternatively,
by earliersubculturaltheoristswho may have idealized the gang as a
noblefraternity
characterizedonlyby camaraderieand we-feeling.
Overin theirlevelsof involveall, we findthatyouthswho are verydifferent
mentin delinquencyare neverthelessquite similarin the ways in which
in the
theyview theirfriendshiprelations.There were no differences
averagelengthoftimerespondentsreportedbeingfriends(stability)or in
the ongoingfrequencyof theirinteractions(contact).Delinquentswere
somewhatmore likelythan theirless delinquentcounterpartsto share
privacieswith one another(self-disclosure)
and were about as likelyto
believe that they can "be themselves"while in the companyof these
friends(self-confirmation).
Contraryto the centralassertionof the cold
and brittlerelationshipsargument,delinquentswere no less likelythan
othersto believethattheyhave thetrustoffriendsand thatthesefriends
"reallycare about themand what happensto them."
At the same time,delinquentsdid reportsignificantly
higherlevels of
disagreements(conflict)with friends.This findingis importantin a
methodological
sense,in thatit demonstrates
thatthedelinquentrespon17 A significant
delinquency-by-raceinteractionis the resultof thehigherloyaltyscores
of white delinquents.
1191
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
dents(as well as othersubgroups,includingmales, who scorehigheron
conflict)are capable of perceivingand describingtheirrelationships
in a
relativelycomplicatedway instead of adaptinga uniformly
positiveor
negativeresponseset. That these adolescentsare willingto talk about
some negativeaspects of theirrelationshipsgives added weightto the
validityoftheirmorepositiveresponses.But beyondthismethodological
note, what meaning should be attached to these conflictscores? One
interpretation
is thattheseresponsesrepresentat leastpartialsupportfor
the controltheoryconceptionof delinquentfriendships.But thereis an
alternativeexplanation.Viewed in the contextsof the levels of contact,
and caring and trustalready reported,the somewhat
self-disclosure,
higherconflictlevels may be taken as additionalindicatorsof intimacy
and the importanceof the friendships
to theseyouths.A more neutral
conclusionis thattheseresultssimplyreflecta different
friendship
style
thatshould not be viewed as eitherpositive(attached)or negative(cold
in styleis reflected
and brittle).Anotherdifference
in thegreateremphasis
on loyaltyin responsesofthemoredelinquentyouths.A stronger
beliefin
suchvalues as lyingto protectfriendsmaybe based moreon theirgreater
familiarity
withsuchsituationsthanon a particularly
deep fraternal
bond
or code (a la Thrasher).
An examinationof what we mightcall the interactivecontentof the
friendships
also revealsa mixedpattern.Delinquentsindicatedthatthey
weremorelikelyto reap certaintangibleas well as social (status-striving)
extrinsicbenefitsfromtheirfriends,but therewere no significant
differences foundforthe extrinsicrewardsscale measuringschool concerns.
The higherscores on two extrinsicrewardsscales do not indicatethat
more important(or extrinsic)than
delinquents'relationsare inherently
are thoseofless delinquentyouth.We obviouslydid notexhaustthelistof
all the extrinsicrewardsyouthmightderive fromfriends.But taken
together,the scores of delinquentyouthon bothintrinsicand extrinsic
factorssuggestthat delinquents,at least as much as otheradolescents,
derivea varietyof significant
benefitsfromtheirfriendship
relations.
On the basis of thesefindings,we would reverseHirschi'spreviously
and influence.
quoted statementconcerningthelinkbetweenattachment
It is reasonableto conclude that when adolescentfriendships
are relativelywarm and intimateand provide some combinationof intrinsic,
extrinsic,and identitysupportfunctionsforthe participants,actorsare
likelyto exertconsiderableinfluenceupon each other.An examinationof
is illustrative.On
the patternof responsesof group 1 (thenonoffenders)
almosteverydimension,membersofthismostconforming
groupare least
attachedto friends.Generally,theyhave thelowestlevelsofinteraction,
the lowestlevels of caringand trust,the lowestrates of self-disclosure;
also, theyare the least likelyto admit to the group'sinfluenceon their
1192
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
own behavior.In contrast,themoredelinquentgroupsare morelikelyto
believethattheymaybe influenced
byfriendsand thattheyexertconsiderable influenceon the group. This findingis important,in thatit sugwithinthesegroupsratherthana
gestsa patternofmutualreinforcement
one-wayinfluenceprocess.The controltheoryconceptionofpeerattachmentmayhave emergedlargelyas a reactionagainsttheearlierimageof
the delinquentgroupas beingall powerfulin its effects.The individual
actorseemedto be a passive agentwho, on enteringthegroup,tookon its
attitudesand values.
The findingspresentedhereare consistentwiththe contentionof controltheorythat individualsmay bringcertaindelinquentvalues to the
group at the outset. This would explain youths'respondingthat they
oftenpressuretheirfriends,as well as thereverse.Althoughwe have not
dealtdirectlywiththeimportant
issueofwhatcauses theinitialattraction
to the delinquentor any group, most of the developmental/friendship
in values (value homophily)
literaturefavorsthe conceptof a similarity
precedingentranceintothefriendship.
But we believewithKandel (1980)
thatthereare importantprocessesthatcontinueto workoncetheindividuals have been drawn together;elementsof mutual reinforcement
and
influencethatstrengthen
delinquentpatternsbeyondwhat would be expectedfromtheirinitialvalues. As Hirschihimselfconcludes,"thereare
groupprocessesimportantin the causationof delinquencywhose automaticoperationcannotbe predictedfromthe characteristics
of persons"
(1969,p. 230). This attemptto modifyhisearlierpositionabouttheroleof
peersis notlinkedto controltheoryor quotedas oftenas are his data that
a lack ofattachmentin delinquentfriendships.
empiricallydemonstrated
The data presentedhere seriouslycontradictthe lack-of-attachment
thesis.Othercontributions
of controltheory(namely,the view thatimportantprocesses such as familyattachmentprecede the group's influence)mightbe integratedwiththesomewhatmorecomplicatedimages
of the delinquentand his friendsthatemergefromthisstudy.
Sex Differences
It is importantto view even thesignificant
differences
acrossdelinquency
categoriesagainstthe oftenmoredramaticvariationsin friendship
style
and influenceattributableto the basic demographicvariableof sex. The
experiencesassociated with genderin our societyappear to framethe
styleand contentof friendships.Females, regardlessof theirlevel of
involvementin delinquency,are likelyto be involvedin moreintimate
relationships.In addition to casting doubt generallyon the cold and
brittlerelationshipsargument,thesedata also questionthe imageof the
femaledelinquentas a lonelyand asocial misfit,unable to establishade1193
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
quate peerrelations.The similaror generallyhigherscoresoffemaleson
dimensionsalso call intoquestiona popularexplanation
thesefriendship
in criminality,
fortheoverallgenderdifferences
namely,thatfemalesare
"Sutherlandmaintains
notas peerorientedas are theirmale counterparts:
thatmostcriminalbehavioris learnedwithinintimatepersonalgroups.
... The familyhas traditionally
been the groupwithwhichfemalesare
mostintimately
connected,even duringadolescence.They are also likely
to be more carefullysupervisedwithinthe family.If crimeis learned
withinintimatepersonalgroups,and formostfemalesthecrucialprimary
groupis a restrictive
family,theyare muchless likelyto learn criminal
behavior"(Leonard 1982, p. 107).
The data here (and in otherdevelopmentalstudies)would lead us to
suggestthatthestatementabove somewhatoversimplifies
theimportance
of peer friendshipsto adolescentgirls. Females spend as much (if not
more,countingphoneconversations)
timein thecompanyoftheirfriends
as do males. They reap manyrewards,bothintrinsicand extrinsic,from
these friendships.Therefore,ratherthan discounttheirpeer involvements,we must discovermore about the differential
dynamicswithin
thesenetworksthatseem to amplifydelinquencyin thecase of boysbut
generallyinhibitit among girls(but not always, because some girlsdo
develop delinquentpatterns).These data offersome clues. It will be
recalledthat males as a groupwere morelikelyto believe that friends
exerteda varietyof pressureson them;theirconflictscores were also
higher.If we conceiveof delinquentacts as a set of behaviorsrequiring
somerisktakingor daring-that is, somesortofpush (Shortand Strodtbeck 1965)-it maybe thatthefriendship
stylesofmalesare mostconducive to thekindof groupprocessesthatmove individualmembersto the
pointof collectiveaction. In contrast,femalesare less likelyto indicate
that theseovertpressuresand conflictscharacterizetheirrelationships.
in the normsgoverningappropriatefriendshipstyle
These differences
may serveto inhibitdelinquencyamongfemaleseven in the presenceof
otherfactorsthatmightotherwisepromoteit (e.g., poorfamilyrelations,
economicmarginality,contactwith delinquentopportunities,
and the
like). Females who do becomeinvolvedin delinquentacts, then,would
have to adopt both a set of attitudesin whichtheysaw delinquencyas
appropriate,possible,or desirablebehavior(Harris1977;Giordano1978)
and a friendshipstylein which theywould encourageeach otheras a
groupto act on theseorientations.
Race Differences
Variationsin thefriendship
patternsof black and whiterespondents
also
suggestthepossibilityof ethnicdifferences
in theroleof groupprocesses
1194
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
in the etiologyof delinquentbehavior.At the riskof oversimplification,
we would describetheresponsesof blacks as beingsimilarin manyways
to those of whitesbut also as reflecting
a less intensefriendshipstyle.
That is, blackswereless likelyto believethattheirfriendspressuredthem
to behave in certainways,theywereless likelyto believethattheywould
lie to protectfriends,and theyscoredsignificantly
loweron thecaringand
trustscale. At thesame time,theirfriendships
had greateraveragestabilitythanthoseofwhites.As notedpreviously,thereis verylittleliterature
examiningthefriendship
patternsofblack adolescents,butwhatwe have
foundappears consistentwith thesefindings.Berg and Medrich(1977)
observedfriendship
blacklow-incomeneighpatternsin a predominantly
borhoodand in an affluent
whiteneighborhood.They foundfriendships
in theformerto be less exclusiveand morespontaneousthanthoseofthe
whiteyouths.Iscoe and Harvey(1964) foundthatblack youthswereless
likelyto respond(in an experimentalsetting)to peer pressureon a task
thatinvolvedcountinga metronomeclick than were whites.They also
foundthatthepeak age at conformity
occurredalmostthreeyearsearlier
forblack than forwhite subjects. Finally,Billy and Udry (1983), in a
studyoffactorsaffecting
sexual behavior,foundthatamongwhitesthere
was a moredirectrelationship
betweenthesexualintercourse
behaviorof
same-sexfriendsand respondentsthanforblacks in theirsample. These
fewfindingssuggestthe need to develop morerefinedtheoriesthattake
intoaccountdifferences
in the salienceof friendships
and in the styleof
beingfriendsforblack in contrastto whiteyouth.Our speculationis that,
even thoughdelinquentacts mightbe equally likelyto occur withina
group context,peer pressurewill play a more directrole in the delinquencyof whitethanin thatof black adolescents.However,moreinterview-observational
data on the specificsituationaland social contingencies that producedelinquentaction among blacks, in comparisonwith
whitesor otherethnicgroups(as well as formalesin contrastto females),
is obviouslyneeded.
Toward a Sociologyof Nerds
Anotherlineofresearchsuggestedbythepresentstudyis a moresystematicfollow-upoftheadolescentswho reallycan be classifiedas lonersand
misfitswithinthe adolescentstratification
system.We have shownhere
thatdelinquentsappear to be fairlysuccessfulin negotiatingand maintaininga set of social relationships(even if, as Sherifand Sherif[1964]
note,it makessome people uncomfortable
to thinkthat"antisocial"people can actuallybe quite social). However, more researchneeds to be
focusedon thoseyouthswho scoreverylow on theseor similarfriendship
dimensions.We need to know a great deal more about the social (as
1195
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
opposedto purelyclinicalor psychological)factorsthatgive riseto being
consideredmarginalor, worse, being ignoredentirely.Followingfrom
this, we need to assess the consequencesof this lack of peer support,
as it may be connectedwithemotionaldistressor illness.We
particularly
would hypothesizethat these outcomesare more likelyresultsof poor
peer adjustmentthan is delinquencyinvolvement.'8
APPENDIX A
Dimensionsof Friendship
The factoranalysisof all the pooled itemsyieldedthe followingdimensions,each witha simplefactorialcomplexity.
1. Intrinsicrewards
How oftendo you talk to yourfriendsabout the
a. Self-disclosure:
followingthings?Questionsor problemsabout sex; how yourparents treatyou; whetheryourparentsunderstandyou; thingsyou
have done about whichyou feelguilty(a revisedversionof West
scale, for use with adolescent
and Zingle's [1969] self-disclosure
samples).
"I feelcomfortable
callingmyfriendswhenI have a
b. Caring/trust:
problem;I can trustthem-I can tell them privatethingsand
know theywon't tell otherpeople; theycare about me and what
happensto me; they'reeasy to talk to."
2. Extrinsicrewards
have a contentand
Because extrinsicrewards,almostby definition,
rewardsscale
are notvalue free,we attemptedin theoriginalextrinsic
to include a wide range-from the more pro-social("They help me
with my school work") to the more antisocial("They get drugsfor
18 Termsused by adolescents
to referto thesemarginalyouth,alongwiththeirown
of who belongsin thesecategoriesoffera glimpseintothe potentially
descriptions
devastating
consequencesforindividualsso labeled.Termsinclude:wastesor waste
products,glugs,grimers,
dirtballs ("didn'tdo drugsor alcoholbut shouldhave"),
scum,trashheads,
grubbers,
nobodies("theystayoutofeveryone's
way");speds("the
specialeducationstudents-theycan be foundhelpingthejanitorin thecafeteria");
queers("nonathletic
maleswithwimpybodieswho are notintopartying");
thebores
("theymakehighgrades,go to schooldances,weardrabclothes,areusuallyuglyand
just kindofthere");nerds("don'tdressin style,scummylooking,no one knowswho
theyare,theydon'tknowwhatalcoholanddrugsare");loners("peoplewhodon'thave
anyfriends
at all,justfaces");losers("theyjustdon'tfitin anywhere.
Butpeopledon't
physically
abuse them.Mostlytheylike to intimidate
themand makefunof them
while kiddinglytryingto be theirfriendto impressotherstudents");"we never
bothered
to givethema name;theywerejustleftcompletely
alone"(thesedescriptions
weretakenfromdiagramsmade by undergraduates
who wereasked to depictthe
stratification
systems
withintheirhighschools).
1196
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
me"). Other items were more neutralin that both delinquentand
nondelinquentyouthmight,e.g., "talk about problemsat school."
The factoranalysisresultedin threedistinctclustersof rewards:
a. Tangibles:"They geta car forus to use; theygetbooze forme; they
get drugsforme; when theirparentsgo out, we hang out at their
house."
b. Help with school: This includeshow oftenrespondenttalks with
friendsabout how well he or she getsalongwithhisorherteachers;
how oftenrespondenttalkswithfriendsabout problemshe or she
has at school,and how oftenfriendshelp withschoolwork.
c. Status striving:This includes how oftenrespondenttalks with
friendsabout job plans forthe future,gets theirhelp in meeting
peopleto date, feelsthatpeoplelook up to himmorebecause ofhis
3. Identitysupport
a. Self-confirmation:
"I can't reallybe myselfifI wantto stayfriends
withthesepeople."
4. PatternsofInteraction
a. Contact:How oftenduringtheweek do you spendtimewithyour
friendsotherthan at school? How oftenduringthe week do you
speak to yourfriendson the telephone?
b. Stability:In general,how manyyearshave you been friendswith
mostof thesepeople?
c. Peerinfluence(group-- actor):Sometimespeopleare influenced
by
theirfriendsas to how theyact and thinkabout things.Please tell
me the numberon thiscard thatindicateshow muchyou agreeor
disagreewitheach of thesestatements:I sometimesdo thingsbecause my close friendsare doingthem;I sometimesdo thingsbecause that'swhat the popular kids in schoolare into;I sometimes
do thingsso my friendswon't thinkI'm chicken;I sometimesdo
thingsbecause myfriendsgive me a hardtimeor hassleme untilI
do them; I sometimesdo thingsso my friendswon't thinkI'm
immature;I don'tlikebeingdifferent
or stickingoutin a crowdso I
sometimesgo along with thingsforthat reason; I sometimesdo
thingsnot because myfriendspressureme butjust because I think
it will impressthem;I sometimesdo thingsbecause I don'twantto
lose the respectof myfriends.
d. Peer influence(actor-- group):"I probablypressuremyfriendsto
do thingsmorethantheypressureme; I sometimestalkmyfriends
intodoingthingstheyreallydon't want to do."
5. The vicissitudesoffriendship
a. Conflict:How oftendo you have disagreements
or argumentswith
yourfriends?How oftendo you purposelynottalk to yourfriends
because you are mad at them?
1197
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
b. Imbalance: "Sometimestheyjust won't listento me or my opinion." "I thinkI like mostofthepeoplein mygroupmorethanthey
like me." "Some people in the groupare always tryingto impress
in the
people outsideour group.""There is too muchcompetition
group.""There is too muchjealousy in the group."
If you foundthatyourgroupof friendswas leadc. Loyalty/trouble:
ing you into trouble,would you stillhang around withthem?If
yourfriendsgotintotroublewiththepolice,would you be willing
to lie to protectthem?(Minor,n.d.)
1198
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
4-0 coo
%0I0)
0N co)
-4C-4)
00
00~~~~~~~~~0
0)0
00)
00
-13
00
O
0
0~~~~~4
~
4)
I
CI)
0)
\
0~~~~~~~~~~~~~0
z
H.!
0)
40
c
0
:
:1
0
4
4)~~~~~~~~~4
0)
0
t>40C)
C
4-')
4)
E-4~~~~~~~~~~~~~~~~~~~4
0
~~~~~~
C'3~
zE-4~~~~~~~~~~~~~~~~~~~4
~ ~ ~ ~ ~ ~ ~ ~~~4
4)
l
0
40
.
44
0)
4)
0
~~)-4
0
0
4-')4-')
~~~~80)
4-z
)~~~~~~~~~c') ~~C4)
z
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
REFERENCES
Bell, Robert R. 1981. Worlds of Friendship. Beverly Hills, Calif.: Sage.
Berg, Mary, and Elliott A. Medrich. 1977. "Children in Five Neighborhoods."Report
by the Children's Time Study, School of Law, Universityof California, Berkeley.
Bigelow, Brian J., and John J. LaGaipa. 1980. "The Development of Friendship
Values and Choices." Pp. 15-44 in Foot, Chapman, and Smith, eds.
Billy, 0. J., and Richard Udry. 1983. "Patternsof AdolescentFriendshipand Effects
on Sexual Behavior." Paper presented at the annual meeting of the American
Sociological Association, Detroit.
Bowker, Lee H., H. S. Gross, and Malcolm W. Klein. 1980. "Female Participationin
Delinquent Gang Activities."Adolescence 15 (59): 509-19.
Burgess, R. L., and T. L. Huston, eds. 1979. Social Exchange in Developing Relationships. New York: Academic Press.
Campbell, Anne C. 1980. "Friendship as a Factor in Male and Female Delinquency."
Pp. 365-89 in Foot, Chapman, and Smith, eds.
Cernkovich, Stephen A., Peggy C. Giordano, and Meredith D. Pugh. 1983. "Serious
Chronic Offenders:The Missing Cases in Self-ReportMeasures of Delinquency."
Paper presented at the annual meeting of the American Society of Criminology,
Denver.
Chadwick-Jones,J. K. 1976. Social Exchange Theory:Its Structureand Influencein
Social Psychology.New York: Academic Press.
Chaikan, Alan L., and Valerian Derlega. 1976. "Self-Disclosure." In Contemporary
Topics in Social Psychology,edited by JohnH. Thibaut, JanetSpencer, and Robert
Carson. Englewood Cliffs,N.J.: General Learning Press.
Cohen, Albert K. 1955. Delinquent Boys. New York: Free Press.
Davidson, Bernard, Jack 0. Balswick, and Charles F. Halverson. 1980. "Factor
Analysis of Self-DisclosureforAdolescents." Adolescence 15 (60): 947-57.
Douvan, Elizabeth, and Joseph Adelson. 1966. The Adolescent Experience. New
York: Wiley.
Elliott,Delbert S., David Huizinga, and Suzanne S. Ageton. 1982. "Explaining Delinquency and Drug Use." National Youth Survey Project Report no. 21. Boulder,
Colo.: Behavioral Research Institute.
Empey, LaMar T. 1967. "Delinquency Theory and Recent Research." Journal of
Research in Crime and Delinquency 4:28-42.
. 1982. American Delinquency: Its Meaning and Construction.Homewood,
Ill.: Dorsey.
Foot, Hugh C., AntonyJ. Chapman, and Jean R. Smith, eds. 1980. Friendship and
Social Relations in Children. New York: Wiley.
Giordano, Peggy C. 1978. "Girls, Guys and Gangs: The Changing Social Context of
Female Delinquency." Journal of Criminal Law and Criminology69 (1): 126-32.
. 1983. "Sanctioning the High Status Deviant: An AttributionalAnalysis."
Social PsychologyQuarterly46 (4): 329-42.
Glaser, Daniel. 1956. "CriminalityTheories and Behavioral Images." AmericanJournal of Sociology 61 (March): 433-44.
Harris, Anthony. 1977. "Sex and Theories of Deviance: Toward a Functional Theory
of Deviant Type-Scripts." American Sociological Review 42:3-15.
Hartup, W. W. 1975. "The Originsof Friendship." In Friendship and Peer Relations,
edited by M. Lewis and L. A. Rosenblum. New York: Wiley.
. 1983. "Peer Relations." Pp. 1-282 in Carmichael's Manual ofChild Psychology,4th ed., vol. 4. Edited by P. H. Mussen and E. M. Hetherington.New York:
Wiley.
Hirschi, Travis. 1969. Causes of Delinquency. Berkeley: Universityof California
Press.
1200
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
Delinquency
Honess, Terry. 1979. "Children's Implicit Theories of Their Peers: A Developmental
Analysis." BritishJournalofPsychology
70:417-24.
Iscoe, I., M. Williams, and J. Harvey. 1964. "Age, Intelligence,and Sex as Variables
in the ConformityBehavior of Negro and White Children." Child Development
35:451-60.
Jensen,Gary F. 1972. "Parents, Peers, and Delinquent Action:A Test of the Differential Association Perspective." American
JournalofSociology78:562-75.
Jourard,
SidneyM. 1971. Self-Disclosure:
AnExperimental
AnalysisoftheTransparent Self. New York: Wiley Interscience.
Kandel, Denise. 1980. "Drug and Drinking Behavior among Youth." Annual Review
of Sociology 6:235-85.
Kimmel,D. C. 1974. Adulthoodand Aging:An Interdisciplinary,
Developmental
View. New York: Wiley.
Kinney, David. 1984. "Adolescent Peer Influence: A MultidimensionalApproach."
M.A. thesis, Bowling Green State University,Bowling Green, Ohio.
Klein, Malcolm W., and Lois Y. Crawford. 1967. "Groups, Gangs and Cohesiveness."
JournalofResearchin Crimeand Delinquency4:63-75.
Kleinman, Paul H., and Irving F. Lukoff. 1978. "Ethnic Differencesin Factors Related to Drug Use." Journal of Health and Social Behavior 19:190-99.
Konopka, Gisela. 1966. The Adolescent Girl in Conflict. Englewood Cliffs, N.J.:
Prentice-Hall.
Leonard, Eileen. 1982. Women,Crime and Society: A Critiqueof Theoretical
Criminology.New York: Longman.
Liebow, Elliot. 1967. Tally'sCorner:A StudyofNegroStreetCornerman.
Boston:
Little, Brown.
McCall, George, and J. L. Simmons. 1978. Identities and Interactions. New York:
Free Press.
Mannarino, AnthonyP. 1980. "The Development of Children'sFriendships."Pp. 4563 in Foot, Chapman, and Smith, eds.
Marks, J., and Peggy Giordano. 1978. "A Qualitative Analysis of Cross-Sex Friendship." Paper presentedat the annual meetingof the Midwest Sociological Society.
Matza, David. 1964. Delinquency and Drift. New York: Wiley.
Miller, Walter B. 1973. "Race, Sex and Gangs." Society 11:32-35.
Minor, William. n.d. "Maryland Youth Survey." College Park: Universityof Maryland, Instituteof Criminal Justiceand Criminology.
Pedersen, Darhl, and Kenneth L. Higbee. 1969. "Personality Correlates of SelfDisclosure." JournalofSocial Psychology
78:81-89.
Poole, Eric D., and Robert M. Regoli. 1979. "Parental Suport, Delinquent Friends,
and Delinquency: A Test of Interaction Effects." Journal of Criminal Law and
70 (2): 188-93.
Criminology
Reisman, John M. 1979. Anatomyof Friendship. Lexington,Mass.: Lewis.
Ridgeway, Cecilia. 1983. The Dynamics of Small Groups. New York: St. Martin's.
Rittenhouse, Ruth. 1963. "A Theory and Comparison of Male and Female Delinquency." Ph.D. dissertation,Universityof Michigan, Ann Arbor.
Roll, Samuel, and Levenett Millen. 1979. "The Friend as Representedin the Dreams
of Late Adolescents: Friendship without Rose-colored Glasses." Adolescence 24
(54): 255-75.
Rubin, Zick. 1980. Children's Friendships. Cambridge, Mass.: Harvard University
Press.
Seltzer,Vivian Center.1982. AdolescentSocial Development:
DynamicFunctional
Interaction. Lexington, Mass.: Heath.
Sharabany, Ruth, Ruth Gershoni,and JohnE. Hofman. 1981. "Girlfriend,Boyfriend:
Age and Sex Differencesin Intimate Friendship." DevelopmentalPsychology17 (6):
800-808.
1201
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions
AmericanJournalof Sociology
Shaw, Clifford.1966. The Jackroller.Chicago: Universityof Chicago Press.
Sherif,Muzafer, and Carolyn W. Sherif. 1964. ReferenceGroups: Exploration into
Conformityand Deviation of Adolescents. New York: Harper & Row.
Short,JamesF. 1957. "Differential
Associationand Delinquency."Social Problems4 (3):
233-39.
. 1963. The Gang. Chicago: Universityof Chicago Press.
Short,James F., Jr., and Fred L. Strodtbeck. 1965. Group Process and Gang Delinquency. Chicago: Universityof Chicago Press.
Singleton, L. C., and S. R. Asher. 1979. "Racial Integrationand Children's Peer
Preferences:An Investigation of Developmental and Cohort Differences."Child
Development50:936-41.
Thrasher, Frederick M. 1963. The Gang. Chicago: Universityof Chicago Press.
Wattenberg,William. 1956. "DifferencesbetweenGirl and Boy Repeaters."Journalof
Educational Psychology47 (3): 137-46.
West, Lloyd, and Harvey W. Zingle. 1969. "A Self-DisclosureInventoryforAdolescents." Psychological Reports 24:439-45.
Wiatrowski, M. D., D. B. Griswold, and M. K. Roberts. 1981. "Social Control
Theory and Delinquency." American Sociological Review 46 (5): 525-41.
1202
This content downloaded from 129.1.62.221 on Thu, 10 Oct 2013 19:20:21 PM
All use subject to JSTOR Terms and Conditions