Social Well-Being

Author(s): Corey Lee M. Keyes

Source: Social Psychology Quarterly, Vol. 61, No. 2 (Jun., 1998), pp. 121-140
Published by: American Sociological Association
Stable URL: http://www.jstor.org/stable/2787065 .
Accessed: 25/09/2013 08:31

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 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SocialPsychology
Quarterly
1998,Vol. 61,No. 2,121-140

SocialWell-Being*
COREY LEE M. KEYES
EmoryUniversity

The proposal of fivedimensionsof social well-being, social

social integration,
contribution,social coherence, and social acceptance,is the-
social actualization,
oreticallysubstantiated.The theoreticalstructure, constructvalidity,and the
sourcesofthedimensions
socialstructural ofsocial well-being in
are investigated
twostudies.Itemand confirmatory factoranalysesin bothstudiescorroborate
thetheoreticalmodelof social well-being. The newscales correlateconvergently
withmeasuresof anomie,generativity, perceivedsocial constraints,community
involvement and neighborhood quality.The newscales correlatediscriminantly
withmeasuresof dysphoria,global well-being, physicalhealthand optimism.
Multivariateanalysesin bothstudiessubstantiate theclaimthatsocial well-being
is an achievement, by educationalattainment
facilitated and age. The stateand
directionofthestudyofadultfunctioning arediscussed.

What is the natureof a well-livedlife? genciesand information.

Does positivementalhealth includesocial Role theoriesand conceptsdelineatethe
challenges and criteria? Inquiry into the strainsand incongruities betweenpersonal
nature of well-beingshould embrace the and social expectations(Biddle 1986;Heiss
divisionof lifeintopublicand privatetasks, 1981) or focuson the waysin whichpeople
a distinctionthathas pervaded social psy- managetheincongruity betweenprivateand
chologicaltheory. public lifeand images (Goffman1959;also
The self,for example,is both a public see Shawand Costanzo1982).
processand a privateproduct(James1890; Finally,as adults age theypurportedly
Mead 1934). Individualdifferences in neu- encountertasksthatforcethemto chooseto
roticism and extraversion (Costa and adapt throughprivateresignationor public
McCrea 1980), self-awareness(Duval and social involvement (i.e., generativity)
Wicklund1972;Fenigstein, Scheierand Buss (Erikson1950;also see Adler 1979).The pri-
1975), self-conception (Greenwald and vate and the public sides of life are two
Pratkanis1984;Trafimow, Triandisand Goto potentialsources of life's challenges,with
1991;Triandis1989) and esteem(Luhtanen possiblydistinctconsequencesforjudginga
and Crocker 1992) characterizepeople as well-livedlife.
eitherattentive to situationalor internalexi- Despite the distinctions betweenpublic
and privatelife,the leadingconceptionsof
*Thisresearchwas supported bytheJohnD. and adultfunctioning portraywell-beingas a pri-
CatherineT. MacArthurFoundation Research marilyprivatephenomenon. The clinicaltra-
Network on SuccessfulMidlifeDevelopment (MID- dition tends to operationalize well-being
MAC), whosedirectoris Dr. OrvilleGilbertBrim.
Data forStudy1 are takenfroma local studyof throughmeasures of depression,distress,
PositiveSocialFunctioning,
supported byMIDMAC; anxiety,or substanceabuse (see, e.g.,Thoits
data forStudy2 come fromMIDMAC's national 1992).Well-beingtherefore is theabsenceof
study.I am indebtedto Hazel Rose Markus,AliceS. negativeconditionsand feelings,the result
Rossi, Dov Shmotkin,Richard Shweder,and
ElizabethJ.Thomsonfortheirsupportandfeedback of adjustment and adaptationto a hazardous
on Study1 and an earlierdraftof thispaper.The world.The psychologicaltraditionopera-
feedbackreceivedfromtheSPQ reviewers and co- tionalizeswell-beingas thesubjectiveevalu-
editor,Dr. LynnSmith-Lovin, improvedthepaper ation of lifevia satisfactionand affect(e.g.,
measurably. Specialthanksgo to JaneAllynPiliavin Andrewsand Withey1976;Bradburn1969;
and Carol D. Ryfffortheirindefatigable feedback
and guidance. Direct correspondence to the Campbell 1981; Campbell, Converse,and
Department of Sociology, 1555 Pierce Drive, Rodgers 1976; Diener 1984; Gurin,Veroff,
Tarbutton Hall,Atlanta,
GA 30322. and Feld 1960) orpersonalfunctioning (Ryff
121

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 122 SOCIAL PSYCHOLOGY QUARTERLY

1989; Ryffand Keyes 1995). Accordingto neighborhood), as well as the degree to

thisview,emotionalwell-beingis an excess which theyfeel that theybelong to their
of positiveover negativefeelings;personal communitiesand society.Social integration
psychological functioningis the presenceof draws on conceptions of social cohesion
more positivethannegativeperceivedself- (Durkheim), cultural estrangementand
attributes suchas personalgrowth. Although social isolation (Seeman), and class con-
the existingmodels emphasizeprivatefea- sciousness (Marx). In Durkheim's view,
tures of well-being, individuals remain social coordinationand healthreflectindi-
embeddedin socialstructures and communi- viduals' connectionsto each otherthrough
ties,and facecountlesssocial tasksand chal- normsand indicatetheirfondnessforsoci-
lenges.To understandoptimalfunctioning ety.Accordingto Seeman (1959,1983,1991;
and mental health, social scientists also also see Merton 1949), culturalestrange-
should investigateadults' social well-being mentis the cleavage of self fromsociety.
(also see Larson1992,1996). Estrangementis the rejectionof societyor
The purposeof thisstudyis to substanti- the realizationthatsocietydoes not reflect
ate and testa socialmodelof well-being that one's own values and lifestyle. Social isola-
reflectspositivesocialhealth.Therefore I dis- tionis the breakdownof personalrelation-
cussthesocialnatureoflifeanditschallenges, ships that provide meaning and support.
becausesuchchallengesmightbe criteria that Like Marx's conceptionof class conscious-
individualsuse to assess the qualityof their ness,social integration entailsthe construal
lives.I propose operationaldefinitionsand ofcollectivemembership and fate.
indicatorsof social well-being.In two repre- Social acceptanceis theconstrualofsoci-
sentativesamples,I examinethe theoretical ety throughthe characterand qualities of
structure of each scale,as wellas
and validity other people as a generalized category.
some of thesocial structural originsof each Individualswho illustratesocial acceptance
dimension ofsocialwell-being. trustothers,thinkthatothersare capable of
kindness,and believe that people can be
Symptomsof Life's Social Challenges industrious. Socially acceptingpeople hold
favorable views of human nature (see
Social health,or at least its absence,is a Wrightsman 1991) and feelcomfortable with
preeminentconcernin classic sociological others(Horney 1945). Social acceptanceis
theory.Despite the importanceof anomie the social analogue to personalacceptance:
and alienation, Durkheimand Marxalso dis- People who feelgood about theirpersonali-
cussed severaldimensionsof positivesocial ties and accept both the good and the bad
health.Amongthepotentialbenefitsofpub- aspectsof theirlivesexemplify good mental
lic lifeare social integrationand cohesion,a health (Fey 1955; Ryff1989). Therefore
senseofbelongingand interdependence, and social acceptance of others mightbe the
a sense of sharedconsciousnessand collec- socialcounterpart to self-acceptance.
tive fate (Durkheim1951). The benefitsof Social contribution is the evaluationof
social lifeprovidea foundationfora global one's social value.It includesthe beliefthat
definition ofa socialversionofwell-being. one is a vitalmemberof society,withsome-
Social well-being is the appraisal of thingof value to give to the world.Social
one's circumstance and functioning in soci- contribution resemblesthe conceptsof effi-
ety.Below I propose and describeseveral cacy and responsibility. Self-efficacyis the
social challenges that constitutepossible beliefthatone can performcertainbehav-
dimensions ofsocialwellness. iors(Bandura1977) and can accomplishspe-
Social integrationis the evaluation of the cificobjectives(Gecas 1989). Social respon-
qualityof one's relationshipto societyand sibilityis thedesignationof personalobliga-
community.Healthy individualsfeel that tions that ostensiblycontributeto society.
they are a part of society.Integrationis Social contribution reflectswhether,and to
thereforethe extentto whichpeople feel whatdegree,people feelthatwhateverthey
theyhave something in commonwithothers do in theworldis valuedbysocietyand con-
who constitutetheirsocial reality(e.g.,their tributesto the commonweal.This construct

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SOCIAL WELL-BEING 123
is consistentwithMarx's thesisthatpeople Social coherenceis analogousto mean-
are naturallyproductive(Israel 1971). inglessness in life(Mirowskyand Ross 1989;
Alienationis the economiccounterpart Seeman 1959,1991),and involvesappraisals
to the diminution of the perceivedvalue of thatsocietyis discernable, sensible,and pre-
one's lifeand everydayactivities. The devel- dictable.Psychologically, healthierindividuals
opmentalanalogue of social contribution is see theirpersonal lives as meaningfuland
generativemotivesand behavior.Midlife, coherent(Ryff1989).The sense of personal
accordingto Erikson(1950),is a periodwhen coherence,accordingto Antonovsky(1994),
adultscan acton theirdesireto contribute to maybe a markerof health:individualswho
societyby moldingthe nextgenerationinto have coherenceattemptto maintaincoher-
productive membersofsociety(forexample, encewhenfacedwithunpredictable and trau-
bymentoring). maticlifeevents.
Social actualizationis the evaluationof In sum,perspectivesrootedin philoso-
the potentialand the trajectoryof society. phy,social psychological theoryand cultural
This is the beliefin the evolutionof society analysisarguecogentlyforemphasizingthe
and thesensethatsocietyhas potentialwhich social equally withthe personal nature of
is beingrealizedthroughits institutions and well-being.Themes of integration,social
citizens.Healthierpeople are hopefulabout involvement,and public consciousness in
theconditionand futureof society, and they classicalsociologicaltheory, whichare reiter-
can recognizesociety's potential.Socially ated in social psychologicalconceptionsof
healthierpeople can envisionthatthey,and selfin society,suggestthe social challenges
people like them,are potentialbeneficiaries facedbyadults.
of social growth.The evaluation of social
progressis consistentwith Srole's-(1956) Social Structural SourcesofSocial Well-Being
interpretation of anomieas theevaluationof
society'spotentialby the evaluationof the People do not begin or maintain the
characterof society'scustodians(forexam- quest forsocial well-beingwith the same
ple, publicofficials).Similarly, social actual- assets.Social structureshould constrainor
ization resemblesfatalisnm (Lefcourt1982) facilitate individuals'abilityand opportunity
andpowerlessness (Seeman1991). to respondsuccessfullyto the social chal-
Parallel to self-determination,social lengesof life.Social stratification and aging
actualizationis the sense thatsocietycon- are prominentstructural aspectsof lifeand
trolsitsdestiny. The focuson the realization havebeenthesubjectofpriorresearchon the
of social potential also is similar to the sources of otherdimensionsof health and
themeof potentialand its developmentas well-being. In particular,
educationalattain-
self-realization (Maslow 1968), eudaimonic mentand processesof aginggenerallyaffect
happiness (Waterman1993), and personal instrumental resourcesand self-conceptions.
growth(Ryff1989). Optimalfunctioning, in Educationalattainmentlaunchesyoung
part, is openness to experience and the adults into specificoccupations,affecting
desireand effortto growcontinually. Social whattheyearn and theconditionsand peo-
actualizationcapturestheseideas of growth ple theyencounterdaily.Indirectly, through
and development. monetarysequelae, education determines
Social coherenceis theperceptionof the the qualityof one's housingand neighbor-
quality,organization,and operationof the hood (Karabel and Halsey 1977;Kohn 1969;
social world,and it includesa concernfor Kohn and Schooler1982;Sewelland Hauser
knowingabout the world.Healthierpeople 1975). Lower socioeconomic status (SES)
not only care about the kind of world in has been linkedconsistently to diminished
whichtheylive,but also feel thattheycan physicaland mentalhealth(N. Adler et al.
understand whatis happeningaroundthem. 1994),partiallybecause lifeat lowersocioe-
Such people do not delude themselvesthat conomiclevelsappearsto impairhealth-pro-
theylivein a perfectworld;theyhave main- moting self-conceptions (Mirowsky and
tainedor promotedthedesireto makesense Ross 1989).
oflife. Whereas higher levels of education

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
124 SOCIAL PSYCHOLOGY QUARTERLY

shouldpromotesocial well-being,the rela- (e.g.,generativity)thatmeasureshow great-

is equiv-
tionshipof age to social 'well-being ly individualsperceivethattheyare valued
ocal. Aging and age differencesinclude by others,are used as social resources(such
numerousaspectsthatare notalwaysconsis- as sourcesof advice),and contribute to oth-
tent.For example,despitetheapparentcon- ers' well-being.
strictionof social positions and activity People who feel socially integrated,
(Carstensen 1995; Riley,Kahn, and Foner close to and derivingcomfortfromothersin
1994), and the decline in physical health theircommunity, shouldfeelthattheylivein
with age, research illustrateshow adults a vitaland healthyneighborhood.Socially
seem to age successfullythroughseveral integratedindividualsthereforeshouldper-
mechanisms.(For a discussionof possible ceive theirneighbors as trustworthy and
self-mechanisms,see Heidrich and Ryff theirneighborhoods as safe.Adultsintegrat-
1996.) Moreover,while some self-concep- ed intosocietyalso should be likelyto vol-
tions(such as personalcontrol)apparently unteerto maintaintheirneighborhood, per-
decline with age (Mirowsky 1995), other haps because theyfeel theiractionswillbe
studiesfindthatadults,as theyage,feelhap- valued by othersand because theywantto
pier and moresatisfiedwiththeirlives and maintainthequalityoftheirneighborhood.
reporthigherlevels of some dimensionsof Finally,people who see life as socially
psychological well-being(Heidrichand Ryff coherentshouldalso feel thattheirprivate
1996; Ryffand Keyes 1995). The abilityto lifeis coherent.Thereforeindividualswith
age successfully suggeststhatsome facetsof higherscoreson the social coherencescale
social well-being,like psychologicalwell- shouldfeelthattheirpersonalfivesare more
being,mayincreasewithage. predicable,morecontrollable, and thusgen-
In thetwoprobability samplesI investi-
erallysensibleand understandable.
gate the hypothesisof the social natureof
In short,insofaras the new scales mea-
well-beingand the hypothesisof the social
suresocialwell-being, sociallyhealthierindi-
structural sourcesof variationin socialwell-
vidualsshouldnotregardsocietyand itscus-
being,and subjectthese hypothesesto the
todiansas unsavory, should perceivethem-
standardof replication.I operationalizethe
selves as social resources,should care for
dimensions sampleand
in a local probability
and feel safe in their communities,and
replicatethemin a larger,nationallyrepre-
shouldlead coherentpersonallives.
sentativesample of adults. Confirmatory
factoranalyses permitassessment of the Regressionanalysisof the disparitiesin
utilityof thehypothesized five-factormodel social well-being,by educationand age,per-
ofsocialwell-being. I thencorrelatethenew mitsinvestigation of thehypothesized struc-
scales withextantindicatorsand scales of tural sources of social well-being. Prior
constructs and processesthatreflectindivid- researchon thestructural sourcesofpsycho-
uals' perceptionsof positivesocial function- logicalaspectsofmentalhealthsuggeststhat
ing.In the singleexceptionI correlatethe social well-beingshouldincreaseas thelevel
new scales with a scale of anomie, which of educationrises.Aging,however,involves
measures the absence of social health. losses and the abilityto adapt and age suc-
Insofar as all of the new scales measure cessfully.Therefore,althoughsocial well-
social wellness,theyshouldcorrelatenega- being mayincreasewithage, I explorethe
tivelywithappraisalsofanomie. possibility of a nonlinear relationship
In particular,social actualization and between age and social well-being. For
social acceptanceshould correlatestrongly example,some dimensionsof social well-
with anomie because the content of the being could increasemore rapidlybecause
Srole anomiescale focuseson theperceived adults face the same social challengesthat
qualityof society(as does social actualiza- subside withage or adults adapt withage.
tion) and on the characterof otherpeople However,social well-beingcould increase
like public officials(as does social accep- less quicklywithage because thesocial chal-
tance). Social contribution,on the other lengesof lifethatmightintensify cannotbe
hand,shouldcorrelatestrongly witha scale offsetbyadults'attempts to age successfully.

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SOCIAL WELL-BEING 125

METHODS theself-administered questionnairephase,or

a combinedresponserateof61%.
Samples Table 1 reportsthedemographic charac-
Study1 is a random-digit-dialingsample teristicsof each sample.The data fromthe
(Waksberg1978) of adults age 18 or older nationalsampledata are weightedto adjust
livingin Dane County,Wisconsin.Madison forunequalprobabilities of householdselec-
is theseat ofDane Countyand thecapitalof tionand unequalprobabilities of respondent
the state.All interviews, utilizingcomputer- selection withinhouseholds. The sample
assistedtechnology, were conductedon the thesampleto match
weightalso poststratifies
telephoneby trainedpersonsat the Letters theOctober1995CurrentPopulationSurvey
and ScienceSurveyCenterat theUniversity proportions of adultson thebasis of gender,
of Wisconsin,Madison.Afterthe pretesting age,race,education, andmaritalstatus, as well
and interviewertraining,interviewswere as theproportions of adultslivingin metro-
conducted between April and June 1994. politan(nonmetro)areas and variousregions
When determining householdcomposition, (northeast, midwest,southand west) of the
interviewers selectedthehouseholdmember United States.The mean age of the Dane
whohad celebratedhisor herbirthday most Countysampleis 43.6 (SD = 15.9),compared
recently (see O'Rourke and Blair 1983).The witha mean age of 45.3 (SD = 13.5) in the
interview lasted30 minuteson average.The nationalsample.In theDane Countysample,
sample includes 373 adults,fora response fewerrespondentsare marriedand more
rateof63%. have attainedslightly higherlevelsof educa-
Study2 is a random-digit-dialingsample tion.Moreover,the Dane Countysampleis
of noninstitutionalizedEnglish-speaking mostlyCaucasian, and more households
adultsage 25 to 74, livingin the48 contigu- reportslightlyhigherhousehold incomes
ous states,whosehouseholdincludedat least thanin thenationalsample.
one telephone.1 In thefirststageof themul-
tistagesamplingdesign,the investigators Measures
selected householdswithequal probability
via telephonenumbers.At thesecondstage, Social well-being.Theoretical concep-
theyused disproportionatestratifiedsam- tionsin Study1 (Dane County)promoted
plingto selectrespondents. The samplewas operationaldefinitions thatguidedthe cre-
stratifiedby age and sex; males between ation of indicators(Appendix A). On the
ages 65 and 74 wereoversampled. basis of linguisticclarityand face validity,I
Field procedures were initiated in retained10 items(fivepositiveand fiveneg-
January1995 and lasted approximately13 ative) per scale and placed themin a ran-
months. The respondentswere contactedby dom sequence. Aftersix backgroundques-
professional personnel;thosewho agreedto tionswereasked,the50 itemswereadminis-
participate in the entirestudytookpartin a tered at the beginning of the telephone
computer-assisted telephoneinterview lasting interview. Throughthe unfoldingtechnique
30 minuteson average. Respondentsthen (Groves 1989),respondentsstatedwhether
were mailed two questionnaire booklets theyagreed or disagreed,and thenjudged
requiring about 1.5 hourson averageto com- how much(strongly, moderately, or slightly).
plete.As incentivesforparticipation in the The responseoptionsthereforerangefrom
completestudy, each respondent was offered 1,strongly disagree,to 7,strongly agree.
$20,a commemorative pen,periodicreports Interviewers weretrainedto refrain from
ofstudyfindings, and a copyofa monograph
convertingvoluntaryexpressionsof uncer-
on the study.The sample consistsof 3,032
tainty(i.e.,"don'tknow"and "neitheragree
adults,witha 70% responserateforthetele-
nor disagree") into substantiveanswers.A
phone phase and an 87% responserate for
totalof 87 respondents expresseduncertainty
1 Thissamplewas assembledbyMIDMAC,the at leastonce in response to theitems.retained
JohnD. and CatherineT. MacArthurFoundation
in Study 1. Expressions of "neitheragreenor
Research Network on Successful Midlife disagree" are coded to the midpointof the
Development. scale;"don'tknow'texpressionsare assigned

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
126 SOCIALPSYCHOLOGYQUARTERLY
Samplesfrom
ofSpecific
andComparison
Table1.Characteristics theMacArthur Studiesof
Foundation's
Midlife(Percentages)
Successful

Study1 Study2
Dane CountySample National(Weighted)Sample
(N = 373) (N = 3,032)

Agea
Young 44.2 40.5
Midlife 38.9 40.3
Older 16.9 19.2
Gender
Male 41.2 43.5
Female 58.8 56.5
MaritalStatus
Married 51.5 68.1
Nevermarried 27.9 8.6
Other 20.6 23.3
Race
Caucasian 92.2 83.8
African-American 3.0 11.4
Other 4.8 5.7
Education
Up to 12years 30.0 51.5
13to 15years 30.5 25.5
16yearsormore 39.4 23.0
Income
$14,999orless 12.4 18.3
$15,000to$24,999 19.3 16.3
$25,000to$34,999 15.7 12.2
1$35,OOOto $49,999 16.0 19.5
$50,000ormore 36.6 33.6

Young= 25-39;midlife
= 40-59;older= 60-89.Nationalsample:
aDane Countysample:Young= 18-39;midlife
= 40-59;older = 60-74.

themeanofa respondent's existingitemson a ticipation,social responsibility, and social

scale. In cases in whichtheyagreed or dis- networks.
agreedbutdidnotreportthestrength oftheir Validationcriteria, Study1. I measured
feelings,respondentsare assignedthemeanof anomie (Srole 1956) witha scale consisting
the"agree"or "disagree"side oftheresponse of threeitemsused in the General Social
scale. Thus,forexample,respondentswho Survey (GSS) (Davis and Smith 1994).
agreedwithan itembut did not knowhow Respondentsindicatedwhethertheyagree
strongly theyfeltwereimputedthemean of or disagreethat(1) "Mostpublicofficials are
the"agree"(strongly, moderately or slightly) not reallyinterestedin the problemsof the
sideoftheresponsescale. averageperson,"(2) "The lot of theaverage
Each scale in Study2 consistsof only personis gettingworse,not better,"and (3)
three items. (Because of concern about "It is notfairto bringchildrenintotheworld
respondents'fatigue,I did not include all withthewaythings look forthefuture." The
items.)In keepingwiththeself-administeredinternal(alpha) consistencyof the anomie
format, all pronounsare firstperson(in con- scale is .57,whichis identicalto theaverage
trast to second person in Study 1). The reliability (Ma = .57,SD = .04) of thesame
responseformatrangesfrom1,strongly dis- scale oversixyearsoftheGSS.2
agree, to 7, stronglyagree. Respondents Two indicators measureaspectsofglobal
weregivena midpointoptionlabeled"don't psychologicalwell-being.Respondentsindi-
know."The social well-being items were catedwhethertheyfeltvery,somewhat, a lit-
orderedrandomlyand embedded in a sec-
tionof questionsinquiringabout social par- 2 GSS data are taken from1984,1985,1987,1988,
1989,and 1990.

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SOCIAL WELL-BEING 127
tle,to notat all,satisfiedwiththeirlivesright sometimesfeelI am beingpushedaroundin
now.On the same scale, respondentswere mylife."The internal(alpha) reliabilityof
instructed to pretendfora momentthatthey theperceivedconstraint scale is .86.
knew theirlife would end next week, and Twelveitemsmeasuredtheexperienceof
thento indicatehow happytheyfeltabout dysphoricsymptoms. Respondentsindicated
the lifetheyhad been able to live.Finally, howoften, duringthepast30 days,theyexpe-
respondents indicatedwhethertheyhad ever riencedeach symptom: "noneof thetime,""a
workedwithotherpeoplein theircommunity littleofthetine,""someofthetime," "mostof
to solve a problem in their community. thetime,"or "all of thetime."The symptoms
Respondentswho had been involvedthen are feeling(1) "so sad nothingcould cheer
indicatedwhethertheircommunity activity you up," (2) "nervous,"(3) "restlessor fid-
occurredduringthepast12months. gety,"(4) "hopeless,"(5) "thateverything was
Validationcriteria, Study2. Respondents an effort," (6) "worthless," (7) "cheerful," (8)
indicatedwhethersix indicatorsof genera- "in good spirits," (9) "extremely happy,"(10)
tivitydescribedthema lot,somewhat,a lit- "calmand peaceful,"(11) "satisfied," and (12)
tle,or notat all.The internal(alpha) reliabil- "fulloflife."The internal(alpha) reliability of
ity of the modified Loyola Generativity thedysphoria scale is .92;a higherscoreindi-
Scale (based on McAdamsand de St.Aubin catestheexperience ofmorenegativeandless
1992) is .84.Higherscoresrevealindividuals positivesymptoms.
who feelthattheyhave made contributions Finally,duringthe telephoneinterview,
to society,have impartedskillsor advice to respondentsindicatedtheirphysicalhealth
others,like to teach thingsto others,have as poor,fair,good, verygood, or excellent.
had a good influenceon others,and feel Also duringthetelephoneinterview, respon-
needed by others.A scale measuringper- dentsindicatedwhether"being optimistic"
ceivedneighborhood healthconsistsof four describedthema lot,somewhat,a little,or
indicatorsintendedto measure feelingsof notat all.
neighborhoodtrustand safety.Respondents
indicatedwhetherthe followingstatements RESULTS
describetheirsituationsa lot,some,a little,
or not at all: (1) "I feelsafe beingout alone In Study1 1 retaineditemswitha correct-
in myneighborhood duringthedaytime," (2) ed item-to-scale correlationof .30 or higher
"I feelsafebeingout alone in myneighbor- as indicatorsof theirrespectiveconstructs.
hood duringthenight," (3) "I could call on a Similarly,in Study2 I retaineditems(three
neighborforhelp if I needed it," and (4) per scale) thatdid not suppressthe internal
"People in my neighborhood trusteach reliability estimateas indicators of thelatent
othe'r." The internal(alpha) reliability of the construct. One indicatorof social coherence
neighborhoodhealth scale is .66; a higher in Study2 suppressedthescale internalcon-
score indicatesa strongerfeelingof neigh- sistency;therefore I omittedit fromfurther
borhoodtrustand safety. analysis.The retaineditemsfunctionedas
Respondents also completed a scale indicators oftheirrespective latentconstructs
measuring perceived constraints,which in confirmatory factor-analytic modelsexam-
reflectshow much individuals perceive iningthe relativeexplanatorypowerof the
obstaclesand unpredictable contingencies in theoreticalmodelof socialwell-being. I used
their lives (see Lachman and Weaver PRELIS 2 (Joreskogand Sorbom1993a) to
Forthcoming). Respondents indicated create the varianceand covariancematrix,
whethertheyagree or disagree (strongly, and I estimatedtheparameters and indicesof
somewhat, or a little)witheightself-descrip- fitofthemeasurement modelsusingLISREL
tive statements, includingthe followingas 8 (Joreskog and Sorbom1993b).
examples: (1) "There is little I can do to To assess indicator-to-construct overlap
changethe important thingsin mylife,"(2) in Study1,I fittedthefive-factor theoretical
"What happens in my life is oftenbeyond modelto thevariance-covariance matrixand
mycontrol," (3) "Thereare manythingsthat observedthecompletely standardizedmodi-
interfere withwhatI wantto do," and (4) "I ficationindex(CSMI). Wheneach indicator

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 128 SOCIAL PSYCHOLOGY QUARTERLY
and latent constructis standardized,the fitand thedescriptive indicesoffitofa series
CSMI is the predictedcorrelationof each of models fittedto the variance-covariance
indicatorwitha latentconstructotherthan matricesfromStudy1 and Study2. As a rule
itsown.An indicatorwitha CSMI of .40 or of thumb,values of goodness of fit and
higheriga potentially overlapping item. adjusted goodness of fitindices at .90 or
Next I refittedthe theoreticalmodel highersignify models thatfitthe data very
estimatingthe effectof the unrelatedcon- closely.The criticalN estimatesthe sample
structon thepotentially overlappingitemto size neededto rejectthenullhypothesis fora
the variance-covariancematrix.One item givenmodel at a givenlevel of alpha (.05)
purportedlyindicatingsocial actualization (see Bollen 1989).CriticalNs exceeding200
overlapped(.40, or > completelystandard- represent adequatemodels;increasingly larg-
ized loading);therefore I omittedit. er CriticalNs representeven better-fitting
Examination of higher-ordersample models. I do not entertaintheoriesabout
moments revealedunivariate andmultivariate measurement errorcorrelations; thatis,mea-
skewnessand kurtosisin both Study1 and surement erroris presumedto be random.
Study2. Because of thesmallsamplesize in Models 1 through 3 serve as null
proportionto the numberof indicatorsin hypotheses. Model 1 assertsthatthereare as
Study1, I identified item-distribution trans- many latent constructsas indicators(i.e.,
formationsthat minimized the value of indicatorsare not correlated).In contrast,
Mardia's multivariate testof skewnessand Model 2 postulates that the relationship
kurtosis(see Bollen 1989:423).The conclu- amongindicatorsis theresultof a singlefac-
sionswerethesame regardless of itemtrans- tor,whichmeansthatall indicatorsintercor-
formations, but the transformations slightly relate at about the same level. Model 3,
enhancedthefitofeachmodel.Consequently, however,positsthatthe relationship among
I show the resultsof maximumlikelihood indicatorsis explained by a negative-item
(ML) estimationbased on the exponential and a positive-item factor.Model 3 is artifac-
transformation, whichcaused the greatest tual, because it posits thatpeople respond
reduction inmultivariate skewandkurtosis. not to thecontentbut to thevalenceof the
Because thesamplein Study2 was large scale items,agreeing with all positively
enough, I derived parameter estimates phraseditemsand disagreeing withall nega-
through weighted leastsquares(WLS). Using tively phrased items.
PRELIS 2 to estimatethe asymptotic vari- Models 4 through7 representtests of
ance-covariancematrix,I then employed whether someoftheproposeddimensions of
WLS to weight the variance-covariance social well-being are distinct.Model 4 posits
matrixbytheinverseof the asymptotic vari- twolatentconstructs: One causestheindica-
ance-covariance. As a result,I achieved(in torsofsocialintegration and socialcontribu-
theory)distribution-free,asymptotically unbi- tion,and the othercauses the indicatorsof
ased,efficientparameterestimates.In other socialcoherence, socialacceptanceand social
words,WLS produces more accurate esti- actualization.3 Model 5 posits threelatent
matesthanML as samplesize increases, espe- constructs to test the distinctness of social
ciallywhen the indicatordistributions are coherence indicators from social acceptance
highlynonnormal(see Bollen 1989;Browne and social actualizationindicators. Thus one
1984). Because of the largesamplesize,the of these constructs still
causes social integra-
multivariate skew of the indicator distribu- tion and social contribution, but another
tionsprovedhighlyinfluential. The fitof all causes onlysocial coherence,whilethethird
models was substantiallybetterwithWLS construct causessocial acceptanceand social
thanwithML, althoughtheseestimators pro- actualization. Model 6 positsfourlatentcon-
ducedidenticalconclusions. Therefore I show structs to testthe theorythatsocial integra-
onlytheestimates based onWLS.
3 Thereduced models
theoretical do notexhaust
all possiblecombinations.Thereduced theoretical
Validation:Structure models, basedonconceptual arethebest
similarity,
theoreticalhunchesabout whatconstructsmightnot
Table 2 presentsthe chi-squareindexof be distinct.

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SOCIAL WELL-BEING 129
Table 2. Fit EstimatesofAlternativeand TheoreticalConfirmatory
Models of Social Well-Being

Study1: Dane CountySample

Telephone Interview
(N= 373)
MaximumLikelihooda

Chi-
Model Square df GFI (AGFI) CN

1. Independence 3,747 464 .36 (.27) 54

2. Single Factor 1,556 464 .73 (.69) 130
3. Two-FactorArtifacts 1,521 463 .73 (.69) 132
4. Two-FactorReduced Theory 1,259 463 .79 (.76) 160
5. Three-FactorReduced Theory 1,178 461 .81 (.78) 170
6. Four-FactorReduced Theory 1,044 458 .84 (.81) 190
7. Five-FactorTheory 914 454 .86 (.84) 216

Study2: National (Unweighted)Sample

Self-Administered Questionnaire
(N = 2,887)
WeightedLeast Squares

Chi-
Model Square df GFI (AGFI) CN

1. Independence 3,372 77 .76 (.67) 94

2. SingleFactor 1,677 77 .88 (.84) 188
3. Two-FactorArtifact 1,446 76 .90 (.86) 216
4. Two-FactorReduced Theory 1,211 76 .91 (.88) 257
5. Three-FactorReduced Theory 1,002 74 .93 (.90) 304
6. Four-FactorReduced Theory 740 71 .95 (.92) 396
7. Five-FactorTheory 705 67 .95 (.92) 397

Note:GFI(AGFI) = Goodness of fitindex (adjusted goodness of fitindex); CN = criticalN.

Analysesin bothsamplesare based on imputationof mean froma respondent'sown extantitems.
aAnalysisbased on exponentialtransformationof itemdistribution.

tionis a latentconstructdistinctfromsocial indices of fitsuggestthat the theoretical

contribution. Model 7 investigateswhether five-factormodel in Study 1 fitsthe data
socialacceptanceand socialactualizationare onlyadequately.Simulationstudies,howev-
distinctlatent constructs.In other words, er,suggestthatthe averagevalues of many
Model 7 is theproposedfive-factor theoreti- descriptiveindicesdecrease as sample size
cal modelofsocialwell-being. decreases and as the numberof indicators
To examine how well a model fitsthe increases(see Bollen 1989:277).The theoret-
data,I comparedeach modelwiththeimme- ical model in Study 1 thereforefitswell
diatelypreceding model,usingthechi-square despite the smallersample size (N = 373)
contrast.The differencein chi-squarevalues and thelargenumberofindicators (32).
betweenmodelsis distributed approximately Study2 replicatesthesuperiorfitof the
chi-square;degreesof freedomare equal to theoreticalfive-factormodel. Again, chi-
the differencein the degrees of freedom squarecontrasts illustrate
thateachsuccessive
betweenmodels. and morecomplexmodelfitsthedata more
In Study1,all chi-squarecontrastsshow closely.In theend,thetheoretical five-factor
that each successive and more complex model providesan excellentfitto the data.
model providesa betterfitto the data. In The chi-squarecontrastin the theoretical
fact,the theoreticalModel 7 is the best-fit- model,in comparison withModel 6,is 8.75(p
tingmodel,witha chi-squarecontrastof32.5 < .005).Moreover,thegoodnessof fitindices
(p < .001) to Model 6. The descriptive exceed.90andtheCritical N is nearly400.

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
130 SOCIAL PSYCHOLOGY QUARTERLY
The descriptivestatistics oftheindicatorsand In Study1,all newscales correlatenega-
thescalesalso revealthequalityof thetheo- tivelywithanomie,but social actualization
reticalmodel.Onlyone estimateofthecorre- and social acceptancecorrelatemorestrong-
lationof the indicatorwithits standardized ly,as predicted,withthescale of anomie.As
latentconstruct fails
(i.e.,validitycoefficient) the evaluation of the quality of society,
the.40ruleofthumb(see Appendix
to satisfy anomierelatesto thosedimensionsof social
B). Whereastheconfirmatory factoranalyses well-beingwhichindicatethe evolutionand
supportedthe theoryof fivedistinctlatent potentialofsocietyand theperceptionofthe
factors,the descriptive statisticsfor the characterand qualitiesofgeneralizedothers.
observedscales displayedin Table 3 show Feelingsof social integration and social con-
that all scales intercorrelate positively. tribution in Study1 are higheramongindi-
Moreover, nearlyall scales,eventhereduced- vidualswho have recentlybeen involvedin
itemscalesin Study2, exhibitrelatively high theircommunities (i.e.,in community action)
internalconsistency.The scale ofsocialcoher- than among people who have never been
ence exhibitslower internalreliabilityin involved. As expected,integration is associat-
Study1 but relativelyhigherreliabilityin ed withmaintaining recentprosocialties in
Study2. On theotherhand,thescaleofsocial one's community.
acceptancewas internally in Study
consistent An unanticipated finding was the
1 buttheset of itemsdid not coherewellin strength of theassociationofrecentcommu-
Study2. These lower internalconsistencies nityactionwithsocial contribution. Viewed
maybe lowerbecause of samplingvariation in retrospect, attempting to solve a problem
and becauseI added a new itemto thescale in one's community also couldJieconsidered
ofsocialacceptancein Study2 to achievebal- productive activity thatshouldbe connected
ancebetweennegativeandpositiveitems witha senseofcontribution. Study2, howev-
er,revealsthe distinctionbetweenintegra-
Validation:
Correlations tion and contribution. As expected,social
contribution correlatesstrongly withgenera-
Table 4 displaysthe correlationsof the tivity:People who feelthattheyare socially
newscaleswithvalidition I use thez-
criteria. valuedalso feelthattheypossesstheperson-
statisticto assesswhethera particular scale of al resourcesand qualities needed to guide
socialwell-being correlatesmorestrongly, as others.Social integration, on theotherhand,
hypothesized,with a validation criterion. correlatesmorestrongly withthe perceived
When I employ the average sample size healthof one's neighborhood. Thus people
acrossall correlations in Study1 (Mn = 360), who feelclose to othersin theircommunity
the standarddeviationis .075;therefore the also feel thattheirneighborhoodsare safe
difference betweentwoz-transformed coeffi- and thattheirneighborsare trustworthy.
cientsin Study1 mustbe at least.15to be sta- socialcoherencecorrelates
Finally, strong-
tistically at the.05 alpha level.4In
significant ly,as predicted,withthe scale of perceived
Study2,thestandarddeviationofthez-statis- constraints. Individualswho findthe social
ticis .026;thez-ratiotherefore mustmeetor worldmoreunpredictable, morecomplex,and
exceed .052 to be statistically significant at moreincoherent also tendto viewtheirown
the .05 alpha level. In the discussion of livesas complexaffairs, fullofinsurmountable
results,I focusonlyon whetherthehypothe- obstaclesandunpredictable contingencies.5
sizedvalidationcorrelations are confirmed.
5 Somenewscalesin Study1 are unbalanced. The
imbalance,however, does notappearto systematical-
4 Samplesize variesforthecorrelations ofcom- lyinfluencethecorrelations.Social coherenceis the
munityactionwiththe scales of social wellness. mostimbalanced scale,consistingofonlyonepositive
Particularly is thecorrelation
interesting of recent item;theanomiescale,forexample, consistsofthree
community inwhichadultswhowereinvolved
action, negativelyphraseditems.Social actualization,a bal-
intheircommunities during thepastmonth (N = 128) ancedscale,correlates morestrongly, however, than
are contrastedwithadultswho have neverbeen socialcoherencewiththescaleofanomie.Recall,too,
involved(N = 167).As a result,thestandard devia- thattheconfirmatory two-factorartifact
model(neg-
is .082,and thez-ratiomust
tionof thez-statistic therela-
ativeand positive)did poorlyin explaining
equal or exceed.164to be statistically at
significant tionshipbetweentheitems, comparedwiththeseries
the.05alphalevel. modelsofsocialwell-being.
oftheoretical

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SOCIAL WELL-BEING 131
Table3. Descriptive (Study1 inUpperDiagonal,Study2 inLowerDiagonal)
Statistics

1 2 3 4 5

1.SocialCoherence .45 .25 .42 .45

2. SocialActualization .45 .41 .40 .53
3. SocialIntegration .23 .33 - .57 .42
4. SocialContribution .35 .31 .42 - .20
5.SocialAcceptance .23 .48 .35 .23

Study1: Dane CountySample

M 25.0 32.1 33.0 30.3 28.3
SD 4.1 5.7 6.2 5.2 6.4
NumberofItems 5 7 7 6 7
AlphaReliability .57 .69 .81 .75 .77

Study2: NationalSample
M 8.9 11.8 14.0 15.5 13.0
SD 3.3 4.3 4.4 3.8 3.5
Numberof Items 2 3 3 3 3
AlphaReliability .64 .64 .73 .66 .41

Note:Forallrs,p < .01(two-tailed)

The scales of social well-beingcorrelate, Validation: Social Structural Sources

but do not appear to overlap,with-extant
measuresreflectingpsychologicalconcep- Table 5 presentsthe regressions of each
tionsof mentalhealth.The social well-being scale of social well-beingon age and educa-
scales correlatepositively and modestlywith tion.Each model estimatesa quadratic(and
the indicatorsof life satisfactionand life linear)parameterforage to investigate the
happinessin Study1. possibleaccelerationor decelerationin the
In Study2,thesescalesexhibitconsistent- predictedincreaseof social well-beingwith
ly modestand positivecorrelations withthe age. (Only the statistically significant effects
scaleofdysphoria. Because ofthesamplesize shown, however.) All models also adjust for
andstatisticalefficiencyofStudy2,socialinte- gender, race, and marital status, which could
grationcorrelatesslightly stronger withdys- structure educationaloutcomesand possibly
phoria than does any otherscale of social social well-being. The remaining background
well-being.On theotherhand,althoughdys- variables (e.g., income) either follow from
phoria correlatesrelativelystronglywith education or explain the hypothesized age dif-
physicalhealth,all scales of socialwell-being ferences, and therefore are omitted from
displaysmall but positivecorrelationswith these models.The interceptrepresentsthe
theindicator ofsubjective physicalhealth.The predictedlevel of social well-beingforthe
newscalesofsocialwellnesstherefore appear average18-year-old (Study1) or 25-year-old
to measureaspectsof mentalhealthand life (Study 2).7
qualityrelatedto,butdistinct from, depressive
feelingsandphysicalhealth. smallcorrelations withthesocial well-being scales
None of thesocial well-beingscales cor- may reflect,in part, the mode of administrationas
wellas theeffectof thetimelag betweenthetele-
relatesverystrongly withthe perceptionof phoneinterview and theself-administeredquestion-
oneselfas optimistic (rangeofr = .17 to .23). naire.
Thisfinding suggeststhatappraisalsofsocial 7 In Study2, I showthe multivariate analyses
well-being are not unduly influencedby basedon unweighted databecausetheresults arethe
optimismor possiblythe tendencyto exag- same regardlessof sampleweighting. The compo-
nentsofthesampleweights in Study2 are notfunc-
geratepositivequalities.6 tionsoftheoutcomesin thisstudy. Therefore multi-
variatemodelsbasedon theunweighted datashould
6 Because physicalhealthand optimismare ascer- thereforebe unbiasedand efficient
ifthemodelis
tained during the initial telephone interview,the specified
correctly
(see WinshipandRadbil1994).

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 132 SOCIALPSYCHOLOGYQUARTERLY
Table4. Product-Moment
Correlations
ofSocialWell-Being
ScaleswithValidation
Criteria

1 2 3 4 5 6

Study1:Dane CountySample
1. AnomieScale -.21 -.04ns -.18 -.16
2. RecentCommunity Action(Never=0) 0.0 .27 .08ns
3. Community ActioninPast(Never=0) .04ns -.06ns
4. GlobalLifeSatisfaction .46
5. GlobalHappinesswithLife
6. SocialActualization -.55 .20 .OSns .26 .16
7. SocialAcceptance -.49 .1Sa .05ns .21 .14
8. SocialIntegration -.33 .41 .14a .32 .31
9. SocialContribution -.28 .46 .22 .27 .20
10. SocialCoherence -.28 .24 .02ns .20 .16

M 1.2 .43 .32 3.4 3.5

SD 1.1 .50 .47 .74 .67

Study2:NationalSample
1. Generativity
Scale .14 -.23 .17 .10 .24
2. HealthofNeighborhoodScale -.21 -.24 .16 .14
3. Perceived
Constraints
Scale .49 -.28 -.30
4. DysphoriaSymptomsScale -.40 -.30
5. SubjectivePhysicalHealth .15
6. Self-Described
Optimistic
7. SocialActualization .17 .21 -.32 -.25 .16 .18
8. SocialAcceptance .12 .24 -.22 -.22 .13 .17
9. SocialIntegration .28 .39 -.26 -.33 .19 .23
10. SocialContribution .49 .16 -.36 -.22 .15 .16
11. SocialCoherence .18 .13 -.41 -.28 .15 .21
M 17.0 13.6 18.7 24.9 3.5 3.3
SD 3.8 2.2 8.0 7.1 1.0 .77

Notes:Forall rs,p < .01(two-tailed)

unlessnoted.One-hundredandtwenty-eight
respondents
engaged
recently incommunity 78 engagedincommunity
action, actioninthepast,and167neverengagedincommu-
nityaction.
at alpha= .05.
ap < .05; ns= notsignificant

The replicationof age and educational older adults,youngeradults may findthe

differences in social well-beingstrongly sug- world coherent because they can view a
gests that social well-beingis shaped by worldthattendsto reflecttheirownpopular
prominentsocial forces.For the average culture.Insofaras age has a nonlinearrela-
adult, acquisition of education evidently tionshipto social well-being,the evidence
paves the way forsuccessfulresponsesto favors the "deceleration viewpoint":
social challengesof life.Withmore educa- Althoughsocial actualizationand contribu-
tion,each dimensionof social well-being
tionincreasewithage,the increasedeceler-
increases.Thus social weliness,like all other
ates witheach added year of life in both
aspects of health (N. Adler et al. 1994), is
Study1 and Study2. In otherwords,social
gradedbyprocessesofsocialstratification.
Each dimensionofsocialwell-beingalso well-beingis heterogeneous acrossages.This
is distributed by age. Four dimensions findingis consistentwithstudies showing
increasewithage. Social coherence,howev- the diversityof psychological well-being
er, decreases with age, perhaps in part acrossage groups.For example,olderadults
because the intelligibility of the world in reportmoreenvironmental masterybutless
Americansocietyis skewedtowardthecele- purpose in life than youngeradults (Ryff
brationof youthful culture.Comparedwith and Keyes 1995).

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SOCIAL WELL-BEING 133
DISCUSSION link,but do not overlap,withextantmea-
suresof mentalhealth,whichtendto reflect
Life is purportedly dividedinto private a psychologicalconception of well-being.
and public sides, each making its own Each scale of social well-being,even social
demands and producing its own conse- actualization,which emphasizes society's
quences.Althoughwe recognizethe public potentialand progress, correlatesonlymini-
demandson individuals, social psychologists
mallywithself-defined optimism.In short,
have not alwaystranslatedsocial challenges
thisstudysuggeststhatlifeincludesnumer-
intocriteriathatindividualsuse to evaluate
ous social challenges;thereforewell-being
theirwell-being. We have conceivedofsocial
includes social dimensionssuch as coher-
healthless as thepresenceof positivecrite-
ence,integration, actualization,contribution,
ria than as the absence of negativecondi-
and acceptance.
tions such as anomie and alienation.
Society,too, is a source of variationin
Moreover,the emphasison social structure
social wellness.Social structuralprofilescor-
mayskewattention towardthepublicside of
roboratethetheorythatsocialwellnessis an
lifeas a sourceof variationin social health
accomplishment. Social well-beingincreases
ratherthaninspiring us to explorethesocial
with education and, in general, with age.
natureof themeaningof health.Ultimately,
current research relies on and possibly Like othermeasuresof mentalhealth and
reflectsa bias towardpsychologicalconcep- well-being,it is graded by socioeconomic
tionsof well-being.Researcherseitherfocus standing.It remainsto be seen, however,
on clinicalsymptomatology such as depres- whetherthe same processes(e.g., self-con-
sion or use global measuresof lifesatisfac- ception) and dimensions (e.g., agency)
tionand happiness.Multidimensional mod- explainhow each aspect of social structure
els (Ryff1989) emergefromclassicperson- affectseach typeof mentalhealthand well-
alitytheoriesthatconceiveof theselfas pri- being.
marilyprivate. The relatively strong association of
The studiesreportedhere presentevi- prosocial communityinvolvement with
dence givingcredibilityto the theorythat aspectsofsocialwell-beingalso supportsthe
well-being includes social dimensions. I posited achievement of a well-lived life.
operationalizedand validated fivedimen- Individualsinvolvedin theircommunities
sionsof socialwell-beinggroundedin classic duringthe past 12 months,but not people
sociologicaltheoryand currentsocial psy- involvedmore than 12 monthspreviously,
chologicalperspectives. Data fromtwostud- reportedfeelingmore socially integrated
ies involving cross-sectionsof adultsand uti- and sociallycontributivethan people who
lizing differentmodes of administration had neverbeen involved.
(telephoneand self-administration) provide Despite the unityof measuresof social
strongevidenceforthe replicabilityof the wellness,it is also clear that social well-
latentstructureof the new scales and the being,like psychological well-being,is relat-
social structuralprofilesof dimensionsof ed to age in diverseways.Some aspects of
socialwell-being. social well-beingdecrease linearly;others
Confirmatoryfactor analysis showed increase linearly;stillothersincrease,but at
thatthe hypothesizedfive-factor model of a decelerating rate.The results therefore
social wellness fitsthe data best in both suggestthattheresources, skills,and experi-
studies. The findingssuggestthatputative ence gained through education and its
social challengesare manifoldand distinct. sequelae,as well as throughagingprocesses,
The new scales in Study1 correlateconver- are instrumentalto negotiatingthe chal-
gentlywithanomieand community involve- lengesofsociallife.
ment.In Study2, the new scales correlate Accordingto Nisbet (1953:15), and as
convergentlywithgenerativity, neighbor- quoted by Seeman (1959), the varioussyn-
hood health,and perceived constraints. In onymsforalienationtestifyto the impor-
bothstudies,thesescales correlatewithglob- tance of the constructof alienation. The
al indicatorsof lifesatisfaction, happiness, social scientificconceptionof a well-lived
and dysphoria.Thus the new scales clearly lifealso relieson the image of the isolated

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 134 SOCIAL PSYCHOLOGY QUARTERLY
Table 5. OLS Regressionsof Social Well-BeingScales on Age and Education

Study1 Study2
(N = 368) (N = 2,977)
Dane CountySample National (Unweighted)Sample

b f b f

Social Actualization
Age .17* .49 .06** .19
Age2 -.003** -.58 -.001** -.18
Up to 12 yearseducation
13 to 15 yearseducation 2.2** .18 .73** .08
16 or more yearseducation 3.1** .27 1.7** .19
Alpha 29.0 10.5
Social Acceptance
Age .21* .52 .04** .16
Age2 -.003* -.43 .00 .00
Up to 12 yearseducation
13 to 15 yearseducation 2.2* .16 .50** .07
16 or more yearseducation 2.5** .19 1.2** .15
Alpha 24.9 12.0
Social Integration
Age .07** .19 .06** .17
Up to 12 yearseducation
13 to 15 yearseducation 2.3** .17 .26 .03
16 or more yearseducation 3.2** .25 1.3** .14
Alpha 30.7 12.9
Social Contribution
Age .14* .49 .05* .17
Age2 -.003* -.49 -.002** -.26
Up to 12 yearseducation - - -
13 to 15 yearseducation 1.8** .16 1.1** .14
16 or moreyearseducation 3.4** .32 2.9** .35
Alpha 27.6 14.3
Social Coherence
Age -.03* -.13 -.03** -.09
Up to 12 yearseducation
13 to 15 yearseducation 1.3** .15 .90** .13
16 or more yearseducation 2.5** .30 1.9** .26
Alpha 24.7 8.0

Note:Age = Age - lowerage bound (age 18 in Study1; age 25 in Study2). All estimatesadjustedbygender,
race,and maritalstatus.
* p < .05; ** p < .01 (two-tailed).

individual.That is,theparagonsof thegood onlythe conceptualterritory of social well-

lifeare happyand satisfiedwiththeirprivate beingthatmeasuressociety'strajectory and
lives, and theypossess personal qualities potential(social actualization)and thechar-
indicativeof psychologicalwell-being.We acter of societythroughpeople's character
are leftto wonder,however,aboutthequali- (social acceptance).As Nisbetstates,lifeis a
tyof suchindividuals'livesas livedwithand quest to belong.The quest forthegood life,
forotherpeople and forsociety.By opera- however,has manymeanings, some ofwhich
tionalizingwell-beingthroughsatisfaction in are representedin the proposedfive-factor
domains of life such as familyand work, modelofsocialwell-being.
sociologistshave implicitlyquestionedthe In addition to exploringlife formore
conceptionof theindividualas a disconnect- social dimensions of well-being, future
ed social entity,but have not transformed researcherson positivementalhealth and
thisnotion.Even anomie,theclassicconcep- well-beingfacenumeroustasks.An immedi-
tion of the absence of social health,covers ate task is to investigatethe relationships

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SOCIAL WELL-BEING 135
betweenmeasuresofmentalhealthand well- 37:62-83.
beingthatpurportedly reflectdistinctcon- Carstensen,Laura L. 1995."Evidencefora Life-
structs.For example,are measuresofpsycho- Span Theoryof SocioemotionalSelectivity."
logical well-beingtrulydistinctfrommea- CurrentDirectionsin PsychologicalScience
4:151-56.
sures of social well-being?At stake is the Campbell,Angus.1981. The Sense of Well-Being
hypothesisthatthe challengesand labor of inAmerica:RecentPatterns and Trends.New
adulthoodoriginatein thefundamental divi- York:McGraw-Hill.
sionofpublicand privatelife. Campbell,Angus,PhilipA. Converse,and Willard
If life,in both its publicand its private L. Rodgers.1976. The Qualityof American
details,is the source of the dimensionsof Life.NewYork:Sage.
wellness,to whatextentare the social and Costa, Paul T. and Robert R. McCrea. 1980.
psychologicalchallengesbound by culture? "InfluenceofExtraversion and Neuroticism
on Subjective Well-Being: Happy and
What,in turn,are themediatorsand thevar- Unhappy People." Journalof Personality
ious structural sourcesof each typeof well- and Social Psychology38:668-78.
being? I repeatThoits's(1995) call forthe Davis,JamesA. and TomW.Smith.1994.General
use of multipleoutcomemeasuresin studies Social Surveys, 1972-1994: Cumulative
of healthand well-being. Progressdepends, Codebook. Chicago: National Opinion
however,on the developmentand use of ResearchCenter,University ofChicago.
modelsof well-beingthatmeasurethearray Diener, Ed. 1984. "Subjective Well-Being."
of humanfunctions, rangingnot onlyfrom PsychologicalBulletin95(3):542-75.
Durkheim,Emile. 1951.Suicide.New York:Free
negativeto positivebut also frompsycho-
Press.
logicalto social. Duval, Shelleyand RobertA. Wicklund.1972.A
Theoryof ObjectiveSelf-Awareness. New
REFERENCES York:AcademicPress.
Erikson,Erik H. 1950. Childhood and Society.
Adier, Alfred 1979. Superiority and Social NewYork:Norton.
NewYork:Norton.
Interest. Fenigstein,Allen,MichaelF Scheier,and Arnold
Adler, Nancy E., Thomas Boyce, MargaretA. H. Buss. 1975. "Public and Private Self-
Chesney,Sheldon Cohen,Susan Folkman, Consciousness:Assessmentand Theory."
RobertL. Kahn,and S. LeonardSyme.1994. Journal of Consulting and Clinical
"Socioeconomic Status and Health." Psychology 43:522-27.
AmericanPsychologist 49(1):15-24. Fey,W.F. 1955. "Acceptance of Others and Its
Andrews,FrankM. and StephenB. Withey. 1976. Relationto Acceptanceof Self and Other:
Social Indicatorsof Well-Being:Americans' A Reevaluation."JournalofAbnormaland
Perceptions of Life Quality. New York: Social Psychology 50(2):274-76.
Plenum. Gecas, Viktor.1989. "The Social Psychologyof
Antonovsky, Aaron. 1994. "The Sense of Self-Efficacy."AnnualReviewof Sociology
Coherence: An Historical and Future 15:291-316.
Perspective."Pp. 3-40in Senseof Coherence Greenwald, Anthony G. and Anthony R.
and Resiliency:Stress,Coping,and Health, Pratkanis.1984. "The Self." Pp. 129-78 in
edited by H.I. McCubbin,E.A. Thompson, HandbookofSocial Cognition Vol.3, edited
A.I. Thompson,and J.E.Fromer.Madison: by R.S. Wyerand T.K. Srull.Hillsdale,NJ:
UniversityofWisconsinPress. Erlbaum.
Bandura,Albert.1977."Self-Efficacy: Towarda Goffman, Erving.1959.ThePresentation ofSelfin
UnifyingTheory of Behavioral Change." EverydayLife.NewYork:Doubleday.
PsychologicalReview84(2):191-215. Groves,Robert.1989.SurveyErrorsand Survey
Biddle,Bruce J.1986."Recent Developmentsin Costs.NewYork:Wiley.
Role Theory."Annual Reviewof Sociology Gurin,Gerald, JosephVeroff,and Sheila Feld.,
12:67-92. 1960.AmericansViewTheirMentalHealth.
Bollen, KennethA. 1989. StructuralEquations NewYork:Basic Books.
withLatentVariables. NewYork:Wiley. Heidrich,Susan M. and Carol D. Ryff. 1996."The
Bradburn,Norman M. 1969. The Structureof Self in Later Years of Life."Pp. 73-102in
PsychologicalWell-Being. Chicago:Aldine. Aging in the Twenty-FirstCentury:A
Browne, Michael W. 1984. "Asymptotic Developmental Perspective,edited by L.
DistributionFree Methods in Analysisof Sperryand H. Prosen.NewYork:Garland.
Covariance Structures." BritishJournalof Heiss,Jerold.1981."Social Roles." Pp. 94-129in
Mathematicaland StatisticalPsychology Social Psychology:
SociologicalPerspectives,

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 136 SOCIAL PSYCHOLOGY QUARTERLY
edited by M. Rosenbergand R.H. Turner. Control." Social Psychology Quarterly
NewYork:Basic. 58:31-43.
Horney, Karen. 1945. Our Inner Conflicts:A Mirowsky,Johnand Catherine E. Ross. 1989.
Theoryof Neurosis.New York:
Constructive Social Causes of Psychological Distress.
Norton. NewYork:Aldine.
Israel,Joachim.1971.Alienation:FromMarx to Nisbet,Robert.1953. The Questfor Community.
Modern Sociology. Boston: Allyn and NewYork:OxfordUniversity Press.
Bacon. O'Rourke, Diane and Johnny Blair. 1983.
James,William.1890. Principlesof Psychology. "Improving RandomSelectioninTelephone
NewYork:Holt. Surveys."Journalof MarketingResearch
Joreskog, Karl G. and Dag Sorbom. 1993a. 20:428-32.
PRELIS 2: A ProgramforMultivariateData Riley,Matilda,R.L. Kahn and A. Foner.1994.Age
Screening and Data Summarization. and StructuralLag: Society's Failure to
IN: Scientific
Mooresville, Software,
Inc. ProvideMeaningfulOpportunities in Work,
. 1993b. LISREL 8: A Guide to the Program Family, and Leisure.NewYork:Wiley.
andApplications. Chicago:SPSS, Inc. Ryff,Carol D. 1989."HappinessIs Everything, or
Karabel, Jeromeand A.H. Halsey. 1977. Power Is It? Explorations on the Meaning of
and Ideology in Education. New York: Psychological Well-Being." Journal of
OxfordUniversity Press. Personality and Social Psychology
Kohn, Melvin. 1969. Class and Conformity. 57(6):1069-81.
Homewood,IL: Dorsey. Ryff,Carol D. and Corey Lee M. Keyep.1995.
Kohn, Melvin and Carmi Schooler. 1982. "Job "The Structure of PsychologicalWell-Being
Conditionsand Personality: A Longitudinal Revisited." Journalof Personality and Social
Assessmentof Their Reciprocal Effects." Psychology 69(4):719-27.
AmericanJournalofSociology87:1257-86. Seeman, Melvin. 1959. "On the Meaning of
Lachman, Margie E. and Suzanne L. Weaver. Alienation."AmericanSociologicalReview
Forthcoming. "The Sense of Control as a 24:783-91.
Moderatorof Social Class Differencesin .1983. "Alienation Motifs in Contemporary
Health and Well-Being." Journal of Theorizing:The Hidden Continuityof the
Personality and Social Psychology 74(3): 76- Classic Themes." Social Psychology
73. Quarterly 46:171-84.
Larson, James S. 1992. "The Measurement of __. 1991. "Alienation and Anomie." Pp. 291-
Social Well-Being." Social Indicators 374 in Measuresof Personalityand Social
Research28:285-96. Psychological Attitudes,edited by J.P.
. 1996. "The World Health Organization's Robinson, P.R. Shaver, and L.S.,
Definitionof Health:Social versusSpiritual Wrightsman. San Diego:AcademicPress.
Health."Social IndicatorsResearch38:181- Sewell, William and Robert Hauser. 1975.
92. Education, Occupation, and Earnings:
Lefcourt,Herbert M. 1982. Locus of Control: Achievement in theEarlyCareerNew York:
Current Trendsin Theoryand Research.2nd AcademicPress.
ed. Hillsdale,NJ:Erlbaum. Shaw,MarvinE. and Philip R. Costanzo. 1982.
Luhtanen,Riia and Jennifer Crocker.1992. "A Theoriesof Social Psychology. 2nd ed. New
Collective Self-Esteem Scale: Self- York:McGraw-Hill.
Evaluation of One's Social Identity." Srole,Leo. 1956."Social Integration and Certain
Personality and Social PsychologyBulletin Corollaries: An Exploratory Study."
18(3):302-18. AmericanSociologicalReview21:709-16.
Maslow,AbrahamH. 1968.Towarda Psychology Thoits,PeggyA. 1992. "IdentityStructuresand
ofBeing.2nded. NewYork:Van Nostrand. Psychological Well-Being: Gender and
McAdams,Dan P. and Ed de St.Aubin.1992."A Marital Comparisons."Social Psychology
Theoryof Generativity and Its Assessment Quarterly 55(3):236-56.
throughSelf-Report, BehavioralActs,and __. 1995. "Stress, Coping, and Social Support
Narrative Themes in Autobiography." Processes: Where Are We? What Next?"
Journal of Personality and Social Journal of Health and Social Behavior
Psychology 62(6):1003-15. (extraissue):53-79.
Mead, George H. 1934. Mind,Self,and Society: Trafimow, David,HarryC. Triandis, and SharonG.
FromtheStandpoint of a Social Behaviorist. Goto. 1991."Some Testsof the Distinction
Chicago:University ofChicagoPress. betweenthePrivateSelfand the Collective
Merton,Robert.1949.Social Theoryand Social Self." Journalof Personalityand Social
Structure. Glencoe,IL: FreePress. Psychology 60:649-55.
Mirowsky,John.1995. "Age and the Sense of Triandis,Harry C. 1989. "The Self and Social

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SOCIAL WELL-BEING 137
Behavior in DifferingCulturalContexts." "Sampling Weights and Regression
PsychologicalReview96(3):506-20. Analysis." Sociological Methods and
Waksberg,Joseph.1978."SamplingMethodsfor Research23(2):230-57.
Random Digit Dialing." Journal of the Wrightsman,Lawrence S. 1991. "Interpersonal
AmericanStatistical
Association73:40-46
Trust and Attitudes toward Human
Waterman,Alan S. 1993. "Two Conceptionsof
Nature." Pp. 373-412 in Measures of
Happiness: Contrasts of Personal
Expressiveness(Eudaimonia) and Hedonic Personality and Social Psychological
Enjoyment."Journalof Personalityand Attitudes,edited by J.P.Robinson, P.R.
Social Psychology
64:678-91. Shaver,and L.S., Wrightsman.San Diego:
Winship,Christopherand Larry Radbil. 1994. AcademicPress.

CoreyLee M. Keyesis an AssistantProfessorin theDepartment of Sociologyand theRollins

Schoolof PublicHealth(Department ofBehavioralSciencesand HealthEducation).He is an
associate of theMacArthurFoundationMIDMAC researchnetwork.He also is a recent
inducteeinto a researchnetworkon "PositivePsychology"organized and led byAPA
PresidentMartinSeligmanand MihalyiCsikszentmihalyi to build and promotea fieldof
inquiryintoPositiveHealth.His researchexploresthesocial psychologyof
social-scientific
healthand well-beingand successful
aging.

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
138 SOCIALPSYCHOLOGYQUARTERLY
A. Indicators
Appendix Constructs
ofSocialWell-Being

Social Integration
Study1
2. Y6u don't feelyou belong to anythingyou'd call a community(-).
14.You feel like you're an importantpartof yourcommunity(+).
17. If you had somethingto say,you believe people in yourcommunitywould listento you (+).
27. You feelclose to otherpeople in yourcommunity(+).
30. You see yourcommunityas a source of comfort(+).
44. If you had somethingto say,you don't thinkyourcommunitywould take you
seriously(-).
50. You believe otherpeople in societyvalue you as a person(+).
Study2
2. I don't feel I belong to anythingI'd call a community(-).
6. I feelclose to otherpeople in mycommunity(+).
11. My communityis a source of comfort(+).
Social Acceptance
Study1
7. You thinkthatotherpeople are unreliable(-).
9. You believe thatpeople are kind(+).
18.You believe thatpeople are self-centered(-).
26. You feel thatpeople are not trustworthy
(-).
33. You thinkthatpeople live onlyforthemselves(-).
41. You believe thatpeople are moreand moredishonestthese days (-).
46. You thinkthatpeople care about otherpeople's problems(+).
Study2:
3. People who do a favorexpectnothingin return(+).
10. People do notcare about otherpeople's problems(-).
14. I believe thatpeople are kind(+).
Social Contribution
Study1
4. Your behaviorhas some impacton otherpeople in yourcommunity(+).
5. You thinkyou have somethingvaluable to give to theworld(+).
37. Your dailyactivitiesdo notproduceanythingworthwhileforyourcommunity(-).
40. You don't have the timeor energyto give anythingto yourcommunity(-).
42. You thinkthatyourworkprovidesan importantproductforsociety(+).
49. You feelyou have nothingimportantto contributeto society(-).
Study2
4. I have somethingvaluable to give to theworld(+).
7. My dailyactivitiesdo not produceanythingworthwhileformycommunity(-).
15. I have nothingimportantto contributeto society(-).
Social Actualization
Study1
6. You believe thatsocietyhas stoppedmakingprogress(-).
8. Societyisn'timprovingforpeople like you (-).
19.You don't thinksocial institutionslike law and governmentmake yourlifebetter(-).
21. You see societyas continuallyevolving(+).
25. You thinkour societyis a productiveplace forpeople to live in (+).
38. For you there'sno such thingas social progress(-).
45. You thinkthe worldis becominga betterplace foreveryone(+).
Study2
5. The worldis becominga betterplace foreveryone(+).
9. Societyhas stopped makingprogress(-).
13. Societyisn'timprovingforpeople like me (-).

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 SOCIAL WELL-BEING 139
Social Coherence
Study1
3.Theworldis toocomplexforyou(-).
10.Scientistsaretheonlypeoplewhocanunderstand howtheworldworks(-).
13.You cannotmakesenseofwhat'sgoingon intheworld(-).
15.Mostcultures areso strangethatyoucannotunderstandthem(-).
20.You thinkit'sworthwhile theworldyoulivein(+).
to understand
29.You findithardtopredict whatwillhappennextinsociety(-).
Study2
1.Theworldis toocomplexforme(-).
8. I cannotmakesenseofwhat'sgoingon intheworld(-).
12.I finditeasytopredictwhatwillhappennextinsociety(+).

Notes:A negativesigninparentheses indicatesthattheitemis reversescored,whereresponse

optionsrangefromstrongly,
moderately, orslightlydisagreetoslightly, orstrongly
moderately,
agree.(2) Theitemnumbers correspond totheitemnumbers andtheirmeasurementqualities
reported inTable2 (Appendix).In study2,item12 thatindicates SocialCoherencewasomitted
fromStudy1 becauseitsuppressed theinternal ofthescale.
reliability

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions
 140 SOCIAL PSYCHOLOGY QUARTERLY
AppendixB. Estimatesof ParametersofTheoreticalSocial Well-BeingMeasurementModel

Dane CountySample National (Unweighted)Sample

Metric Validity Metric Validity

Construct Item Loading Coefficient Loading Coefficient

Social 6 fl.0 .56 1.8 .46

Actualization 8 1.0 .62 1.4 .78
45 .54 .48 fl.0 .62
19 .74 .46
21 .75 .47
25 .89 .57
38 1.0 .62

Social 7 fl.0 .56

Acceptance 9 .99 .58 1.8 .41
46 .84 .60 4.3 .70
18 .62 .51
26 1.1 .68
33 1.1 .60
41 .80 .58
43a fl.0 .18

Social 2 fl.0 .46 fl.0 .59

Integration 27 1.2 .66 1.2 .78
30 1.1 .57 1.1 .74
14 1.3 .65
17 1.2 .66
44 1.2 .64
50 1.1 .65

Social 5 fl.0 .59 fl.0 .58

Contribution 37 1.2 .65 1.3 .57
49 1.2 .74 1.5 .80
40 1.1 .64
42 1.2 .60
4 .86 .49

Social 3 fl.0 .43 fl.0 .61

Coherence 13 1.2 .47 1.3 .77
15 1.3 .55
10 .90 .44
20 .92 .46

Notes:fl.0 = indicatoris used to fixthe metricof the latentconstruct.

The itemnumberrefersto theorderof
the itemin the 50-itempool in Study1; the itemnumberalso referencesthe itemin AppendixA.
aThe indicatorused to set themetricof Social Acceptancein Study2 was not retainedin Study1, butwas used
in Study2 in an attemptto achieve a morebalanced scale.

This content downloaded from 192.236.36.29 on Wed, 25 Sep 2013 08:31:05 AM

All use subject to JSTOR Terms and Conditions