You are on page 1of 24

Effects of Sexual Activity on Adolescent Social and Psychological Development

Author(s): John O. G. Billy, Nancy S. Landale, William R. Grady and Denise M. Zimmerle
Source: Social Psychology Quarterly, Vol. 51, No. 3 (Sep., 1988), pp. 190-212
Published by: American Sociological Association
Stable URL: http://www.jstor.org/stable/2786919 .
Accessed: 22/06/2014 15:14

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to
Social Psychology Quarterly.

http://www.jstor.org

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
Social Psychology Quarterly
1988, Vol. 51, No. 3, 190-212

Effectsof Sexual Activityon AdolescentSocial and


PsychologicalDevelopment*
JOHNO.G. BILLY
BattelleHumanAffairs
ResearchCenters
NANCY S. LANDALE
University
of Chicago
WILLIAM R. GRADY
BattelleHumanAffairs
ResearchCenters
DENISE M. ZIMMERLE
BattelleHumanAffairs
ResearchCenters

Usingpanel datafroma juniorhighschoolsystemin an urbanarea ofFlorida,thisstudy


investigatesthe short-term social and psychologicalconsequencesof early sexual
involvement whenpregnancy does notresult.Fifteenoutcomesrepresentingcomponents of
an adolescent'spersonality, behavior,and environment systemsare analyzedseparatelyby
race-sexsubgroupas consequencesofintercourse behavior.Six generalconclusionsresult
fromthestudy.Adolescent premarital coitus1) does notprecipitate
overwhelming changes
in an adolescent'ssocial psychologicalframework overa subsequenttwo-year interval;2)
has moresignificant effects on thesubsequentattitudesand behaviorsof whitesthanof
blacks;3) leads to morepositiveattitudes towardsexfor all race-sexsubgroups;4) has a
strongnegativeeffecton the self-reported academic grades of whitemales; 5) affects
negativelytheimportance placed on goingto collegeamongwhitefemales;and 6) has a
strongpositiveeffecton theselectionbywhitemalesand whitefemalesoffriendswhoare
sexuallyactive.

The increasingrate of sexual activity 1975, 1977;Jorgensen and Sonstegard1984),


amongteenagersin theUnitedStates,espe- andsocialcharacteristics (Billy,Rodgers,and
cially amongwhitefemales,has been well Udry 1984; Billy and Udry 1985a, 1985b;
documented (Darling,Kallan, and VanDusen Daughertyand Burger1984; Fox and Inazu
1984; Pratt,Mosher, Bachrach,and Horn 1980; Hogan and Kitagawa 1985; Moore,
1984; Udry, Bauman, and Morris 1975; Peterson,and Furstenburg 1984; Mott 1984;
Zelnik and Kantner1977; Zelnik, Kantner, Sack, Keller, and Hinkle 1984; Shah and
andFord1981). In keepingwiththisincrease, Zelnik 1981; Thorntonand Camburn1983)
theperiodsince1970has seena largevolume havebeenfoundto be important predictors
of
of scientificresearchdirectedat examining the onset and extentof sexual activity.A
the causes and consequencesof premarital numberof later-life consequencesof adoles-
sexual activityamongadolescents.The ante- cent sexual behavioralso have been estab-
cedentsof adolescentsexualinvolvement are lished. Yet almost all the researchon the
becomingwell understood. Adolescents'bio- consequencesof adolescentsex pertainsonly
(Billy and Udry1983; to theindirect
logical characteristics effectsthrough pregnancy and
Morris,Mallin,and Udry1982; Udry,Billy, childbearing. Indeed,theincreasein premar-
Morris,Groffand Raj 1985; Udry,Talbert, ital adolescentpregnancyis the impetusfor
and Morris1986), psychologicalcharacteris- muchof theresearchon boththecauses and
tics (Christopherand Cate 1984; Gilbert, theconsequencesof adolescentsexualbehav-
Bauman,and Udry 1983; Jessorand Jessor ior. The focuson premarital pregnancy stems
fromits consistentlynegativeconsequences
fortheadolescent.'
* This researchwas supportedby the Office of
Population
Affairs,DHHS, underGrantNo. APR-000918-
01-1. Address correspondenceto John O.G. Billy, ' For example, premaritaladolescent pregnancies
BattelleHuman AffairsResearchCenters,4000 N.E. generallyare unplannedand unwanted(Furstenberg
41stStreet,Seattle,WA 98105. The viewsand opinions 1976; Presser 1974; Vincent 1961) and often are
expressedin this paper do not necessarilyreflectthe terminatedby abortion (Campbell, 1980). Further,
viewsor policiesof theOfficeof PopulationAffairs,
the pregnant adolescentswho chooseto carrytheirpregnan-
BattelleMemorialInstitute,
ortheUniversity ofChicago. cies to termsufferdisadvantageous consequenceswith

190

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
SEXUAL ACTIVITY AND ADOLESCENT DEVELOPMENT 191
Althoughit is commonlybelieved that Shafer 1985) and cervical cancer (Rotkin
earlyinitiation of sex has a harmful effecton 1967); 2) Kinsey, Pomeroy,Martin,and
youth,previousresearchdoes not indicate Gebhard's (1953) pioneeringstudy,which
whether premarital adolescentsexual activity speculatesthatpremarital coituselicitslittle
affectslater life-courseevents that do not "psychologicaldisturbance" and mayin fact
operatethrough pregnancy and childbearing. precipitatehealthiernonsexualrelationships
Thisimportant topichas beenneglectedin the and increasedmaritalsatisfaction and happi-
research literaturebecause of the more ness; and 3) Gilbertet al. (1983), who finda
immediate and moreapparentneed to under- positivecausal linkbetweenadolescentsexual
standtheconsequencesof adolescentparent- activityand the subjectiveexpectedutility
hood. This studyinvestigates the short-term(SEU) forengagingin sex.
social and psychologicalconsequences of Despite the paucityof empiricalresearch
early sexual involvementwhen pregnancy
on thisissue,theoretical support existsforthe
does not result.We examineoutcomesthat
belief that premarital sexual activity at a
occurwhilethe individualis stillan adoles-
centbecause "whattranspires in thisperiod youthful age will have negative social and
has inescapableimplications and continuing psychological consequences. Most notably,
consequencesforlaterlife"(JessorandJessor JessorandJessor(1975, 1977) haveapplieda
1975, p. 5). In view of the importance of middle-range theoryof thesocial psychology
adolescence as a formativeperiod, our of problembehaviorto sexualactivity during
analysis is the logical first step toward adolescence. This framework rests on the
understanding theconsequencesof adolescent premise that the psychological and social
sexualinvolvement foroutcomesunrelated to development of individualsis a coherentand
pregnancy and childbirth. holisticprocess; behavioralchangesdo not
To identify specificshort-term social and occur in isolation from other aspects of
psychologicalconsequencesof early sexual development.Thus, observedchanges in a
involvement, we use panel data collectedin particular behaviorare inevitably a partof a
1980 and 1982 frompublic school students broader process of change involvingthe
(Grades 7 through9) in an urbanarea of personality, theenvironment, andotherbehav-
Florida.This data set includesmeasuresof iors.
sexual experienceand a wide varietyof Fundamental to thisperspective is theidea
outcomesat bothpoints;thusitis a recentand that societies are stratified by age. Social
highlyappropriate data set foraddressing the normsdefinethe behaviorsappropriate for
issuesfundamental to ourstudy. personsof a givenage and thegeneralrange
of ages at which life transitions should be
BACKGROUND
made (Hogan and Astone 1986). The prema-
tureadoptionof behaviorsordinarily reserved
Verylittleresearchhas beenconductedon foradultsbothreflectsand has implications
either the short-termor the long-term forotheraspectsof adolescentdevelopment.
nonpregnancy-related consequences ofpremar- Particularly important are "transition" behav-
ital sexual behavior.The exceptionsare 1) iorswhichmark,bothfortheadolescentand
recentreportslinkingadolescentsexual in- for others,a change in age-gradedstatus.
volvementto such healthrisks as sexually Initial engagementin such behaviorsmay
transmitted diseases (McGregor1985; U.S. representa claim on adult status; as a
CentersforDisease Control),pelvic inflam- consequenceitmaylead toattitudinal changes
matorydisease (Washington,Sweet, and and the adoption of additionalbehaviors
consideredinappropriate foradolescents.
respect to completededucation (Furstenbera1976), The initiation of sexual activityis such a
occupationalattainment(McLaughlin1977), and marital transitionbehavior. Despite the increased
stability
(McCarthy and Menken1979). Teenagemothers
also are mor-elikelyto relyon some formof public prevalence(Darlinget al. 1984) and accep-
assistance than are older mothers(Haggstrornand tance(GallupReport1985) of sexualactivity
Morrison1981; Moore and Hofferth1978:; Mott and outsidemarriage in theUnitedStates,adoles-
Maxwell 1981). The consequencesof adolescentparent- cent
foradolescentmaleswhomarTyearlyas
hoodare simnilar
premarital coitus, especiallyduringthe
a resultof a pregnancy (Freedmanand Coombs 1966; earlyteenageyears,remainsunacceptable in
Ha;gstromand Morrison1981. the eyes of mostadults(Reiss 1970). Thus

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
192 SOCIAL PSYCHOLOGY QUARTERLY
participation in sexual intercourseduringthe thistaxonomy to exploretheconsequencesof
earlyadolescentyearsmarksa departure from earlysexual activity.Specifically,we exam-
social normsdefining acceptablebehaviorfor ine a set of outcomesrepresenting compo-
thatage or lifestage.Accordingly itbecomes nentsof these systemsas consequencesof
importantfor the young person and has intercourse behavior.The predictedrelation-
implications forotheraspectsof adolescent shipbetweensexualinvolvement and each of
development. theseoutcomesis detailedbelow.
The applicability of the problembehavior
theoryto sexual activityin adolescenceis Personality System
supported byJessorandJessor's(1975) study
of thetransition to nonvirginity amonghigh FollowingJessorand Jessor(1977), we
school studentsin a smallcityin theRocky classify the elements of the personality
Mountainregion.Conductedbetween1969 systemintothreecomponent structures:
1) the
and 1972, thestudyexaminedcross-sectional motivational-instigation structure,2) theper-
relationships betweenintercourse status(virgin sonal belief structure, and 3) the personal
vs. nonvirgin)and measuresof attitudinal controlstructure. The motivational-instigation
conventionality and devianceproneness.The structure encompassesvalues and goals that
associationssuggestedthatadolescentswho are importantfor youth. One such goal
had engagedin sexual intercourse were less examinedinthisanalysisis academicachieve-
conventional and moreproneto transition or ment;a highvalue on academicachievement
deviance than those who were virgins. is consistentwith societal expectationsand
Further, thestudyexaminedthetransition to thusreflects a conventional orientation.
Early
intercourse overa one-yearperiodand found sexual activity,however,is considereda
thatantecedentmeasuresof conventionalitycontranormative behaviorand mayindicatea
withrespectto bothattitudesand behaviors shiftaway fromconventionality. Therefore
were significant predictorsof the onset of the model predictsthatsexual activitywill
sexual activity. Thus the findingswere lowerthevalue placed on academicachieve-
consistent withtheexpectations generated by ment.
thetheoretical perspective.The studydidnot, The personal belief structureincludes
however,examinetheconsequencesof sexual beliefsaboutselfand aboutselfin relationto
activity duringtheadolescentyears. society.Locus of controland self esteem,
Ourresearchis notintended as a testof the whichfallwithinthiscategory,are variables
Jessorand Jessortheoryof "problembehav- includedin our analysis.In general,sexual
ior." Insteadtheirmodel providesa useful involvement is expectedtojeopardizepositive
starting pointforourexamination of potential self-esteembecause the behavior violates
negativeconsequencesof early adolescent conventionalsocial norms. Sexual activity
sexual activity.Althoughthe model was also is expectedto lead to a shifttowardan
developedinitiallyto explainthe acquisition externallocus of control.This expectation
of problembehaviors,the mechanismsused followsfromthe factthatan externallocus
applyto consequencesas well as to causes. eliminatesthe notion of "inappropriate"
Bidirectionalcausality is implied by the behaviorbecauseone's behavioris dependent
model, as Jessorand Jessormake clear in on externalforces.
theirstatement that "transitions definedby The personalcontrolstructure consistsof
the timingof onset of various problem key personality controlsthatreflectpersonal
behaviorshave reverberating implications for moralstandards. Theseincludesexualpermis-
the restof adolescentdevelopment"(1977, siveness,religiosity, and positiveand nega-
pp. 4-5). tive reasonsforengagingin sex (subjective
expectedutility).Earlysexualinvolvement is
expectedto be relatedpositivelyto sexual
THE MODEL
permissiveness and negativelyto religiosity.
Accordingto the model developed by These relationships followfromour concep-
Jessorand Jessor(1977), the adolescent's tualization of adolescentsexualbehavioras a
social-psychological framework encompasses transition in statusawayfromconventionality
threedevelopmental systems:thepersonality andnormative adherence, andfromthenotion
system,the behaviorsystem,and the per- of a coherentpatternof change in an
ceived environment system.Our modeluses adolescent'ssocial-psychological system.That

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
SEXUAL ACTIVITY AND ADOLESCENT DEVELOPMENT 193
is, societyin generalandorganizedreligionin The specificconventionalbehaviorsexam-
particularview sexual behavior as most ined in thisstudyare churchattendance and
appropriateduring adulthood and within academic performance. The model predicts
marriage(Marshalland Suggs 1971). Adher- thatearlysexual activitywill reducechurch
ence to this belief is inconsistentwith attendance and will impedeacademicperfor-
premarital adolescentsexualactivity.Subjec- mance. These relationshipsare expected
tiveexpectedutility(SEU) is a conceptused because in general,problembehaviorsand
to representtheweighingofexpectedpositive conventional behaviorsare viewedas incom-
andnegativeconsequencesof a behavior,and patible.To thedegreethata contranormative
oftenhas beenusedas a predictorofengaging behaviorsuchas sexualactivity has occurred
in a behavior (see, e.g., Bauman 1980; and has become important,such activity
Bauman and Udry 1981; Beach, Campbell, should "interferewith, preempttime and
and Townes 1979; Bowman and Fishbein opportunity for,and locate one in a context
1978). Althougha positiveutilityforsexual less conduciveto" engagingin conventional
activitymay precipitatethe onset of the behavior(Jessorand Jessor1977, p. 36).
behavior,we argue here thatthe behavior Further,as noted above, the violationof
itselfmay weakenthe adolescent'spersonal normsrelatedto sexual activitymay reduce
controlstructure and may lead to furtheror theadolescent'ssensitivity
to normsrelatedto
continuedsexual involvement. Thus sexual otherbehaviors.
involvementis expected to be associated
positivelywithSEU. Environment
System
A basicassumption ofthesocialpsycholog-
BehaviorSystem
is thatan adolescent'sdevelop-
ical literature
The behaviorsystemis dividedintothetwo mentis conditioned by his or hernetwork of
componentsidentified by Jessorand Jessor social relationships. Two primary but diver-
(1975, 1977): problembehaviorsandconven- gent types of relationshipsare important
tionalbehaviors.A majortaskin thestudyis duringadolescence:those withparentsand
to investigatehow sexual activity,itselfan those with peers. Referencegroup theory
elementof thebehaviorsystem,affectseach holds thatwhen individualsare faced with
of theotherelementsof thesystem. two or morealternative guidesto behavior,
The model implies that early sexual theywill conformto theexpectations of the
involvement (itselfa problembehavior)will groupperceivedas havingthegreatest control
inother over valuedresources.Duringchildhoodthe
increasetheadolescent'sparticipation
contranormative behaviors.As notedthrough- familyservesas theimportant referencepoint
outthissection,forexample,sexualinvolve- in guidingbehavior,but duringadolescence
mentmarksa claimon a morematurestatus. close friendsassume an increasingly impor-
Thus once this transitionhas occurred,the tant role (Reiss 1967, 1970). Our model
adolescentmay view otheradult behaviors impliesthatearlysexual activitywill hasten
(generallyconsideredinappropriateforadoles- thattransition. We expectthisto occur for
cents) as appropriate.This possibilityis two reasons. First,sexual involvement is a
consistentwith the positionthat transition statustransition anda claimon a moremature
behaviorsare partof a coherentpatternof status.A shiftin orientation fromparentsto
change in the adolescent'sbehavioraland peers is consistentwith such a transition.
system.Sexual behavioralso may Second, because theparentalenvironment
attitudinal is
determine the typeof social settingselected likelyto be more conventional,we expect
bytheadolescent(e.g., a "fastcrowd"),thus sexual activityto lead to a greaterinvolve-
affectingopportunities to engage in other mentin a peer context,wherethereis less
deviantbehaviors.Further,the violationof commitment to conventional normsand less
one age-gradednormmay reduce the per- controlovertransgression.
ceived costs of violatingotherage-graded Even so, thedecreasein theimportance of
norms.For thesereasons,then,we expecta parentalrelationships and theincreasein the
positiverelationshipbetweensexual involve- importanceof peer relationshipsneed not
mentand "devianceproneness,"a construct always occur simultaneously(see, e.g.,
that incorporatesa number of problem Davies and Kandel 1981). Thus we also
behaviors. dividetheenvironment systemintoitsparent

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
194 SOCIAL PSYCHOLOGY QUARTERLY
and peer componentsand investigatehow foundto have morepermissivesexual atti-
insexualactivity
participation relation- tudesthanfemales(Reiss 1967; Staples1978)
affects
ships withineach component.Specifically, and are morelikelyto have reported premar-
we examinetheeffectsof sexualbehavioron ital intercourse (Sorensen 1973; Vener and
how close the adolescentfeels to his or her Stewart 1974; Zelnik and Kantner1980).
parentsand peers. Some researchers interpretthesedifferences
In addition,we investigate the effectsof as evidenceof culturaldifferences in norms
sexual activityon the prevalenceof models surrounding thetransition to adulthood.Reiss
forsexual behavior.Our data are uniquein (1967, p. 38), for example commentsthat
thattheypermitus to examinethe actual "racialand sexualdifferences are notsimply
attitudesand behaviors reportedby the differences in degreebutreflect a ratherbasic
respondent'sfriendsas well as the respon- differencein the socioculturalsituationof
dent's perceptionof his or her friends' these groups." Similarly,Goode (1961),
behavior.Our theoretical framework implies Bernard(1966), Rainwater(1970), and Sta-
that being sexually active will lead to ples (1978), althoughdisagreeingon the
increasedavailabilityof actual models for origin of the difference,believe that the
sexual behavior.Similarity val- greater
in attitudes, sexualpermissiveness andhigherrates
ues, and behaviors should be expected of premaritalsex among blacks reflecta
eventuallyto obtainamongfriendsbecause difference in sexualnorms.
similaritymakes interactionbetween two Because behaviorsthatare not negatively
personsmore rewarding;that is, similarity sanctionedmaynotcarrythesignificance of
increaseseach member'sapproval of the transition markers,theymay be less impor-
other.Further,by reducingthe numberof tant and may have fewer implicationsfor
potentialpointsof conflict,similarity mini- otherareas of the adolescent'ssocial and
mizes the hostilityand the psychological psychologicalframework. To theextentthat
discomfort causedbycognitiveinconsistency.group differencesin sexual attitudesand
Finally, similarityprovides a means for behaviorsreflect differences in normsregard-
members ofa dyadto validatetheirrespective ing sex, we wouldexpectthosegroupswith
identities(Hallinanand Tuma 1978). more permissiveattitudesand behaviorsto
As for the adolescent's per-ceptioni of exhibitfewerconsequencesof sexualinvolve-
changesin his or herenvironment, we expect ment.Blacksand males,then,areexpectedto
adolescentswho are sexuallyactive to be evidencefewerconsequencesthanwhitesand
more likely to view theirenvironment as females. Accordinglywe analyze the out-
havinga highsexual content,even when it comes separately by race-sexsubgroup.This
does not. As notedby Davies and Kandel procedureavoids the presumptionof the
(1981), individualstendto distorttheirviews operationof majorityculturalvalues and
of othersto makethemconsistent withtheir permitsan explicitrecognition of subcultural
own attitudes and behavior.Mirande(1968) differencesin the significanceof sexual
argues that this misperceptionallows a involvement as a transition marker.
"rationalization" of the adolescent'ssexual To summarize,sexual involvementis
behavior;to theextentthatfriendsengagein expectedto affectcomponents of theadoles-
thebehavior,the adolescent'sown behavior cent's personality system,behaviorsystem,
is sanctioned. and environment system.The strengthof
theserelationships also is expectedto differ
by race and by sex. Figure 1 presentsour
Racial anidSexualDiffer-ences
conceptualmodel of the consequencesof
We expectstrongracialand sexual differ- adolescentsexual activity;as noted in the
encesinthemagnitude oftheeffects ofsexual figure,controlvariablesare introduced when
behavioron theoutcomesstudied.It is well theserelationships are examinedwithineach
documented thatblackshavemorepermissive racialgroup.Thesecontrolswillbe discussed
sexual standardsthan whites (Reiss 1964, in thefollowingsection.The curveddouble-
1967; Staples 1978) and are morelikelythan headedarrowson the right-hand side of the
whites to have had premaritalintercourse figuredenotetheinterrelatedness of thethree
(Finkel and Finkel 1975; Prattet al. 1984; systems. Our strategyfor dealing with
Sorensen1973; Udryet al. 1975; Zelnikand relationships amongtheoutcomevariablesis
Kantner1980). Further,males have been outlinedin themethodssection.

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
SEXUAL ACTIVITY AND ADOLESCENT DEVELOPMENT 195

PersonalitySystem
1. Sexual permissiveness
2. Subjectiveexpected utility
3. Locus of control
4. Self-esteem
AdolescentSexual S. Value on academic achievement
Activty 6. Religiosity

BehaviorSystem
1. Academicgrades
2. Churchattendance
3. Deviance proneness

Controls /\ F}!
Cntos\ EnvironmentSystem
1. Closeness to parents
2. Closeness to peers
3. Parent-peerorientation
4. Actualmodels forsexual activity
6. Perceivedmodels forsexual activity

Figure1. ConceptualModel of theConsequencesof AdolescentSexual Activity

DATA AND METHODS Approximately two years afterthe initial


Data Set Description interview, between January andApril1982,a
follow-upinterview was administered to the
The firstround of data collectionwas originalsampleof adolescents.Of the 1405
begunin February1980 and lasteduntilearly respondents fromthefirstround,1148 (82%)
June.The eligiblepopulation was all seventh-, completedquestionnaires in Round2. The 18
eighth-,and ninth-grade studentsat four percentwho did not participate in Round 2
selectedpublic schools in an urbanarea of includethosewho were ineligible(who had
Florida. Schools were selectedfirston the moved or had been institutionalized), who
basis of the availabilityof public lists of refused,orwhoseparentsrefusedpermission.
students andtheiraddresses.In addition,they At thetimeof thesecondinterview mostof
were selectedto ensurea sufficient sample theadolescentswerein Grades9, 10, and 11.
size and to be generallyrepresentative of They rangedin age from13 to 19, witha
otherschoolsin Floridacities. mean age of 15.9. Only those adolescents
Interviewswere conductedin the respon-
who participated in bothroundsof the data
dents' homes; an interviewer completedthe
collectioneffortare used here. In addition,
demographic background portionof theques-
tionnaireand the adolescentfilled out the femaleswho reporteda pregnancyat either
remainderof the instrumentin private. Round 1 or Round 2 (N 28) are excluded
=

Completionrateswere unusuallyhighfora from this study. Thus 1120 respondents are


studyof sexual behavior,especiallyin view available for analysis, including 390 white
of the respondents'youth.A totalof 1405 males, 163 black males,413 whitefemales,
adolescents,representing 73 percentof those and 154 blackfemales.
in thestudy,completed A particular
eligibleto participate strength of the data set is its
questionnaires.Refusalswere dividedabout longitudinal design, a rarityin studies of
equallybetweenparents'refusalsto allow the adolescentsexual behavior.Anothermajor
adolescentsto participateand refusalsby advantage is the wide range of social
eligiblerespondents. Theywere also distrib- psychologicaland behavioralcharacteristics
uted approximately equally by grade, race, in the adolescents included. The survey
and sex. Respondents rangedfrom11 to 17 instrument measuredeach adolescent'sstage
yearsof age, witha meanage of 14.1, and of pubertaldevelopment,sexual behavior,
representa wide range of socioeconomic utilitystructure, social values,parentalsanc-
levels. tions,friendshipgroup,psychological makeup,

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
196 SOCIAL PSYCHOLOGY QUARTERLY
and generaldemographic background.2Fur- The modelincludesone endogenousvari-
ther,in almostall cases identicalmeasuresof able (the outcomevariableat Time 2) and
these characteristics were obtainedat both seven exogenousvariables.The exogenous
points. The purposivesamplingdesign al- variablesare
lowedadditionaldata on social networks and
behavior-attitude relationships amongfriends 1) Time 1 stateof theoutcomevariable;
2) Time 1 intercourse;
to be obtained. Because this is not a 3) intercourse
transitionbetweenTime 1 and
probabilitysample, generalizationof the Time 2;
resultsto any largepopulationis precluded. 4) age;
This loss in generalizability, however, is 5) Time 1 pubertaldevelopment;
offsetpartially by thecontrolof interregional 6) Time 2 pubertaldevelopment; and
and interschool differences thatmightcom- 7) socioeconomicstatusof familyof origin.
bineto hiderelevanteffects.
The last four variables are included as
controlsand are discussed furtherin the
AnalyticStrategy variable measurementsection. No causal
order is assumed among the exogenous
The basic researchquestionis "If adoles- variables.
cents become sexually active, will they The analyticmodelis presentedin Figure
experiencechangesin theirsocial psycholog- 2; one model is specifiedfor each of the
ical framework?" Time 1 (1980) measuresof outcomesdevelopedearlierin theconceptual
sexual intercoursebehavior are used to modelsection.In thefigure,theTime 1 state
predictchanges in outcomesbetween1980 of the outcomevariableis a stabilityterm,
and 1982(Time2). Establishing thistemporal representing the degreeto whichan adoles-
order maximizes our confidencethat the cent'sscoreon theoutcomeat Time 1 is the
causal directionis fromsexualbehaviorto a sourceof his or herscoreon theoutcomeat
givenoutcomeand notthereverse. Time 2 (Wheaton, Muthen, Alwin, and
Summers1977). By includingthis stability
2 A potentiallyimportantvariable that was not term, each remaining coefficient
indicatesthe
context"in extent to which chanige in the outcome
includedin thedata set is the "relationship
whichthesexualactivity of theadolescentoccurred.It is variable, from Time 1 to Time 2, is
possible,forexample.thata sexualexperiencewithina
verycarinaor lovinarelationship mayprecipitatefewer dependent on the associated exogenous
neaativeconsequencesthanone ina relationship inwhich variable.3
thereis littleor no emotionalbond. Yet regardlessof the
levelofcarinathatcharacterizes therelationship
inwhich
thesexualinvolvement occurs,suchinvolvement remains - This holdstruebecausethefollowingtwoequations
contranormative in that it occurs "too early." Early are equivalent:
sexual involvement,then, is expected to precipitate Y, = a + b,Yj + b,X, + u (1)
neaative consequencesreaardlessof the relationship Y = Y, - Y, + a + b,Yj + b,X, + u (2)
contextwithinwhichittakesplace. The onlydifferencebetweentheseequationsis thattheb,

T1 Intercourse
T1 - T2 Interncu
Transition
Ti Outcome Measure

T2 Outcome Measure

Age
SES
Ti PubertalDevelopment ent
T2 PubertalDevelopment
Figure2. AnalyticModel of theConsequencesof AdolescentSexual Activity

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
SEXUAL ACTIVITY AND ADOLESCENT DEVELOPMENT 197
The Ti Intercoursevariablein Figure 2 nonvirgins at Time 1 who became sexually
representsthe adolescent'svirgin/nonvirgin activeearly(1.5 yearsormorebeforethefirst
statusat Time 1. We also includea variable interview)and 2) nonvirgins at Time 1 who
representingwhether theadolescentmadethe becamesexuallyactiverecently (less than1.5
transitionto intercourse betweenroundsof yearsbeforethefirstinterview).4
thestudy.Thus we deriveestimatesof both The model parametersare estimatedby
theeffectof virginitystatusat Time 1 andthe ordinary least-squares (OLS) regression. The
effectof the transition to nonvirginity be- coefficients are presentedin unstandardized
tweenTime 1 and Time 2 on changesin the formbecause correlationsand standardized
outcomesmeasures.Whenbothvariablesare regression coefficients are highlysensitiveto
presentin the model, the Ti Intercourse changes in variancesacross subgroupsand
coefficientindicateswhetheran adolescent over timeand therefore rendercomparisons
who is a nonvirgin at Time 1 differsfroman problematic (Markus1979). A disadvantage
adolescentwho is a virginat bothroundsof of using OLS regressionin analyzingtwo-
thestudyin termsof a changein an outcome wave panel data is thatthecoefficients may
overthetwo-year interval.Thisunambiguous be biased because of the violationof two
temporalorderingallows a clear causal assumptions:1) thatall variablesare mea-
statement regarding the effectof intercourse sured perfectlyand 2) that measurement
on a change in the adolescent's social errorsat Time 1 and Time 2 are uncorrelated
psychologicalframework.The Intercourse (Campbelland Mutran1982; Hanushekand
Transitioncoefficient addresseswhetheran Jackson 1977). LISREL was used during
adolescentwho made the intercourse transi- preliminary analyses because it allows for
tionbetweenroundsdiffers fromone who is explicitestimation of measurement error(and
stilla virginat Time2. Yet becausea change correlations betweenerrors)and leaves struc-
in virginitystatusbetweenroundsmay have tural coefficientsbias-free(Joreskogand
occurredafterthe change in the outcome, Sorbom 1978). Yet the parameterestimates
only a correlationalinterpretation can be producedby OLS regression and by LISREL
made. werevirtually identical;therefore we present
A large part of one's abilityto make resultsderivedfromOLS regression.5
inferences aboutcausalityrestson theproper We notedearlierthatthecomponents ofthe
specification of the causal time lag (Heise adolescent'ssocial psychologicalframework
1970). Because we have onlytwo roundsof are interrelated. Sexual behavior,forexam-
observationspaced approximately two years ple, maylowera respondent's gradesbecause
apart,we mustassumethatthisinterval is the of the friendshe or she keeps or because it
properlengthin whichto observechangein precipitates theonsetof otherdeviantbehav-
each outcomedue to intercourse behavior. iors that affectacademic performance. To
We can, however,adjust the sexual inter- examinethisissue(albeitin a rather rudimen-
coursevariable."Ever had intercourse" is a tary fashion) for each outcome affected
measureof cumulative experiencewhichmay significantly bysexualbehavior,we introduce
haveoccurredfarearlierthantheperiodover as a controltheTime 1 and changemeasure
which we examine outcome changes. An
adolescentwho becamea nonvirgin at a very
earlyage mayalreadyhaveadjustedhisorher 4 The criterionof 1.5 yearsfordefining"early"versus
social psychologicalframework. As a means "recent" initiationof intercourseis based on consider-
ationsof the sample size on eitherside of the cutting
of settingtheintercourse experiencecloserto pointandon howrecenttheintercourse transition
mustbe
the beginningof the two-yearinterval,we to affectan outcomechangein thesubsequenttwo-year
consideran additionalmodel for each out- interval.As in the originalmodel,the sexual behavior
come. In this alternativemodel, Ti Inter- variablesare a seriesof dummyvariables.EarlyTime 1
intercourserepresentsa contrastbetween Round 1
courseis partitioned intotwocomponents: 1) nonvirgin adolescentswho made the intercourse transi-
tion early and adolescentswho were virginsat both
inEquation1 is thestability
coefficient while
coefficient, rounds.Similarly,recentTime 1 intercourse contrasts
theb, coefficient in Equation2 represents theeffectof Round 1 nonvirgins who firsthad intercourse
shortly
Time 1 statuson a changeinthatstatusovertime.The b, beforethe firstinterviewwith adolescentswho were
in Equation2 is equal to theb, coefficient
coefficient in virginsat bothrounds.
Equation 1 minusone. Each remainingcoefficient in 5 A comparison oftheOLS regression
andtheLISREL
EquationI is identicalto itsEquation2 counterpart(see resultsfora selectedsampleofmodelsfromthisstudyis
Bohrnstedt 1969). availablefromtheauthorson request.

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
198 SOCIAL PSYCHOLOGY QUARTERLY
of each remainingoutcomethatis affected developmentvariable representsall effects
by sexual behavior.When the relatedto the hormonaland morphological
significantly
effectof sexual intercourseon an outcomeis process of puberty,includingthose effects
diminishedor renderednonsignificant after whichworkthrough biologicalprocessesand
the introduction of the control, there is thosewhichworkthrough social perceptions.
evidence of an indirecteffectthroughthe The age variable is a proxyfor all social
controlvariable. processes that vary by age, except those
whichare associatedwithpubertaldevelop-
ment.These processeswould includeage-
Var-iableMeasur^emlent gradednormsand expectations,age-graded
parentalcontrolson behavior,opportunity
behavior.Sexual activityis mea- structures,
Sexuatal social skills learned,and knowl-
suredby theadolescent'sTime 1 and Time 2 edge obtainedthroughsame-sexfriendsand
"yes" or "no" response to the question the media. Includingage, pubertaldevelop-
"Have you ever had sexual intercourse?" ment,and changein pubertaldevelopment in
Logically inconsistentresponses between themodelsis therefore a parsimonious wayof
rounds were recoded in accordance with controllingfor virtually"everythingelse"
procedures developedby Rodgers,Billy,and that might renderspurious the effectof
Udry(1982). In the second set of models, intercourse behavioron each outcomevari-
measuresof "early" and "recent"Time 1 able. Socioeconomic status of family of
intercourseare based on the respondent's originis includedas a controlbecauseof the
reportedage at firstintercoursein whole presumedimportance attachedto thisvariable
years.We assumedthatadolescentsmadethe in previousresearchon adolescents.7
intercoursetransitionat mid-age(e.g., 12.5) Age is computed from date of first
and we subtracted thisage at firstintercourse interviewand respondent'sdate of birthto
fromexactage at firstinterview. All Time 1 hundredths of a year. Confirmatory factor
nonvirgin respondents who showeda result- analysis was used to operationalizethe
ing value of 1.5 or more were coded as pubertaldevelopment8 and SES9 measures.
havinghad "early" intercourse. Those with
valuesof less than1.5 werecoded as having
had "recent"intercourse. 7 We shouldnote,however,thatin previousanalyses
In theanalyticstrategy of thisdata set, SES was foundto have verymodest,if
Controlvar-iables.
any, effectson sexual behavior.Further,Jessorand
sectionwe notedthatfourvariableswouldbe Jessor (1977) report weakandsubstantively trivialeffects
controlledin each outcomemodel:socioeco- of SES on mostof theoutcomesused in thisstudy.
nomicstatusof familyof origin,respondent's 8 The pubertal developmentfactor is based on
age, and respondent's Time 1 and Time 2 respondents' self-reports
ofspecificindicatorsofpubertal
pubertaldevelopment.6 This age groupshows development.For males thefactoris composedof 1) a
drawinashowina five different
considerablevariancein thetimingand pace Tanner-type
penis size; 2) A Tanner-type
staaes of
drawina showingfive
of puberty.Age and pubertaldevelopment different amountsof pubic hair; 3) a five-point Likert
thereforemaybe treated as separatevariables, scale of penissize; and4) a five-cateaory
responseto the
and their separate effectson intercoursequestion,"Is yourvoice lowernowthanwhenyouwere
behavior and on each outcome may be in gradeschool?" For femalesthe factorcontains1) a
Tanner-type drawinashowina five different staaes of
examined(see Billy and Udry 1983). When breast size; 2) a Tanner-typedrawina showina five
pubertaldevelopmentand age are entered different amountsof pubic hair; 3) a five-point Likert
simultaneously into a model, the pubertal scale of breastsize; 4) a no" or "yes" responseto
Have you everhad a period(menstruated)?'; and 5) a
five-cateaoryresponseto whether therespondent'sbody
6
When the two pubertaldevelopmentmeasuresare is "aboutas curvyas it was whenI was a child" or "a
enteredsimultaneously into the model, T2 pubertal wholelot morecurvythanwhenI was a child.'
development representstheeffectof changein pubertal 9 The socioeconomic statusoffamilyoforiainfactoris
development on changeintheoutcomebecausetheeffect based on measuresof the father'soccupation,mother's
ofT2 pubertaldevelopment is netoftheTI measure.The education,and totalfamilyincome.Familyincomeis
TI pubertaldevelopment coefficient representsthedirect representedin a set of 11 discreteresponsecateaories
effectof TI pubertaldevelopment (i.e., thatpartwhich rangingfrom "less than $5,000' to '$50,000 + .
does notoperateindirectlythrough theassociationof TI Mother'seducationrangesfrom"no formalschoolincy"
pubertaldevelopmentwithT2 pubertaldevelopment). to g'raduateschool-three or more years. Father's
The total effectof TI pubertaldevelopmentcan be occupationwas recodedto forma set of five status-
retrievedby summinathe two pubertaldevelopment rankedordinalcategoriesranginafrom unskilled"to
coefficients. professional.

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
SEXUAL ACTIVITY AND ADOLESCENT DEVELOPMENT 199
All factorsare derivedby principalcompo- anddeviance-proneness. 13 Higherfactor load-
nentsanalysis with varimaxrotationsepa- ings also indicate an externalversus an
ratelyforeach race-sexsubgroup.A scoreon internal locusofcontroll4 andpeerorientation
each race-sex-specificfactoris computedfor relativeto parentorientation.'5
each respondentand then is used as the Closeness to parents and to peers is
measureincludedin the analyses. A more measuredby six-point Likertscales indicating
positivefactorloadingindicatesa higherlevel how close the adolescentfeelsto his or her
of thevariable. mother,father,and best male or female
Outcomevariables.Six of theoutcomesof friend.Because the surveyinstrument does
adolescentsexualbehavioralso areoperation- notpermitus to assess whether an opposite-
alized withconfirmatory factoranalysis,as sex friend was a sexualpartner, the"closeness
describedabove. Again, a score on each to best male friend"question is used in
race-sexspecificfactoris computedforeach analyzingmales and the "closeness to best
respondent at each timepointandthenis used femalefriend"questionis used in analyzing
as a measureof a particular
outcome.A more females.Value on academicachievement also
positivefactorloadinggenerallyindicatesa
higherlevelof thevariable,as is thecase for
happyareyouwithyourself?" Highfactorscoresindicate
sexual permissiveness,'0subjectiveexpected highself-esteem and low valuesindicatelow self-esteem.
utilityof intercourse(SEU),'' self-esteem,
12 13 The deviance-pronenessfactoris based on respon-
dents'"no" or "yes" repliesto thefollowing:1) "Have
you ever drunkalcohol?"; 2) "Have you ever smoked
'0 The sexual permissivenessfactoris based on four cigarettes?"; and 3) "Have you evercheatedon a test?"
variablesthatreflectattitudinaltoleranceof unconven- These data were recoded to correct for logically
tionalsexualbehavior.The firstthreevariablesrepresent inconsistent responsesover time. Admittedly, perfor-
the adolescent's"I agree," "I don't know," and "I manceof any of thesebehaviorsindicatesonlya mild
disagree"responseto the statements: 1) "Only adults formofdeviance,buteach behavioreitheris unlawful for
shouldhavesex"; 2) "Onlymarried couplesshouldhave thisage groupor lies outsidethetolerancelimitsof most
sex"; and 3) "The sexual behaviorof youngpeople is oftheadolescents'parents.We notefurther thattheseare
theirown business and no one else's." The fourth rather weakdatabecause,likethe"everhad intercourse"
variable is a same-sex-specific statementworded, "I dichotomy, theyare insensitive to differences
betweenan
wouldnotrespecta boy/girl who had sex witha girl/boy individualwho performed a behavioronly once and
he did not love." The responsesrangefrom"strongly anotherwho does so regularly. Thus theanalyticmodels
agree" (1) to "stronglydisagree"(5). These variables will reflectonly ratherlarge and dramaticeffectsof
were coded so thatlow values of the resultingfactor sexualbehavioron a changein thisoutcome.Highfactor
scores representlow sexual permissiveness and high scores indicatemore deviance proneness;low scores
valuesindicatehighsexualpermissiveness. indicateless devianceproneness.
" A listof20 consequencesthatmightbe expectedby 14 The locus of controlfactor is a two-indicator
adolescentsif theyhad intercourse wereincludedin the construct derivedfromtherespondent's "stronglyagree"
Time 1 and Time 2 questionnaires.To measure (1) to "strongly disagree"(5) responseto how he or she
subjective respondents
probability, wereaskedto indicate feelsabouteach of thefollowingsentences:1) "Most of
howlikelytheythought each consequencewas to occurif thetimewhathappensto me is myown doing";and 2)
they had sex in the next month.The five response "Whenthingsturnout well forme, I usuallyfeelthatI
categorieswerecodedfrom"wouldhappenforsure" (4) can take much of the credit." These variableswere
to "sure it wouldn'thappen"(0). To measuredesirabil- recodedso thatlow valuesof theresulting factorscores
ity,each subjectwas askedto indicatehow muchhe or represent an externallocus of controland high values
shewouldlikeor dislikeeach consequenceto occur.The represent an internallocus of control.
fourresponsecategorieswere coded from"like very 15 Parent-peer orientationis derivedfroma factor
much" (+2) to "dislikeverymuch" (-2). Following analysis of threevariables.The firstvariable is the
theprocedure usedbyGilbertet al. (1983), we calculated respondent's"stop seeing thatfriend"(1), "see that
theproductof subjectiveprobability and desirability
for friendless" (2), or "keep seeingthatfriend as usual" (3)
each consequence.The productsof the odd-numbered replytothequestion,"If yourparents didnotwantyouto
consequencesthenweresummedto formone measureof hangaroundwithone of yourfriends,whatwouldyou
SEU and the productsof the even-numbered conse- do?" The secondvariableis theadolescent's"My parents
quencesweresummedto forma secondmeasure.Then have muchmoreinfluence thanmyfriends"(1) to "My
thesetwomeasureswerefactoranalyzedto createa latent friendshave muchmoreinfluence thanmyparents"(5)
construct of SEU; low factorscoresindicatedlow utility replyto, "Rightnow, who has the mostinfluenceon
forintercourse and highscoresindicatedhighutilityfor your day to day behavioraroundyour friends-your
intercourse. parentsor your friends?"The thirdvariable is the
12 The self-esteemfactoris based on theadolescent's respondent's"parentsmuchmore" (1) to "best friend
"nottoo" (1) to "very"(3) answerto fourquestions:1) much more" (5) replyto, "Who would you listento
"How well can you do thethingsthatyou wouldreally when you have a problem-yourparentsor yourbest
liketo do?"; 2) "How muchdo you respectyourself?"; friend?"Low values of the resultingfactorscores
3) "How gooddo youfeelabouthowyouwilldevelopas indicate more parent orientationwhile high values
a personin thenextfewyears?"4) "On thewhole,how indicatemorepeerorientation.

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
200 SOCIAL PSYCHOLOGY QUARTERLY
is basedon a six-point Likertscale indicating coefficientof variation-theabsolutevalueof
how important going to college is to the a variable'sstandarddeviationdividedby its
adolescent.Religiosityis based on a four- mean-is usefulforcomparing differences in
pointscale askinghowimportant religionis in variancesamong subgroupsthathave very
theadolescent'sdailylife. Similarly,church different means (Blalock 1972). Variables
attendance is measuredon a four-point scale derivedby factoranalysis have means of
askinghow oftentheadolescenthas gone to approximately zeroandstandard deviations of
churchin thelastyear.The academicgrades approximately one.
variableis basedon theadolescent'sreportof
whetherhe or she usuallyreceives"below
RESULTS
averagegrades"(1), "averagegrades"(2), or
"above averagegrades"(3). Before we discuss the results, a few
The surveydesignofthedatasetpermits us commentsare in order.First,in additionto
to matcha respondent's sexualbehaviorwith thecustomary procedure of discussingeffects
hisor herfriend'ssexualbehavioras reported thatare statistically significant at thep<.05
bythefriend.Thisfeatureallowsus to assess level, we also discussthosethatare signifi-
theeffectsof therespondent's sexual behav- cant at the p<. 10 level. This step is
ior on a changein his or her prevalenceof particularly important in discussingtheresults
friendsof similar sexual experience.The forblacks;itcompensates forthefactthatthe
measurement of actual models for sexual samplesof blackmalesand blackfemalesare
activity is madeintermsofthe"no" or "yes" roughlyone-third the size of the samplesof
intercourse responseby theadolescent'sbest whitemales and whitefemales.We do not
same-sexfriend.We assess perceivedmodels wantto ignorea coefficient forblacksthatis
forsexual activityby usingtherespondent's as largeas or largerthana comparableeffect
"no" (1), "don't know" (2), or "yes" (3) forwhitessimplybecauseitdoes notmeetthe
replyto thequestion,"Do youthinkthatyour usualcriterion forstatistical significance as a
bestmale (female)friendhas had sex witha resultof small sample size. If an effectis
girl (boy)?" In elicitingtheseresponseswe significant at the.05 level,we concludethata
instructed theadolescentsto referto thesame relationship exists.If a coefficient is signifi-
persons whose IDs were coded in their cantat the. 10 level,we concludethat"some
questionnaires as theirbest male and female evidence" of a relationship exists. Second,
friends. although the coefficientsof the control
Some of the outcomevariablesdescribed variables indicate that other factorsmay
above provideonly crude indicatorsof the predictsignificantly a change in a given
conceptsthattheyaredesignedto measure.In outcome,these effectsare not the central
particular,the deviance-proneness construct focusof attention of thisstudy.Therefore we
includesonly mildlydeviantbehaviors,and do notpresentor discussthem.
thepsychometric propertiesof our measures
of locusofcontrolandparent-peer orientation
Effectsof TimneI Initer-colfr-se
anid Timne
have notbeenestablished.Further, academic 1-Tiune2 Inter-colfr-se
Tranisitioni
performance is based on respondents'self-
reportsand covers broad categories.These Table 2 shows, for each of the four
shortcomings aredueto thetypesofoutcomes race-sexsubgroups,the effectsof Time 1
thatcould be operationalized fromthekinds Intercourseand IntercourseTransitionbe-
ofinformation availableat thetwotimepoints tweenTime 1 andTime2 on changesin each
in the survey. Thus althoughthis study of the 15 outcomesinvestigated, net of the
providesusefulinformation abouttheeffects controlvariables.As shownin thetable,only
of early sexual involvementon such out- two componentsof the personalitysystem
comes,thereadershouldkeepthesemeasure- (Outcomes1 through 6) are affectedconsis-
mentissues in mind when interpreting the tentlyforall race-sexsubgroups.Intercourse
results. increaseseitheran adolescent'ssexualpermis-
Table 1 presentsthe Time 1 and Time 2 siveness,subjectiveexpectedutilityforhav-
means,standarddeviations,and coefficients ing sex, or both. The TI Intercourse row
of variationof each of the variables by shows thatforeverysubgroupexceptblack
race-sexsubgroup.Also includedare mea- males, nonvirgins at Time 1 became more
suresof changein each of thevariables.The sexuallypermissive overthetwo-year interval

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
Table 1. Means,StandardDeviations,and Coefficients
of VariationforAll VariablesUsed in theAnalyses,by Race-SexSubgro

Race-SexSubgroup
WhiteMales Black Males Whit
Variables Mean S.D. V N Mean S.D. V N Mean S.D.
Sexual Behavior
1. Everhad intercourse
Time 1 .25 .43 1.74 390 .77 .42 .54 163 .09 .28
Time2 .47 .50 1.06 390 .91 .28 .31 163 .29 .46
Change .22 .42 1.88 390 .14 .35 2.47 163 .21 .40
2. Earlyvs. recentintercourse
(nonvirginsat Time 1)
Early .48 - - 93 .84 - - 122 .25 -
Recent .52 - - 93 .16 - - 122 .75 -
3. Level of sexualactivity
(nonvirginsat Time 1)
Time 1 -.02 .94 - 88 -.01 .90 - 91 -.02 .75
Time2 .05 .86 - 71 -.08 .90 - 95 -.04 .87
Change .00 1.02 - 65 .00 .99 - 67 -.00 .96
ControlVariables
1. Age
Time 1 13.99 .99 .07 390 14.38 1.07 .07 163 14.00 .97
Time2 15.83 .93 .06 390 16.15 1.06 .07 163 15.81 .92
Change 1.84 .24 .13 390 1.77 .25 .14 163 1.81 .13
2. Pubertaldevelopment
Time I .01 .94 - 383 -.01 .91 - 156 -.01 .94
Time 2 -.01 .90 - 380 -.01 .82 - 158 .00 .89
Change .00 .85 - 374 .00 .85 - 152 .00 .82
3. Socioeconomicstatus
Time I and 2 .00 .69 - 390 .00 .71 - 163 .00 .72
OutcomeVariables
1. Sexual permissiveness
Time I -.01 .90 - 371 -.02 .91 - 155 .00 .83
Time 2 -.01 .91 - 373 -.01 .93 - 153 .00 .86
Change .00 .91 - 355 .00 1.12 - 148 .00 .83
2. Subjectiveexpectedutility
Time I -.01 .92 - 365 .00 .87 - 141 -.01 .94
Time2 .01 .92 - 377 .00 .89 - 156 .01 .95
Change .00 .95 - 352 .00 1.09 - 137 .00 .95
3. Locus of control
Time I -.01 .59 - 383 -.01 .71 - 157 .00 .54
Time2 .00 .51 - 383 .00 .50 - 157 .00 .64
Change .00 .72 - 376 .00 .78 - 153 .00 .70

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
Table 1. Continued

Race-SexSubgroup
WhiteMales Black Males White
Variables Mean S.D. V N Mean S.D. V N Mean S.D.
OutcomeVariables-continued
4. Self-esteem
Time 1 -.00 .84 - 383 .00 .82 - 153 .00 .86
Time2 -.01 .88 - 384 -.04 .91 - 152 -.01 .88
Change .00 .96 - 378 .00 1.03 - 144 .00 .89
5. Value on academicachievement
Time 1 5.03 1.40 .28 386 4.66 1.70 .36 158 5.24 1.14
Time 2 4.76 1.57 .33 385 4.40 1.54 .35 136 5.10 1.27
Change -.26 1.36 5.30 381 -.35 1.85 5.23 133 -.13 1.22
6. Religiosity
Time 1 2.42 .85 .35 389 3.20 .84 .26 158 2.60 .82
Time2 2.44 .87 .36 390 3.21 .84 .26 158 2.59 .83
Change .02 .77 33.24 389 -.01 .83 63.14 153 .00 .70
7. Academicgrades
Time 1 2.31 .63 .27 389 1.88 .65 .35 162 2.51 .56
Time 2 2.28 .59 .26 386 2.01 .43 .22 161 2.38 .52
Change -.03 .60 23.03 385 .13 .78 5.93 160 -.14 .51
8. Churchattendance
Time 1 2.18 1.04 .48 386 2.51 .97 .38 156 2.31 1.04
Time2 2.05 1.05 .51 384 2.29 1.03 .45 154 2.23 1.02
Change -.13 .80 6.38 381 -.26 1.01 3.92 147 -.08 .90
9. Devianceproneness
Time 1 .01 .86 - 379 .04 .84 - 159 .00 .94
Time 2 .01 .76 - 389 .06 .88 - 157 .00 .78
Change .00 .69 - 378 .01 .62 - 155 .00 .68
10. Closenessto mother
Time 1 5.27 1.00 .19 385 5.38 1.34 .25 158 5.13 1.21
Time 2 5.00 1.08 .22 388 5.26 1.30 .25 156 4.91 1.24
Change -.28 1.04 3.75 384 -.12 1.55 13.08 152 -.26 1.18
11. Closenessto father
Time 1 4.92 1.30 .26 383 4.52 1.92 .42 153 4.64 1.39
Time 2 4.56 1.35 .30 379 4.41 1.72 .39 150 4.18 1.48
Change -.36 1.20 3.35 375 -.13 1.92 15.14 142 -.48 1.36
12. Closenessto best
same-sexfriend
Time 1 4.53 1.06 .23 375 4.64 1.50 .32 153 5.29 .91
Time 2 4.79 1.04 .22 375 4.61 1.38 .30 142 5.34 .91
Change .25 1.21 4.78 362 -.03 1.74 58.14 134 .03 1.08

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
Table 1. Continued

Race-SexSubgroup
WhiteMales Black Males Whit
Variables Mean S.D. V N Mean S.D. V N Mean S.D.
OutcomeVariables-continued
13. Parent-peerorientation
Time I .00 .76 - 370 .02 .78 - 155 -.01 .84
Time 2 .00 .81 - 381 -.01 .81 - 158 -.01 .85
Change .00 .81 - 362 .00 .94 - 152 .00 .90
14. Perceivedmodelsfor
sexualactivity
Time 1 1.62 .78 .48 368 2.48 .75 .30 151 1.26 .59
Time 2 1.94 .92 .47 365 2.46 .74 .30 136 1.61 .89
Change .32 .97 3.01 344 -.02 .98 40.80 125 .35 .92
15. Actualmodelsfor
sexualactivity
Time 1 .28 .45 1.61 370 .85 .36 .42 148 .10 .30
Time 2 .45 .50 1.11 353 .87 .33 .38 135 .29 .45
Change .18 .56 3.18 336 .04 .39 9.61 122 .20 .48

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
204 SOCIAL PSYCHOLOGY QUARTERLY
than adolescentswho were virginsat both nonvirgins at Time 1 became less religious
roundsofthestudy.Similarly, theIntercourse thanthosewho remainedvirginsat Time 2.
Transition rowshowsthatforeverysubgroup Even so, the IntercourseTransitioncoeffi-
except black males, those who made the cient, althoughnegative,is nonsignificant.
intercourse transitionbetweenroundsbecame This findingmay indicate that the time
more sexually permissivethan those who neededforsexualactivityto have a negative
remainedvirgins. effecton a whitefemale'sreligiosity is longer
Because of therelativelysmall amountof thanourtwo-year observation period.
variancein the intercourse behaviorof our OutcomeVariables7, 8, and 9 in Table 2
sampleof blackmales,we wouldnotexpect are components of the adolescent'sbehavior
to find many significant effectsof sexual system.As was the case for most of the
behavioron theoutcomesamongthisgroup. personality outcomevariables,thereare no
Even so, bothTI intercourse and makingthe significant effectsof sexual activityon the
intercourse betweenroundsincrease behavioraloutcomesof black males or black
transition
a black male's SEU significantly. Similarly, females.In contrast to thesignificant negative
havingsex eitherbeforeRound 1 or between effects ofintercourse behavioron thevalueof
roundsincreasesa whitefemale'sSEU. For academic achievementfor white females,
whitemales, however,only the Intercourse intercourse has no significant effectson white
Transition coefficient Hencefor females' self-reported
is significant. grades. Yet it has a
white males we can conclude only that verypowerful negativeeffecton whitemales'
intercourse is associatedwithan increasein self-reported grades.Time 1 nonvirgin white
SEU, withoutbeing clear as to causal males and white males who made the
direction.Althoughthereis some evidence intercoursetransitionbetween rounds are
(p<. 10) of a Time 1 intercourse effecton significantly morelikelyto exhibita decline
SEU forblackfemales,thereis no significant in academicperformance over the two-year
difference in the change in utilitybetween intervalthanthosewho remainedvirginsat
thosewhomadetheintercourse and Time 2. Thus forwhitefemalestheeffectof
transition
thosewho remainedvirgins. sexualactivity is on attitudetowardacademic
Withtheexceptionof whitefemales,there achievement, while forwhitemales it is on
is littleevidencethatsexual activityaffects present (self-reported)academic achieve-
othercomponents of an adolescent'sperson- ment.16
alitysystem.Intercourse behaviordoes not Intercourse behaviorhas no effecton white
affectsignificantly a change in locus of males' churchattendance.Althoughthereis
control, self-esteem,value on academic no evidence that Time 1 nonvirginwhite
achievement,or religiosityamong white femalesare morelikelyto showa declinein
males,black males, or black females.There churchattendancethanthosewho remained
is modest evidence (p<. 10) that white virginsat Time 2, those who made the
females who were nonvirginsat Time 1 intercourse transition betweenroundsdiffer
shiftedtoward a more externallocus of significantly fromthose who remainedvir-
controlthanthose who remainedvirginsat gins. Because we cannotbe sure that the
Time 2. Yet we do not detectan effectof intercourse transition precededthedeclinein
makingthetransition to intercoursebetween churchattendance,we may conclude only
roundsof the studyon a changein a white thatintercourse behaviorand churchatten-
female's locus of control,nor is there a dance are relatednegativelyfor white fe-
significanteffectof sexual activityon a males. Similarly,making the intercourse
changein a whitefemale'sself-esteem. With transitionis relatedpositivelyto deviance
respectto value on academic achievement, pronenessamongwhitemales,butno strong
whitefemaleswhowerenonvirgins at Time 1 causal inferencecan be made. For white
decreasedtheirappraisalof theimportance of females,sexual activityhas no effecton a
goingto collegemorestrongly thanthosewho changein devianceproneness.
were virginsat both roundsof the study.
Similarly, thereis someevidencethatmaking 16
the intercourse transition betweenroundsis It is unlikelythatdifferential
changein academic
grades between white males and white females is
relatednegativelyto a change in value on responsibleforthisfindingbecause the stabilitycoeffi-
academicachievement forwhitefemales.We cientsin the whitemale and whitefemalemodelsare
also find that white females who were similar.

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
SEXUAL ACTIVITY AND ADOLESCENT DEVELOPMENT 205
Table 2. EffectsofTime 1 Intercourse
andTime 1-2Intercourse
Transition
on Each OutcomeVariable,byRace-Sex
Subgroup

Race-SexSubgroup
OutcomeVariable WhiteMales Black Males WhiteFemales Black Females
1. Sexual Permissiveness
TI intercourse .26** -.02 .27* .52**
Intercoursetransition .34*** -.10 .47*** .56***
(N) (343) (129) (376) (119)
2. SubjectiveExpectedUtility
TI intercourse .11 .83*** .70*** .40*
Intercoursetransition .27** .70** .57*** .24
(N) (340) (129) (357) (111)
3. Locus of Control
TI intercourse -.01 -.08 -.21* .07
Intercoursetransition .02 -.09 .05 -.10
(N) (361) (142) (383) (125)
4. Self-Esteem
TI intercourse -.06 -.38 .17 .16
Intercourse
transition -.06 -.26 .02 .31
(N) (364) (133) (379) (123)
5. Value on AcademicAchievement
TI intercourse -.23 .26 -.44** -.33
Intercourse
transition -.10 .27 -.27* .26
(N) (360) (123) (379) (122)
6. Religiosity
TI intercourse -.05 .16 -.24** -.01
Intercoursetransition -.08 .38 -.11 .16
(N) (365) (136) (382) (117)
7. AcademicGrades
TI intercourse -.26*** .00 -.10 -.01
Intercoursetransition -. 19*** -.25 -.06 .11
(N) (360) (123) (379) (122)
8. ChurchAttendance
TI intercourse -.04 -.24 -.03 .25
Intercoursetransition .03 -.20 - .26** .31
(N) (365) (136) (382) (117)
9. DevianceProneness
T I intercourse .09 .08 -.02 .13
Intercoursetransition .16** .02 .05 .10
(N) (357) (131) (384) (119)
10. Closenessto Mother
TI intercourse .03 .17 .04 .18
Intercoursetransition -.39*** -.77* .08 .09
(N) (359) (127) (372) (111)
II. Closenessto Father
TI intercourse .06 -.55 -.26 .08
Intercoursetransition -.11 -.67 -.09 .22
(N) (359) (127) (372) (111)
12. Closenessto Best Same-SexFriend
TI intercourse .26* .08 -.16 -.22
Intercoursetransition .28** -.31 -.02 .02
(N) (319) (108) (325) (87)
13. Parent-Peer Orientation
TI intercourse .21 -.22 -.04 .i2
Intercoursetransition .16* .01 .09 - .36*
(N) (349) (141) (370) (123)
14. PerceivedModel forSexual Activity
TI intercourse .99*** .01 1.02*** .61***
Intercoursetransition .88*** -.07 .86*** .48***
(N) (332) (117) (371) (103)
15. ActualModels forSexual Activity
TI intercourse .22*** .01 .40*** .02
Intercoursetransition .15** .05 .23*** .07
(N) (323) (112) (354) (112)
Effectsare unstandardized
ordinary least-squares coeficients
regression and are netof controlsforage, pubertal
development,changein pubertaldevelopment, SES, and theTime 1 outcomemeasure.
* p<.10.
** p<.05.
np<.Ol.

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
206 SOCIAL PSYCHOLOGY QUARTERLY
The remaining outcomevariablespresented same-sexfriend'sintercourse behavior,it has
in Table 2 arecomponents of theadolescent's no effecton actualsame-sexfriendbehavior.
environment system.Intercourse experience We maysummarize theresultspresented in
does not alterhow close an adolescentfeels Table 2 as follows:Becomingsexuallyactive
towardhis or her fatherfor any race-sex increasesan adolescent'ssexual permissive-
subgroup.Amongmales,however,thosewho nessandsubjectiveexpectedutility forhaving
madetheintercourse transitionexperienced a intercourse; with minorexceptions,this is
reduction in perceivedclosenessto motherin true for all race-sex subgroups.For white
comparison withthosewhoremainedvirgins. males, black males, and black females,
BecauseTime 1 Intercourse is notsignificant, havingintercourse has littleeffecton other
it is unclear whethermales become less components of thepersonalitysystem.Sexual
attachedto theirmothersbecause of their activity,however,lowers the value thata
intimacy withanotherfemaleor whether they white female places on futureacademic
become less close to their mothersand achievement and decreasesherreligiosity. It
subsequentlydevelop a sexual relationship also may shifther towarda more external
with a femalepeer. Sexually active white locus of control.
males also become closer to their best In its effecton adolescents'subsequent
same-sexfriends.Consistentwiththese re- behavior,becomingsexuallyactiveat leastis
sultsis some evidence(p<. 10) of a positive associated with a decline in churchatten-
relationshipbetween sexual activity and dance among white females and with an
parent-peerorientation.Making the inter- increasein deviancepronenessamongwhite
coursetransition as opposed to remaininga males. There is ratherstrongevidencethat
virginat Time2 shiftsa whitemale towarda havingintercourse has a negativeeffecton a
peerorientation. whitemale's academicgrades.
Although thereis someevidencethatblack Depending on race-sex subgroup,some
maleswhohad intercourse betweenRounds1 componentsof the adolescent'speer-parent
and 2 of thestudybecameless close to their environment system also are affectedby
mothers, thereis no indicationthatdoingso sexual activity.Among males, makingthe
affectedeitherclosenessto theirbest same- intercourse transition
is relatedto a declinein
sex friend or theirparent-peerorientation;nor perceivedclosenessto mother.Amongwhite
does sexual activityhave anyeffecton these males this decline is accompaniedby an
outcomevariablesforwhitefemales.Among increase in perceived closeness to best
blackfemalesthereis theinexplicable finding same-sexfriend andbyan overallshifttoward
thathavingintercourse betweenTime 1 and a peer (relativeto parent)orientation. With
Time 2 shiftedtheadolescentmoretowarda theexceptionof black males, sexual activity
parentorientation, althoughit had no effect precipitatesthe perceptionthatthe adoles-
on closenessto motheror fatheror to best cent's best same-sex friendalso has had
same-sexfriend.We must note, however, intercourse. Among whitemales and white
thatthecoefficient is significantonly at the femalesthiseffectholdswhentheintercourse
p<.IO level. behavioras reportedby the friendis exam-
When we examinethe last two outcome ined.Thussexuallyexperienced whiteadoles-
variablesin Table 2, we can see thatwiththe centsare likelyto acquirefriendswho also
exceptionof black males, sexuallyexperi- are nonvirgins. There is no evidence,how-
encedadolescentsare morelikelyto perceive ever,thatintercourse statusis a criterion of
that theirfriendsare also sexuallyactive. friendship selectionamongblackadolescents.
Further,sexually active white males and Finally,adolescentpremarital sex has few
white females are more likely to have significant effectson thesubsequentattitudes
same-sexfriendswho reportbeing sexually and behaviorsof black males and black
active. Althoughthe dependentvariablesin females.Thereforeour resultsmay indicate
thetwomodelsare notin thesame metric,it thatsexual behaviorelicitsfewerchangesin
appearsthatthe effectof sexual activityon thesocial psychologicalframework of blacks
thereported intercoursebehaviorof friendsis thanof whitesbutwe mustbe verycautious
weaker than its effecton the perceived in this interpretation. The samplesof black
intercoursebehavior of friends. Indeed, malesand blackfemalesare relatively small,
althoughsexualactivity has a strongpositive makingit difficult for variablesto achieve
effect on black females' perceptionsof statisticalsignificance.In addition,thereis

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
SEXUAL ACTIVITY AND ADOLESCENT DEVELOPMENT 207
littlevariancein the intercourse behaviorof tremelycautiousin concludingthatduration
black males; most alreadywere nonvirgins since intercourse transition has littleeffect.
beforethefirstroundof datacollection. Partitioning Time 1 intercourse intoearlyand
recentcomponentscreatessmall subgroups,
whichthenarecontrasted. The analysesofthe
Effects ofEarlyvs. RecentTimeI
effects of duration sincefirstintercourse push
Intercourse
thedatato theirlimitsin havinglargeenough
To assess theeffectsof durationsincefirst contrastgroups to obtain stable parameter
intercourseon the changes in outcomes estimatesand statistical significance.17
duringtheobservation period,we partitioned
Time 1 intercourse behaviorintotwocompo-
of TimeIntercourse and Time1-Time
nents:1) nonvirgins at Time 1 who became Effects
2 Intercourse Transition, Controlling for
sexuallyactive 1.5 yearsor morebeforethe
OtherOutcomeVariables
firstinterviewand 2) nonvirgins at Time 1
who became sexuallyactive less than 1.5 Below, we examinebrieflythe relation-
yearsbeforethefirstinterview. Each of these ships amongthe componentsof the adoles-
groupsis contrasted withthosewhoremained cent's social psychologicalframework and
virginsat Time 2 (resultsnot shown). We
hypothesized thatsomeof theeffects of Time 17
In an additionalanalysis,we addressedthe ques-
1 intercourse shown in Table 2 mightbe tions,"Does howmuchsexualactivityan adolescenthas
obfuscated by includingnonvirgins at Round had affecthis or her attitudesand behaviors?"We
1 whomadetheintercourse accomplished thisbyrestricting thesamplesto thosewho
(andthe were
transition nonvirginsat Time 1 and by examiningtheeffects
subsequentsocial-psychological adjustment) of levelandchangein thelevelof intercourse activity on
severalyearsbeforethebeginning ofobserva- each of theoutcomes.TI Intercourse Level was defined
tion. This hypothesis receiveslittlesupport; as a multipleindicatorconstruct representing thenumber
the Ti Intercourse-Recent effect is not of times the adolescenthad had intercourseand the
numberof partners he or she had sex within themonth
consistentlylarger than either the Ti precedingthe Time 1 interview.IntercourseLevel
Intercourse-Early effect or the total Ti Changerepresented the difference betweenthe Time 2
Intercourseeffect presentedin Table 2. compositemeasureand its Time 1 counterpart. It was
Further, an evaluationof thedifference in the includedforthesamereasonas we includedthetransition
over the two-yearintervalin the former
models' goodness of fit with and without to intercourse
models.
partitioning of Time 1 intercourse reveals that Because of the severelyrestricted sample sizes, the
thelattermodelsare notsignificant improve- analysisof the effectsof level of intercourse provides
mentsovertheircounterparts in Table 2. only a few additionalinsights.First,althoughlevel of
The resultsof thisanalysis,however,did sexual activityis based on only 21 cases, it has a
significant negativeeffecton a whitefemale'svalue on
provideinsightintotheeffectofthetimingof academic achievement. Thereforeit appearsthatbecom-
thetransition to nonvirginityfortwooutcome inga nonvirgin has a detrimentaleffecton theimportance
variables. First, partitioning the Time 1 a whitefemaleattachesto goingto college.Amongthose
intercoursevariable reveals that the small who make the intercoursetransition,level of sexual
activityexacerbatesfurther the decline in the value
(nonsignificant) negativeeffectof thetransi- placed on academic achievement.Second, the strong
tion on value on academic achievement negativeeffectof sexual behavioron white males'
amongwhitemales,shownin Table 2, is due academic grades is not replicatedwhen level of
to the significant negativeeffectof an early intercoursebehavior is considered. Thus although
transition to intercourse. Thereforeit appears nonvirginwhite males are more likely to decline in
academicperformance thanthosewhoremainvirgins, the
that a ratherlong lag time elapses before nonvirgins degreeof sexual activityappearsto have no
sexualactivity a whitemale's valueon effecton theirgrades.This findingwouldargueagainst
affects
academicachievement. Second, it was found an interpretation of theearlierfindings as beingdue to a
thatregardlessof whena whitemale became whitemale'spreoccupation withsexualencounters andto
consequentdecline in the time devotedto studies.
sexually active, his academic grades were aFinally, Time 1 level of intercourseactivityhas a
affected negatively in theinterval overwhich significantly negativeeffecton the gradesof nonvirgin
we observe change in them. Thus the blackmales.As notedpreviously, almostall blackmales
consequencesof intercoursefor academic in the sample were nonvirginsat Time 1. Perhapsan
performance appear quite robustfor white effectof sexual behavioron academic grades is not
to whitemales,but also extendsto black males
males. A similar effectis notfound,however, unique
and is detectedwhenwe use a measureof sexualactivity
foranyotherrace-sexsubgroup. thatdiscriminates morefinelyamongmembersof this
Despite these results,we must be ex- race-sexsubgroup.

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
208 SOCIAL PSYCHOLOGY QUARTERLY
describethemechanisms thatgive riseto the affectsthisfactorthrough its negativeeffect
relationships describedabove. For each out- on presentschoolperformance.
come affectedsignificantly by intercourse The resultsalso provideinformation about
behaviorwe introduced as a controltheTime how intercourse affectsvalue on academic
1 andthechangemeasureofeach oftheother achievementamong white females. As re-
remaining outcomesthatwereaffected signif- ported previously,sexually active white
icantlybysexualbehavior(resultsnotshown). females acquire same-sex friendswho are
If introducing the controloutcome in the also sexuallyactive.Because controlling for
model reduces significantly the effectof the acquisition of sexually active friends
intercourse on theoutcomeof interest, there negatestheeffectof intercourse behavioron
is some evidence that intercourse operates theimportance of goingto college,it appears
indirectlythrough thecontroloutcome. thatthisfactoris an intermediate linkin the
Generally,regardlessof which outcome causal chain between sex and importance of
variableis introduced as a control,neitherthe college. That is, havingintercourse leads a
magnitudenor the significanceof the Ti whitefemaleto acquirefriendswho are also
Intercourseand the IntercourseTransition sexuallyactive,and this "fastcrowd" peer
effectsis alteredappreciably.An exception involvementprecipitatesa decline in the
occurs when sexual permissivenessand/or value attachedby the adolescentto future
subjectiveexpectedutilityare controlledin academicachievement.
thewhitemale and whitefemalemodels.For We had hoped that an analysis of the
example, when sexual permissivenessis effects of intercourse behavioron the out-
enteredintothewhitemalemodels,theeffect come variables, controlling for each of the
of intercoursetransitionon subjectiveex- othersignificantly affectedoutcomes,would
provideinsightsabout the causal operators
pectedutilityand on deviancepronenessis
weakenedand becomesnonsignificant. This throughwhich the original relationships
obtain. The few instances noted above
finding suggests that having intercourse
providesomeevidencethatone outcomeacts
changesa whitemale's attitudesabout sex
as an intermediate variablein anotherinter-
and thatthese morepermissiveattitudesin
course-outcome causal chain. In mostcases,
turnaffectothercomponentsof his social
however,the introduction of the outcome
psychological framework. Similarly,forwhite variables as controls does not
alter the
femalesit appearsthathavingsex gives rise original That is, the outcome
relationships.
to more sexually permissiveattitudesand variablesarenotso interconnected as to show
expectations, whichin turnaffectvalue on clearlyand conciselyhow intercourse behav-
academicachievement, andchurch ior affectsthe whole nexus of personality,
religiosity,
attendance. behavioral,and environmental outcomes.
The moststriking findingfromthisset of
analysesis thattheeffectof intercourse on a
whitemale's academicgradespersistsdespite SUMMARY AND CONCLUSIONS
the introduction of controlvariables.Hence The basic researchquestionaddressedin
thisanalysisoffersno insightintothecausal this study is whethersexual intercourse
mechanism through whichsex affectsgrades. behavioraffectschangesin variouscompo-
Given our models, we must concludethat nentsof an adolescent'spersonality, behav-
intercourse lowersa whitemale's academic ior, and environment systems.Fifteenout-
performance itdoes notappearto do comes representingcomponentsof these
directly;
so indirectlythroughsuch operatorsas systemsare analyzed separatelyas conse-
makingthe adolescentmoregenerallydevi- quencesof intercourse behavior,by race-sex
ance proneor by affecting thefriendsthathe subgroup.
keeps. Academicgrades,however,appearto We can draw six general conclusions
be the mechanismthroughwhichthe early regardingthe short-term consequences of
transitionto intercourse reducesthe impor- adolescentsexualactivity. Adolescent premar-
tanceof goingto collegeamongwhitemales. italcoitus1) does notprecipitate overwhelm-
Thatis, thefactthattheTI Intercourse-Earlyingchangesin an adolescent'ssocial psycho-
effecton theimportance ofgoingto collegeis logicalframework overa subsequent two-year
reducedsubstantially whenacademicperfor- interval; 2) has moresignificant effects on the
manceis controlledsuggeststhathavingsex subsequentattitudes and behaviorsof whites

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
SEXUAL ACTIVITY AND ADOLESCENT DEVELOPMENT 209
than of blacks; 3) leads to more positive The moststriking findingof our studyis
attitudestoward sex for all race-sex sub- thatintercourse behaviorhas a strongnega-
groups;4) has a strongnegativeeffecton the tive effecton the self-reported academic
self-reportedacademicgradesof whitemales; gradesof adolescentwhitemales. The effect
5) affects
negatively theimportance placedon is almostas strongforthosewho made the
goingto collegeamongwhitefemales;and 6) transitionto nonvirginity long before the
has a strong on theselectionby period of observationas for those who
positiveeffect
whitemalesandwhitefemalesoffriends who becamenonvirgins moreimmediately before
are sexuallyactive. theinitialobservationpoint.Thosewhomade
Takenas a whole,theset of substantivelytheintercourse transition
duringtheperiodof
interpretable, resultsis relatively observation
significant also suffereda declinein grades
small. In particular,adolescentpremarital relativeto those who remainedvirginsat
coitus has few significanteffectson the Time 2. To the extent that adolescent
subsequentattitudesand behaviorsof black premaritalcoitus has long-termeffectson
males and black females.We hypothesized academicperformance, and to theextentthat
that sexual behavior would elicit fewer school performance is a good indicatorof
changesin the social psychologicalframe- success in laterlife, premarital coitus may
workof blacks thanof whitesbecause the havefar-reaching negativeconsequencesfora
behavioris less important as a markerof the whitemale's future well-being.
transitionto adult status among blacks. It is unclearwhy becominga nonvirgin
Althoughourresultssupportthishypothesis, affectsnegatively theself-reported gradesof
we mustbe verycautiousin the interpreta-whitemales.Wheneach oftheotheroutcome
tion.The samplesof black males and black measuresis enteredas a control,theeffectof
females are relativelysmall, making it intercourse on academicperformance is not
difficultto achieve statisticalsignificance. reduced.Thuswe cannotidentify an interme-
Moreover, there is little variance in the diate variable that serves as the causal
intercourse behaviorof black males because mechanismfor the effect.Notablyhaving
most of those in the sample were already intercourse does notappearto lowera white
nonvirginsbefore the firstround of data male's gradesindirectly by makinghimmore
collection.Mostoftheblackmaleswhowere generallydevianceproneor by affecting the
nonvirgins at Time 1 also reportedan early friendswhomhe keeps. Thus in view of our
age at firstintercourse,long before the model specifications,the effectmust be
intervalover whichwe observedchangesin interpreted as direct.Perhapswhen a white
outcomes. male becomes a nonvirginhe becomes
Althoughfew outcomesare affectedby increasingly preoccupiedwithremaining sex-
intercoursebehavior,those changes in an uallyactive.This preoccupation withsex and
adolescent'ssocial psychologicalframework the time devoted to pursuingsubsequent
thatare precipitated by sexual activityare sexual encountersmay detract from his
nontrivial and policy-relevant.For all race- studies.
sex subgroups, behaviorincreases
intercourse Thisfinding meritsfurther investigation.In
eitheran adolescent'ssexual permissiveness particular,we need to specifythe causal
or the subjectiveexpectedutilityforhaving mechanismby which having sex affects
sex, or both. The positive relationship grades.Doing so mayhelpto explainwhyan
betweenhavingsex and attitudesabout the effectobtainsamong white males but not
behavioris important becauseitis at leastpart among white females. The measure of
ofthecausal mechanism bywhichintercourse academicperformance-the adolescent'sself-
behavioraffectschangesin otheroutcomes. report of his or her grades as "above
By no meansdoes havingsex accountforthe average,""average," or "below average"-
total effectof premaritalcoitus on other is rathercrude. An important line of future
changesin an adolescent'ssocial psychologi- inquiry wouldbe to studytheconsequencesof
cal framework, butit appearsthathavingsex adolescentsexualbehaviorforschoolperfor-
gives rise to more sexually permissive mance as measuredby the pupil's actual
attitudesand expectations,whichin turnmay gradepointaverage.
affectsuch outcomesas value on academic Another strikingfindingofourstudyis that
achievement, devianceproneness, religiosity, althoughthe intercourse transition does not
and churchattendance. affecta white female's grades, it affects

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
210 SOCIAL PSYCHOLOGY QUARTERLY
negatively theimportance thatshe attachesto . 1985a. "The InfluenceofMale andFemaleBest
goingto college.To theextentthata sexually Friendson AdolescentSexualBehavior."Adolescence
20:22-32.
activewhitefemalecontinues toplace a lower . 1985b. "Patternsof AdolescentFriendship and
value on academicachievement as the time Effectson Sexual Behavior." Social Psychology
approachesfor decidingwhetherto attend Quarterly48:27-41.
college, the opportunityto achieve her full Blalock,HubertM. 1972. Social Statistics.New York:
potential sexually McGraw-Hill.
mightbe lost.Interestingly,
Bohrnstedt,George W. 1969. "Observationson the
active white femalesare likely to acquire Measurement of Change." Pp. 113-133 in Sociologi-
sexuallyactivesame-sexfriends;in turnthis cal Methodology, editedby Edgar F. Borgatta.San
acquisitionappearsto affectthevalue placed Francisco:Jossey-Bass.
on academicachievement. Bowman,Carol H. and MartinFishbein.1978. "Under-
not only standingPublic Reactionto EnergyProposals: An
Finally, premaritalcoitus does Applicationof the Fishbein Model." Journal of
have a strongeffecton adolescents'percep- AppliedSocial Psychology8:319-40.
tion thattheirfriendshave had intercourse; Campbell, ArthurA. 1980. "Trends In Teenage
amongwhitemalesand whitefemalesit also Childbearingin the United States." Pp. 3-13 in
affectsstronglythe selectionof same-sex AdolescentPregnancyand Childbearing:Findings
friendswho reportbeing sexually active. from Research, edited by C.S. Chilman. NIH
PublicationNo. 81-2077. Washington,DC: U.S.
Thus the peer environmentof sexually Government Printing Office.
experienced whiteadolescentsis filledwitha Campbell, RichardT. and Elizabeth Mutran. 1982.
high (perceivedand actual) sexual content. "AnalyzingPanelData in StudiesofAging."Research
Because similarity among friendsservesto on Aging4:3-41.
Christopher, F. Scott and Rodney M. Cate. 1984.
increase each individual'sapprovalof the "Factors Involved in PremaritalSexual Decision-
otherand to validate each person'sidentity, Making."JournalofSex Research20:363-76.
theacquisitionof sexuallyactivefriends may Darling, Carol A., David J. Kallan, and Joyce E.
stimulateadditionalsexualencounters. VanDusen. 1984. "Sex in Transition,1900-1980."
Our studyis a logical firststep in a full Journalof Youthand Adolescence13:385-99.
Daugherty, Linda R. and Jerry M. Burger.1984. "The
investigationof the consequencesof adoles- Influenceof Parents,Churchand Peerson theSexual
cent sexual involvement thatdo not result Attitudesand Behaviorsof College Students."Ar-
frompregnancyand childbearing. Although chivesofSexualBehavior13:351-59.
intercoursebehaviorhas onlyrelativelyfew Davies, Mark and Denise B. Kandel. 1981. "Parental
and Peer Influenceson Adolescents' Educational
effectson changesin behaviorsand attitudes Plans: Some Further Evidence."AmericanJournalof
that occur while the individualis still an Sociology87:363-87.
adolescent,those outcomesmay have far- Finkel,MadelonL. and David J. Finkel.1975. "Sexual
reachingimplications. Most important, nega- and Contraceptive Knowledge,AttitudesandBehavior
tiveoutcomessuch as a declinein thevalue of Male Adolescents."Family,PlanningPerspectives
7:256-60.
placed on academic achievementand a Fox, GreerL. and JudithK. Inazu. 1980. "Mother-
declinein academicgradesmayalterthelife DaughterCommunication about Sex." FamilyRela-
coursesignificantly. tions29:347-52.
Freedman,Ronald and Lolagene Coombs. 1966. "Eco-
nomicConsiderations in FamilyGrowthDecisions."
REFERENCES PopulationStudies20:197-222.
Furstenberg,FrankF., Jr.1976. Unplanned Parenthood:
Bauman,KarlE. 1980. Predicting Adolescent Drug Use: The Social Consequencesof Teenage Childbearing.
The UtilityStructureand Marijuana. New York: New York:FreePress.
Praeger. GallupReport.1985. "Premarital Sex." No. 237 (June).
Bauman,KarlE. andJ.RichardUdry.1981. "Subjective Princeton, NJ.
ExpectedUtilityand AdolescentSexual Behavior." Gilbert,Marcia A., Karl E. Bauman, and J. Richard
Adolescence16:527-35. Udry.1983. "SubjectiveExpectedUtilityandAdoles-
Beach, Lee R., FrederickL. Campbell,and BrendaD. centSexual Behavior:A Structural EquationAnalysis
Townes. 1979. "SubjectiveExpectedUtilityand the of Panel Data." Unpublishedmanuscript.
Predictionof Birth-Planning Decisions." Organiza- Goode, William J. 1961. "Illegitimacy,Anomie and
tionalBehaviorand HumanPerformance 24:18-28. CulturalPenetration." AmericanSociologicalReview
Bernard,Jessie. 1966. Marriage and Family Among 26:319-25.
Negroes.EnglewoodCliffs,NJ:Prentice-Hall. Haggstrom,Gus W. and Peter A. Morrison. 1981.
Billy, JohnO.G., JosephL. Rodgers,and J. Richard Teenage Parents: TheirAmbitions anidAttaiinmwents.
Udry.1984. "AdolescentSexualBehaviorandFriend- SantaMonica:The Rand Corporation.
shipChoice. Social Forces 2:653-78. Hallinan, Maureen T. and Nancy B. Tuma. 1978.
Billy, JohnO.G. and J. RichardUdry. 1983. "The "ClassroomEffectson Changein Children'sFriend-
Effectsof Age and PubertalDevelopmenton Adoles- ships." Sociologyof Education51:270-82.
centSexual Behavior."Unpublished manuscript. Hanushek,EricA. andJohnE. Jackson.1977.Statistical

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
SEXUAL ACTIVITY AND ADOLESCENT DEVELOPMENT 211
Methodsfor Social Scientists.New York: Academic Mothers:1968 and 1979." FamilyPlanningPerspec-
Press. tives6:287-92.
Heise, David R. 1970. "Causal InferencefromPanel Pratt,William F., William D. Mosher, ChristineA.
Data." Pp. 3-27 in SociologicalMethodology, edited Bachrach, and Marjorie C. Horn. 1984. "Under-
by Edgar F. Borgattaand Gus W. Bohmstedt.San standingU.S. Fertility:Findingsfromthe National
Francisco:Jossey-Bass. Surveyof Family Growth,Cycle III." Population
Hogan, Dennis P. and Nan M. Astone. 1986. "The Bulletin39:1-42.
to Adulthood."AnnualReviewofSociology
Transition Presser,HarrietB. 1974. "EarlyMotherhood: Ignorance
12:109-30. or Bliss?" FamilyPlanningPerspectives 6:8-14.
Hogan,DennisP. and EvelynM. Kitagawa.1985. "The Rainwater,Lee. 1970. Behind Ghetto Walls: Black
Impact of Social Status, Family Structure,and Familiesin a FederalSlum.Chicago:Aldine.
Neighborhood on theFertilityof Black Adolescents." Reiss, Ira L. 1964. "PremaritalSexual Permissiveness
AmericanJournalofSociology90: 825-55. amongNegroesand Whites."AmericanSociological
Jessor,Richardand ShirleyJessor.1975. "Transition Review29:688-98.
from Virginityto Nonvirginity among Youth: A . 1967. The Social Contextof PremaritalSexual
Social-PsychologicalStudyoverTime." Developmen- Permissiveness. New York: Holt, Rinehartand Win-
tal Psychology11:473-84. ston.
_ 1977. Problem Behavior and Psychosocial . 1970. "PremaritalSex as DeviantBehavior:An
Development. New York:AcademicPress. Applicationof CurrentApproachesto Deviance."
Joreskog,Karl G. and Dag Sorbom.1978. LISREL IV: AmericanSociologicalReview35:78-87.
Analysisof Linear Relationships
Structural by the Rodgers, JosephL., JohnO.G. Billy, and-J.Richard
Methodof MaximumLikelihood.Chicago: National Udry.1982. "The Rescissionof Behaviors:Inconsis-
EducationalResources. tentResponsesin AdolescentSexualityData." Social
Jorgensen,StevenR. and JanetS. Sonstegard.1984. ScienceResearch11:280-92.
"PredictingAdolescent Sexual and Contraceptive Rotkin,I.D. 1967. "AdolescentCoitus and Cervical
Behavior:An Applicationand Test of the Fishbein Cancer:Associationof RelatedEventswithIncreased
Risk." CancerResearch27:603-17.
Model." Journal of Marriage and the Family
Sack, Alan R., JamesF. Keller,and DennisE. Hinkle.
46:43-55.
1984. "Premarital Sexual Intercourse: A Test of the
Kinsey, AlfredC., WardellB. Pomeroy,Charles E.
Effectsof PeerGroup,Religiosity, and Sexual Guilt."
Martin,and Paul H. Gebhard.1953. Sexual Behavior
JournalofSex Research20:168-85.
in theHumaniFemale. Philadelphia:Saunders.
Shah, Faridaand MelvinZelnik.1981. "ParentandPeer
Markus, GregoryB. 1979. AnalyzingPanel Data.
Influenceon Sexual Behavior,Contraceptive Use, and
BeverlyHills: Sage.
Pregnancy Experienceof YoungWomen."Journalof
Marshall,D.S. and R.C. Suggs (eds.). 1971. Human
Marriageand theFamily42:339-48.
Sexual Behavior: Variationsin the Ethnographic
Sorensen, Robert C. 1973. AdolescentSexualityin
Spectrum. New York:Basic Books.
Contemporary America.New York:World.
McCarthy,Jamesand JaneMenken. 1979. "Marriage,
Staples,Robert.1978. "Race, Liberalism-Conservatism
Remarriage,Marital Disruptionand Age at First and PremaritalSexual Permissiveness: A Bi-Racial
Birth."FamilyPlanningPerspectives11:21-30.
Comparison."Journalof Marriageand the Family
McGregor,JamesA. 1985. "AdolescentMisadventures 40:733-43."
withUrethritis and Cervicitis"Journalof Adolescent Thornton, ArlandF. and Donald Camburn.1983. "The
HealthCare 6:286-97. Influence of theFamilyon Premarital SexualAttitudes
McLaughlin,StevenD. 1977. "ConsequencesofAdoles- and Behaviors."WorkingPaper. Ann Arbor:Survey
cent Childbearingfor the Mother's Occupational ResearchCenter,Institute forSocial Research,Univer-
Attainment." Final Reportto NICHD, ContractNo. sityof Michigan.
NO1-HD-62832. Udry, J. Richard,Karl R. Bauman, and Naomi M.
Mirande,AlfredM. 1968. "Reference GroupTheoryand Morris.1975. "Changesin Premarital Coital Experi-
AdolescentSexual Behavior." Journalof Marriage ence of Recent Decade-of-Birth Cohortsof Urban
and theFamily30:572-77. AmericanWomen." Journalof Marriage and the
Moore, KristenA. and Sandra Hofferth.1978. "The Family37:783-87.
Consequencesof Age at First Childbirth:Female Udry,J. Richard,JohnO.G. Billy,NaomiMorris,Terry
Headed Familiesand WelfareRecipiency."Working Groff,and Madhwa Raj. 1985. "Serum Androgenic
Paper 1146-05, Washington, DC: UrbanInstitute. HormonesMotivateSexual Behaviorin Adolescent
Moore, KristenA., JamesL. Peterson,and FrankF. HumanMales." Fertility and Sterility43:90-94.
Furstenberg. 1984. "StartingEarly:The Antecedents Udry,J. Richard,LutherM. Talbert,and Naomi M.
of EarlyPremarital Intercourse."Paper presentedat Morris.1986. "BiologicalFoundations forAdolescent
the 1984 annualmeetingsof thePopulationAssocia- FemaleSexuality."Demography 23:217-30.
tionof America. U.S. CentersforDisease Control.1976. VD Fact Sheet
Morris,Naomi M., KatherineMallin, and J. Richard 1975, 32nd ed. Washington, DC: DHEW Publication
Udry. 1982. "Pubertal Developmentand Current No. (CDC) 76-8195.
Sexual Intercourseamong Teenagers." Paper pre- Vener, ArthurM. and Cyrus S. Stewart. 1974.
sentedat the 1982 annualmeetingof the American "Adolescent Sexual Behavior in Middle America
PublicHealthAssociation. Revisited." Journal of Marriage and the Family
Mott, Frank L. 1984. "The Patterningof Female 36:728-35.
Teenage Sexual Behaviors and Attitudes."Paper Vincent,ClarkE. 1961. Unmarried Mothers.Glencoe:
presented at the 1983 annualmeetingof theAmerican FreePress.
PublicHealthAssociation. Washington,A. Eugene, Richard L. Sweet, and
Mott,FrankL. andNan L. Maxwell.1981. "School-Age Mary-AnneB. Shafer. 1985. "Pelvic Inflammatory

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions
212 SOCIAL PSYCHOLOGY QUARTERLY
Disease and Its Sequelae in Adolescents."Journalof in the United States, 1976 and 1971." Family
AdolescentHealthCare 6:298-310. PlanningPerspectives 9:55-71.
Wheaton,Blair, BengstMuthen,Duane F. Alwin,and . 1980. "Sexual Activity,
Contraceptive
Use and
Gene F. Summers.1977. "AssessingReliabilityand Pregnancy among Metropolitan-AreaTeenagers:
in PanelModels." Pp. 84-136 inSociological
Stability 1971-1979."FamilyPlanningPerspectives 12:230-37.
Methodology, editedby D.R. Heise. San Francisco: Zelnik,Melvin,JohnF. Kantner,and KathleenA. Ford.
Jossey-Bass. 1981. Sex and Pregnancyin Adolescence. Sage
Zelnik,Melvinand JohnF. Kantner.1977. "Sexual and Libraryof Social Research.Vol. 133. BeverlyHills:
ContraceptiveExperience ofYoungUnmarried Women Sage.

JohnO.G. Billyis a ResearchScientistat theBattelleHumanAffairs ResearchCentersin Seattle,Washington.His


behaviorsand contextualeffectson the
currentresearchinterestslie in the studyof adolescentfertility-related
reproductive
and contraceptive behaviorofwomenin theUnitedStates.

NancyS. Landale is an Assistant ProfessorofSociologyat theUniversity


ofChicago. Herprimaryresearchinterest
is the social demography patterns.She is currently
of thefamily,especiallymarriageand fertility studying
male
marriagebehaviorin the 19th centuryUnitedStates and the determinants of femalefamilyheadshipamong
contemporary PuertoRicans.

WilliamR. Gradyis a ResearchScientistat theBattelleHumanAffairsResearchCentersin Seattle,Washington.


His
current includefertility
researchinterests adolescentsexandfertility
regulation, behaviors,and theretirement
process
amongoldermenand women.

Denise M. Zimmerleis a ResearchSpecialistin theHealthand PopulationResearchCenterat BattelleHumanAffairs


ResearchCentersin Seattle,Washington. Her researchinterestsincludemaritalbehavior,laborforceparticipation,
andfamilystructurein additionto adolescentsexualbehavior.

This content downloaded from 91.229.229.129 on Sun, 22 Jun 2014 15:14:50 PM


All use subject to JSTOR Terms and Conditions

You might also like