Divorce and Psychological Stress

Author(s): Alan Booth and Paul Amato

Source: Journal of Health and Social Behavior, Vol. 32, No. 4 (Dec., 1991), pp. 396-407
Published by: American Sociological Association
Stable URL: http://www.jstor.org/stable/2137106 .
Accessed: 29/11/2014 16:53

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
Journal of Health and Social Behavior.

http://www.jstor.org

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
 Stress*
Divorceand Psychological

ALAN BOOTH
StateUniversity
ThePennsylvania
PAUL AMATO
ofNebraska-Lincoln
University

of Healthand Social Behavior1991, Vol. 32 (December):396-407

Journal

Whileresearchon adjustment to divorcehas beenextensive,thepaucityofstudies

assessing stress before and after divorce has kept the relation between
psychologicalstressand maritaldissolutionunclear. Conflicting findingscast
some doubton the utilityof usinga crisismodelto studydivorce.Analysisof
three-wavepanel datafroma nationalsampleofpersonsmarriedin 1980 indicates
thatthe crisis model is appropriatefor understanding adjustmentto divorce.
Comparisonsofdivorcedpersonswithmarriedpersonsshowa predivorcerise in
stresswhichthenreturnsto levels comparableto those reportedby married
individuals.No evidencewas found supportingthe idea thata high level of
psychologicalstressis a generalcause of divorceor thatdissolutionresultedin
moreor less permanentelevationof psychologicalstress.Findingssupportthe
hypothesisthatpredivdrceresourcesand outlooksinfluence theamountof stress
experiencedin thetwoyearsimmediately followingdivorce.Belowmedianfamily
incomes,no post-highschool experience,and wifenot in the laborforce put
divorcingindividualsat a disadvantage.Individualsreporting few premarital
troublesand beliefsin the immorality of divorcealso appear to experience
heightenedstressin thetwoyearsfollowingdivorce.

Adjustment to divorcehas beenthesubject The concept of adjustment,however,

of a largenumberof studies(see Kitsonet al. impliesa modelof divorceas a crisiswhere
1989; Kitsonand Morgan1990 forreviews). psychologicalstressis low beforetheevent,
Witha fewnotableexceptions(Doherty,Su, increasesas the dissolutionapproaches,and
and Needle 1989; Menaghan1985; Wertlieb then decreases followingdivorce. As is
et al. 1984; Wheaton1990), theseinvestiga- shown below, thereis some evidence that
tionsbeginwhenindividualsfilea petitionor extendedandinfinite periodsof stressprecede
aregranted a decree.As such,assessments of and followdivorce.This researchcastsdoubt
howan individual'squalitiesinfluence adjust- on the use of a crisis model. Therefore,
mentto divorceare limited.Some studies beforeexaminingfactorsthatinfluencead-
haveused retrospective information on predi- justment,we mustfirstaddressthequestion
vorce conditions,but recall of predivorce
of whether divorcefitsthecrisismodel.
circumstances maybe influenced by interven-
The studyreportedhere allows us (1) to
ing events(e.g. Bower 1981) or thepassage
of time. examine how well the crisis model fits
divorce,and (2) ifitdoes fit,to overcomethe
methodologicaldifficulties of prior studies
thattryto explainadjustment to divorce.We
* This studywas supported in partby National drawon a nationalsampleof married persons
Instituteon Aginggrantno. 5 ROI AG4146. The firstinterviewed in 1980 and again in 1983
contributionsofco-investigatorDavid Johnsonand
of three anonymousreviewersare gratefully
and 1988. Over theeight-year periodof the
acknowledged. Correspondence shouldbeaddressed study,138 of the 1,339 interviewed in all
to Alan Booth,Department of Sociology,Pennsyl- threewaves divorcedor separatedperma-
vaniaStateUniversity, UniversityPark,PA 16802. nently.Information on pre-and post-divorce
396

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
DIVORCE AND PSYCHOLOGICAL STRESS 397
circumstances andpsychological stresspermit thatthepsychological deficitsassociatedwith
us to considertheappropriatenessofthecrisis divorce are not temporaryresponsesto a
model in addition to the influence of crisis,butpersistindefinitely. Some studies'
predivorce factorson thelevelanddurationof resultsare consistentwith this notion.No
psychological stress. relationhas been foundbetweentimesince
divorce and lower stress(e.g. Amato and
Partridge 1987; McLanahan1983).
Adequacyof theCrisisModel Accordingto thethirdexplanation, psycho-
logical problemscan lead to maritaldissolu-
Previousresearchhas clearlydemonstratedtion;indeed,manypeoplecitetheirpartner's
thatdivorceand separatedindividualshave personality problems,drinking, druguse, and
higherlevels of psychologicalstressthando violenceas factorsin theirdecisionto seek a
marriedindividuals.Statisticsrevealthatthe divorce (Bloom, Niles, and Tatcher 1985;
divorcedand separatedhave an elevatedrisk Kitsonand Sussman 1982). Consistentwith
of psychiatric illness,suicide,motorvehicle thisnotion,one prospectivestudy(Doherty,
accidents, alcoholism, homicide, physical Su, and Needle 1989) foundthat married
illness,and overallmortality (Bloom, Asher, individualswho subsequentlydivorcedre-
and White 1978; Gove, Style, and Hughes portedmorepre-divorce symptoms of depres-
1990). Surveysalso revealthatdivorcedand sion thandid those who remainedmarried.
separatedindividuals, comparedwithmarried Similarly,Wertliebet al. (1984) foundthat
individuals,reporthigherlevels of depres- the utilizationof medicaland mentalhealth
sion, anxiety,and generalunhappiness(e.g. facilitieswas higheramongindividualswho
Gove and Shin 1989). later divorcedthan among those who re-
Althoughthe link betweendivorce and mainedmarried.
psychological impairment is well-established, It is quite possible, however, that the
thereare threepossible explanations,only psychological problemsrevealedin theselast
one of which embracesthe crisis model. two studiesresultedfromthemaritalconflict
Whilethethreeexplanations are notmutually and estrangement thatprecededdivorce.The
exclusive, distinguishingamong them is pre-separation periodis oftendescribedas the
usefulforanalyticpurposes.The crisismodel mostdistressing partof the divorceprocess
explanation focuseson thenotionthatdivorce (Kitsonand Morgan1990). Moreover,there
is a stressfullife event (e.g. Tschann, is evidencethatis contrary to theexplanation
Johnston, and Wallerstein 1989); indeed,it is thatpsychologicalproblemscause divorce.
consideredto be one of themoststressful of One prospective studyfoundthatpsychologi-
all life events (Holmes and Rahe 1967). cal distressfollowed,but did not precede,
Implicitin the notion of a crisis is the maritaldissolution (Menaghan1985). And,in
assumptionthatafteran elevationin stress, theDohertyet al. (1989) study,an increasein
individualsreturnto predivorcelevels. For substanceabuse amongfemaleswas apparent
example,Hetherington, Cox, and Cox (1982) after,butnotbefore,divorce.
foundthatmostdivorcedmothersimproved Each of the three explanationsmakes
in functioning afterone or two years. This different predictionsabout the variationbe-
perspectiveassumesthatthe gap in psycho- tweenpsychologicalstressand divorceover
logical stressbetweenthe divorcedand the time. The first perspectiveassumes that
marriedcan largelybe attributed to individu- psychological stressincreasesaroundthetime
als whohavebeendivorcedfora shortperiod of divorce and then declines following
of time;individualswho have been divorced divorce.The secondperspective assumesthat
fora longerperiodof timeare assumedto psychological stressincreasesaroundthetime
showlevelsof stresscomparableto thosewho of divorceand staysat a relatively highlevel
are married. indefinitely.The thirdperspectiveassumes
The second explanationfocuses on the thatthe psychologicalstressis high for a
notionthat divorce involves persistentlife considerableperiod of time beforedivorce
strainsthatresultin psychological stress.The and may remainat a high level following
strainsinclude economic hardship,social divorce,at leastforthepartner afflicted
with
isolation,and,forsingleparents,sole parent- poor mentalhealth.In thepresentstudy,we
ingresponsibilities (McLanahan1983;Pearlin examinehow thepsychologicalstressof the
andJohnson1977). This perspective assumes divorcedand permanently separatedvaries

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
398 JOURNALOF HEALTH AND SOCIAL BEHAVIOR
over timeto see whichperspectiveis most 1987; Colletta 1983; Kitson et al. 1989;
comparablewiththedata. Of course,thedata Kitson and Morgan 1990; Tschann et al.
maybe consistent withmorethanone model. 1989; Weiss 1979).' Divorce is knownto
In orderto testthethreemodelsadequately, decreasethe standardof livingof bothmen
remarriageneeds to be taken into account and women(Duncanand Hoffman1985), but
because it can influencethe adjustment individualswho have a highlevel of income
process. Marriedindividualsenjoy greater priorto divorceare less likelyto experience
well-beingthan non-marrieds (Gove et al. extremeeconomic hardshipthan are those
1990) and our own analysis (not shown) who enterdivorcewithonly a moderateor
indicatesthatremarrieds do not differfrom low levelof income.Educationmayfacilitate
marriedson overallhappinessand a number adjustment partlythroughincreasinga per-
of otherindicatorsof well-being.Given the son's income-earning potential;in addition,
linkbetweenwell-beingandmarriage, remar- well-educatedindividuals,compared with
riage may be important to improvements in poorlyeducatedindividuals,mayhave better
psychologicalfunctioning followingdivorce. cognitivecopingresponses,a stronger sense
If declinesin stressare apparentonlyamong of control,andbettersocial supportnetworks
those who remarryand are not observed forcopingwithdivorce(Kitsonand Morgan
amongthosewhoremainsingle,thestatusof 1990). In relationto employment, women
beingdivorcedmaybe a chronicstressor that whoarein thepaid laborforcebeforedivorce
can onlybe remediedthrough remarriage. On do not have to seek employmentunder
the otherhand, if individualswho remain duress,are in a betterlabormarketposition
singleexperiencelowerstressovertime,this thanare womenwho have been out of the
supportsa view of divorceas a temporarypaid laborforcefora periodoftime,andmay
crisis.We exploretherole of remarriage in have friendsin the workplace who can
post-divorce adjustment. providesupport.Husbandsbenefitbecause
employedwives are less likelyto draw on
theirspouse's incomeforsupport.And with
Pre-divorce FactorsthatAffect Adjustment regardto social support,the availabilityof
networks
supportive is widelythought to be a
Assumingfor the momentthat a crisis major factorin buffering individualsfrom
model is appropriate, people may differin stressfulevents (House, Umberson, and
theirabilityto returnto normallevels of Landis 1988). Finally, youth may be a
functioning followingdivorce.In considering resource,butpreviousstudiesyieldcontradic-
howpre-divorce factorsmightaffectspeedof toryresultsin relationto whetheryounger
adjustment,we turn to familystress and individualsor olderindividualsadjustbetter
copingtheory, firstformulatedby Hill (1949) to divorce(Kitsonand Morgan1990).
and laterextendedby McCubbinand Patter- Althoughpreviousstudieshave supported
son (1982). Thistheorystatesthatadjustment the role playedby theseresourcefactorsin
to a familycrisisdependson threefactors:(1) adjustment to divorce,mostof thisresearch
thenumberof stressorsthatpile up, (2) the has beencross-sectional. In thepresentstudy,
resourcesavailableto deal withtheproblem, panel data is used to testthehypothesis that
and (3) the mannerin which individuals psychologicalstress followingdivorce de-
perceiveand appraisethesituation.Whereas creasesmorequicklyforthosewho, priorto
the originalformulation of thistheory(Hill divorce,possess thefollowingresources:(1)
1949) focused on pre-crisisfactorsin ac- a highlevel of income,(2) a highlevel of
countingfor the severityof the crisis, the education,(3) employment ofwifein thepaid
moreelaborateversionofthetheory (McCub- laborforce,(4) a highlevelof social support,
bin and Patterson1982) takes into account and (5) a youngage.
post-crisisfactors.Because of thepaucityof The thirdfactorin the familystressand
researchin thisarea, we focusprimarily on copingmodelis individuals'appraisalof the
theroleof pre-divorce factors. stress-inducingevent.In thecase of divorce,
In relationto thesecondcomponent of the some people maybe relatively satisfiedwith
model,resourceslikelyto facilitatedivorce theirmarriage,and divorceis likelyto be an
adjustment includea highlevelof incomeand unwelcomeor unanticipated change.But for
education,wife's employment, and strong people who perceive a large numberof
socialsupportnetworks (AmatoandPartridge problemsin theirrelationship, divorcemaybe

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
DIVORCE AND PSYCHOLOGICAL STRESS 399
a relief.Supportforthisnotionwas obtained or improvesmorequicklyin someindividuals
in a longitudinal studyby Wheaton(1990), thanin others.
who found that psychologicaldistressin-
creasedafterdivorceamongpeople who saw
theirmarriageas havingfew problems,but RESEARCH PROCEDURES
decreasedafterdivorceamongpeople who
saw theirmarriageas havingmanyproblems. TheSample
Similarly,Kitsonand Sussman(1982) found
that expressingrelief about the end of a Data for this study comes from an
marriagewas associatedwithfewersignsof eight-yearlongitudinalstudyof a national
mentalhealthproblems.The above findings sample of 2,033 marriedpersons(not cou-
suggestthatindividualswho perceivetheir ples) 55 years of age and underwho were
marriageas troubledadjust betterthan do interviewed by telephonein 1980, 1983, and
otherindividuals. again in 1988. Sample households were
People also differin theirviews of the chosen througha clusteredrandom-digit-
morality of divorce.Thosewhobelievethatit dialingprocedure,and the husbandor wife
is wrong to divorce and then experience was selectedfor interviewusing a second
dissolutionexperiencemore distress than randomprocedure.The completionratewas
thosewho do notholdthisbelief. 65 percent.The refusalratewas 18 percent;
The factthatourdatasetcontainsmeasures theother17 percentincludedthoseindividu-
ofthesevariablespriorto divorceprovidesan als whocouldnotbe reachedaftertenormore
opportunity to examinethesenotionswithout callbacks. These rates compare favorably
having to rely on retrospective accounts. with other studies using this technique
Specifically,we test the hypothesesthat (Grovesand Kahn 1979).
psychologicalstress followingdivorce de- In 1983, 1,592 of the respondents were
creasesmorequicklyforthosewho, priorto successfullyinterviewedfor a re-interview
divorce(1) perceivemanyproblemsin their rate of 78 percent.This is comparableto
marriage,and (2) hold a positive attitude panelstudieswhichuse face-to-face interview
towarddivorce. procedures(Boothand Johnson1985). In the
thirdwave, 66 percentof theoriginalsample
werere-interviewed. The originalsamplewas
comparedwith U.S. Census data and was
AnalysisStrategy foundto be representative
withrespectto age,
race, householdsize, presenceof children,
Our analysis takes place in two steps. homeownership, andregion.The secondand
Giventhelackoflongitudinal researchon this thirdwaves are slightly
unrepresentative with
topic,ourfirstaimis to utilizethethree-wave respectto Blacks,Hispanics,younger respon-
panel data to describethe relationship be- dents,renters,and those with less than a
tweenpsychologicalstressand divorceover college education but no more so than
time.This will inform us of theviabilityof a investigationsusing personal interviews
crisismodelof divorce.To accomplishthis, (Boothand Johnson1985).
thelevel of psychologicalstressis examined
at threepointsin time for threegroupsof
people:thosewhodivorcedbetween1980 and Measures
1983, thosewho divorcedbetween1983 and
1988, and those who remainedin intact Divorce. Amongthose interviewed in all
marriages.Our second aim is to test the three waves, 141 divorcedor obtaineda
hypothesesdescribedabove relatingpredi- permanent separationduringthecourseof the
vorcevariablesto thelengthof timerequired study. Three people who divorced twice
for adjustment.To do this, psychological duringthestudyperiodwereeliminated from
stress is examined over time following study,bringingthe totaldivorcedor perma-
divorceto obtaina betterunderstanding ofthe nentlyseparatedto 138. Between 1980 and
duration overwhichpeopleexperiencestress. 1983, 65 dissolvedtheirmarriage;between
Interactiontermsare thenexaminedbetween 1983 and 1988, 73 did so.
predivorce variablesand lengthof timesince PsychologicalStress. Three measuresof
divorceto see ifpsychological stressis lower psychologicalstresswere used in the study.

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
400 JOURNALOF HEALTH AND SOCIAL BEHAVIOR
The firstwas a measureof happinesswidely or uses drugs,has been in troublewiththe
used in social indicatorsurveys:"Takingall law?" A summatedscale was created in
thingstogether, how wouldyou say you are whichthe presenceof the problem(due to
thesedays?" Those reporting less happiness eitherthespouse's or respondent's behavior)
wereassignedhigherscores.The itemhas a was coded 1 whilethelack of a problemwas
meanof 1.61 and a standard deviationof .58. coded 0. Scores rangedfrom0 to 13. The
The seconditemtaps psychologicaldistress: scale has a meanof 2.7, a standarddeviation
"In thelastthreeyears,werethereevertimes of 2.5, and an alphareliability of .76
whenyou wereextremely unhappy,nervous, Attitudesabout the moralityof divorce
irritable,or depressed?" An affirmativewere measured using a Likert-typescale
response was followed by the question: based on theitems:"Now I'm goingto make
"Have you felt this way recently?"Those some statements aboutmarriageand I would
reportingrecentdistresswerecoded 1, others like you to tell me whetheryou strongly
0. The itemhas a meanof 17 percentand a agree, agree, disagree,or strongly disagree
standarddeviationof 38 percent.The third witheach statement: Couplesareable to geta
measureis a set of threeitems fromthe divorcetoo easilytoday.It's okayforpeople
Langner(1962) scale thatwereaskedonlyin to getmarried thinkingthatifitdoes notwork
1983 and 1988 and is therefore notused in all out, they can always get a divorce. The
analyses.The itemsare: "How oftenin the personalhappinessof an individualis more
last yearwereyou botheredby acid or sour importantthan putting up with a bad
stomach?Would you say often,sometimes, marriage.Marriageis for life, even if the
or never?How oftenin thelastyearwereyou couple is unhappy."The firstand last items
troubled withheadachesor painsin thehead? werereversecoded to formthescale. Scores
How often have you been botheredby rangedfrom6 to 30. The scale has a meanof
breakingout in cold sweats?"Scores ranged 13.1, a standarddeviationof 2.2, and an
from3 to 9 and higherscoresindicategreater alphareliability coefficient
of .51.
stress.The stresssymptoms scale has a mean
of 4.6 and a standarddeviationof 1.1
The threemeasuresare quite independent RESULTS
of one anotheras indicatedby low inter-item
correlations;they range from .12 to .32. We begin with an examinationof the
Therefore,theywere not combinedinto a divorce-as-crisis model. Seven factorsare
singlemeasureof stress. then examined that may influenceinitial
Pre-divorceModerators of Stress. We stressand enhanceor retardthe speed of
examinetwo typesof factorsthathave the adjustment to divorce.
potentialforinfluencing initialstressand the
speedof adjustment to divorce:resourcesand
outlook. Resourcesinclude familyincome, Changein PsychologicalStressBeforeand
the formaleducationalachievementof the AfterDivorce
respondent,whetheror not the wife is
employedforpay outsidethehome,number We examinethelevelsof unhappiness and
of close friendsreportedby the respondent, distressin 1980, 1983, and 1988 forpeople
and the age of the respondent.Outlookon whose marriageremained intact for the
divorceis reflected in thenumber ofproblems eight-yearperiod (n = 1,201), people who
reported forthemarriageand attitudes about divorcedbetween1980 and 1983 (n = 65),
themorality of divorce. and peoplewho weredivorcedbetween1983
Information on problemsthatcharacterize and 1988 (n = 73). Only people withstress
themarriage was obtainedfroma setofitems measuresin all threewaves are includedin
thatasked about 13 potentialtroublespots: theanalysis.Multipleclassification analysis,
"Have you had a problemin yourmarriage a formof multipleregressionanalysiswhich
because one of you gets angryeasily, has makesno assumptions aboutthedistributional
feelingsthatare easily hurt,is jealous, is propertiesof the independent variable,was
domineering, is critical,is moody,won'ttalk used to calculatethemeanswhilecontrolling
to the other,has a sexual relationship with for possible confoundingvariables. The
someoneelse, has irritating habits,is not at independentvariable was a three-category
homeenough,spendsmoneyfoolishly, drinks variable reflectingmaritalstatus over the

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
DIVORCE AND PSYCHOLOGICAL STRESS 401
eight-yearperiod (neverdivorced,divorced groupsof divorcedindividualswereexhibit-
between 1980 and 1983, and divorced inggreaterstressthanthosewhosemarriages
between1983 and 1988). Because females remainedintact(p<.01 forthetwomeasures
are more likelythanmales to reportstress in both categoriesof divorce). In 1988,
symptoms(Dohrenwendand Dohrenwend individualswho divorcedbetween1980 and
1976), and to experiencegreatereconomic 1983 did notshow anymoreunhappiness or
deprivation as a resultof divorce(Weitzman distressthanpeoplewhodidnotdissolvetheir
1985), genderwas controlled.Maritaldura- marriage.On the other hand, those who
tionwas controlled becauseithas beenshown divorcedbetween 1983 and 1988 are still
to influencetheprobability of divorceandthe exhibiting elevatedstress.Whilethedistress
factorsthat precipitatedivorce (White and score is not statistically significantbetween
Booth 1991). Level of educationwas also the divorced and the continuouslyintact
controlledas an indicatorof socioeconomic group, the unhappinessscore is (p<.01).
status,anotherfactorwhichhas been shown Although notshownin thefigures, thepattern
to be relatedto the probabilityof divorce forfemalesand malesis thesame.
(Boothet al. 1985) andto stress(Pearlinet al. The data show that stress appears to
1981). The adjustedmeanlevelsof unhappi- increase prior to divorce and to decrease
ness are shown in Figure 1 and percent following divorce. The divorce-as-crisis
reportingpsychological distressare presented model is supportedby this analysis. Of
in Figure2. course, this analysisdoes not rule out the
When comparedto the individualswhose possibilitythatother,more severeformsof
marriagesremainintact,thosewho divorced psychologicaldistresssuch as acute depres-
between1980 and 1983 were experiencing sion or anxietycontribute to maritalinstabil-
greaterstressin 1980,butthosewhodivorced ityin somecases.
between1983 and 1988 were no different
thanpeople who did not divorceduringthe
courseof thestudy.The differences between Pre-divorceFactorsthatModerate
those who did not divorce and those who Post-divorce Stress
dissolvedtheirmarriagesbetween1980 and
1983 are statisticallysignificant(p<.01 for Beforeexploringtheinfluence of potential
unhappinessand distress). In 1983 both moderatingfactors,we firstexaminedthe

FIGURE 1. UnhappinessBeforeand AfterDivorce*

Unhappiness

1.8-

1.6

1.4
1980 1983 1988
Year of Interview
Marriage Intact | Divorced 80-83 * Divorced 83-88
* AdjustedforGender,Durationand Education.

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
402 JOURNALOF HEALTH AND SOCIAL BEHAVIOR
FIGURE 2. Psychological
DistressBeforeand AfterDivorce*
Percent WithPsychologicalDistress
43

23 -

13
1980 1983 1988
Year of Interview
Marriage Intact ? Divorced 80-83 " Divorced 83-88
* AdjustedforGender,Durationand Education.

patternof post-divorce stress.We createda TABLE 1. Mean PsychologicalStress by

measurethatreflectsthe numberof months MonthsSinceDivorceAdjustedfor
that-have elapsed sincedivorce.Afterexam- Pre-divorce Stress
iningelapsedmonth-by-elapsed monthvaria- Std
tion in the dependentvariables,collapsing Mean Dev N
monthssincedivorceintothreecategories(24 Unhappiness
monthsor less, 25 to 48 months,and 49-plus 24 monthsor less 1.9** .68 31
months)yieldsenoughcases in each category 25-48 months 1.7 .56 39
and at the same timereflectsthechangesin 49 monthsor more 1.5 .59 68
psychologicalstress.Multipleclassification Psychological Distress
24 monthsor less 33%* 47% 31
analysis is used to examine the relation 25-48 months 15% 33% 39
betweentimesincedivorceandpsychological 49 monthsor more 17% 38% 68
stress in 1988. Equations are computed StressSymptoms
expressingtherelationbetweentimeelapsed 24 monthsor less 4.8* 1.3 33
and stresswhilecontrolling forstressin 1980 25 monthsor more 4.3 1.1 39
(1983 in the case of the stresssymptoms * Differencesbetweenmeansstatistically
significant
scale). By controlling forpredivorcestress, at .05 level.
each respondent servesas his/herown con- at ** Differences
.01 level.
betweenmeansstatistically
significant
trol,thereby reducingthechancesof extrane-
ous factorsbiasingour results.2Because we
expectstressto decreasewiththepassage of cases foreach periodrevealedthattheresults
time, a one-tailed F test of statistical obtainedby us werenotdue to an outlieror a
significanceis used. The results of the fewinfluential cases. In anotheranalysis(not
analysisare presented in Table 1. shown)genderwas enteredintotheequations
Examination ofthemeansindicatesthatthe to see whetherit interacts withelapsed time
high stresswhich accompaniesdivorcede- as it relatesto subjectivestress.We foundno
clineswiththepassage of time.For all three evidencethatsome symptoms persistamong
stressmeasures,recoveryapproachesnormal womenwho aretypically moredisadvantaged
levelsin 24 months.Normalis definedas the by divorcethanare men.
stresslevels reportedby the non-divorced To investigatethe role of remarriagein
population.3An examinationof individual post-divorceadjustment,we examinedthe

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
DIVORCE AND PSYCHOLOGICAL STRESS 403

interactionamongremarriage, lengthof time dependentvariables may be due to the

since divorce,and our measuresof psycho- differences in theseverityof stresstappedby
logical stress.The analysisrevealedthatthe our threemeasures.Nearlyeveryonesuffers
interactionswerenotsignificant forunhappi- some unhappiness,depression,and anxiety
ness, psychologicaldistress,or stresssymp- duringthecourseof divorce,butmoresevere
toms.Examination of themeansrevealedthat stressas indicatedby psychosomatic symp-
for all outcomes,remarried individualshad tomsmayonlybe experienced by thosewho
lowerlevelsof stressat all threetimeperiods, aredisadvantaged bylimitedresources,attach
but the same patternof improvement over great value to their marriage,or believe
timewas apparentforbothgroups. divorceis morallywrong.4
The modelin whichcertainfactorsmoder- In relationto resources,age is not an
ate the psychologicalstressexperiencedby advantagenoris social resources,buthaving
theindividual following divorcewas testedby a familyincomeabove themedian,education
creatinginteractionterms(multiplying the beyondhigh school, or an employedwife
1980 moderating variables[1983 whenstress suggeststhateithertheinitiallevels of stress
symptoms is the dependentvariable]by the areless orthatadvantaged individualsrecover
time that has elapsed since divorce) and more quickly.The two outlookvariables-
entering themintoregression equationsalong perceivingthe marriageas troubledand
withthemoderating variableandelapsedtime havingan attitude favoringthedissolutionof
while controllingfor subjectivewell-being unhappymarriages-showa similarpattern.
beforethedivorce.A one-tailedtestwas used In an additionalanalysis (not shown), no
sincewe predictthedirection of themoderat- gender differenceswere observed in this
ingvariableeffects. pattern.The adjustedmeanlevelsof psycho-
One of theinteraction termsis statistically logicalstressforthemoderating variablesby
significant(or nearly so) in the case of time since divorcealong withthe standard
psychologicalstress,none in the case of deviationsand numberof cases are shownin
unhappiness, and fivewhenthestresssymp- Table 2.
toms scale is the dependentvariable. The Severalconsistent emergefromthe
patterns
inconsistencyin the effectsfor the three analysis. First, people in the pre-divorce
TABLE 2. Mean Levels of PsychologicalStressby ModeratingVariables and Time Since Divorce
(24 monthsor less/25monthsor more) AdjustedforPre-divorceStress'
</=24 months (Std Dev/N) 25 + months (Std Dev/N)
Incomex Symptoms
$23,000 or more 4.5 (1.2/19) 4.423 (1.1/22)
$22,999 or less 5.3 (1.5/11) 4.4 (1.2/20)
Educationx Symptoms
Post-highschool 4.4 (1.1/18) 4 a bc (1.1/29)
Highschoolor less 5.3 (1.4/12) 3.9 (1.0/13)
Employment x Symptoms
Wifeemployed 4.5 (1.2/21) 4.4amb (1.1/34)
Wifenotemployed 5.5 (1.5/9) 3.7 (1.2/7)
Problemsx Symptoms
Fouror more 4.5 (1.2/19) 454 (1.3/20)
Threeor less 5.1 (1.6/11) 4.2 (1.0/21)
Problemsx Distress
Fouror more 24% (41%/17) 24% (44%/49)
Threeor less 44% (52%/14) 10% (21%/57)
Moralityx Symptoms
Divorceokay 4.4 (1.0/10) 4.63 (1.2/14)
Divorceimmoral 5.0 (1.5/20) 4.2 (1.1/28)
1 at the.05 levelorbetter(one-tailed
All interaction
termsarestatistically significant test)exceptmorality
ofdivorce
by symptoms, whichis significant at the .06 level.
The superscript "a" indicatesthatthedifference in meansbetweencells A and C is statistically at the
significant
.05 levelor better.Superscript"b" has thesamemeaningforcells C andD, "c" forcells B andD, and "d" forcells
A and B. The lettersA, B, C, and D referto standardcell designations as follows: A B
C D
3The differences betweenA and C, and C and D are marginally significant at the.06 level.
4 The difference betweenC and D is marginally significantat the.07 level.

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
404 JOURNALOF HEALTH AND SOCIAL BEHAVIOR
disadvantaged categoryshowhigherlevelsof Analysisofpaneldatafroma nationalsample
stresstwoyearsaftera divorcethanindividu- of people marriedin 1980 indicatesthatthe
als in theadvantagedcategoryoverthesame crisismodelis appropriate forunderstanding
time period. While only four of the six adjustment to divorcefor most, if not all,
differencesare statisticallysignificant(or individuals. Comparisonsof people who
nearlyso), theremainder are in thepredicted divorcedwith those who did not show a
direction.Second, people in the disadvan- predivorce risein stresswhichthenreturns to
tagedcategory forwhommorethantwoyears levelsof stressreported by marriedindividu-
have elapsedsincedivorcehave lowerlevels als. We foundlittleevidencesupporting the
of stressthandisadvantaged individualswho idea thata chronichighlevelofpsychological
are within24 monthsof a divorce.All six stressis a generalcause of divorceor that
mean differences are statisticallysignificant dissolutionresultedin moreor less permanent
(or nearlyso). Third,people in the advan- levelsof highstress.The inferences we draw
taged categorieswithintwo years of their fromouranalysismustbe regarded withsome
divorceare no morestressedthanadvantaged caution,however,sincewe do nothavestress
individualswho are more than 24 months levels at set periods(e.g. everyyearor two)
away fromdivorce. In no case does the beforeand afterdivorce.
difference even approachstatisticalsignifi- Our resultsappearto contradict Menaghan
cance. Fourth,people in the disadvantaged (1985) who foundno pre-divorce differences
category tend to show less stress than in stress.It is possiblethatthesmallnumber
advantagedindividualsonce divorceis more of divorcedindividuals in herstudy(32) may
than two years old. While this patternis have contributed to herresult.However,we
statistically
significantin onlytwo of thesix thinkthe difference betweenthetwo sets of
cases, thepattern is evidentin fiveof thesix resultsmayrestin themeasuresused. In an
cases. analysis (not shown), the tendencyfor
Giventhefindingthatdisadvantaged indi- unhappinessand psychologicaldistresspre-
vidualsexperiencepsychosomatic symptoms cedes divorceby a longperiodof time(up to
in connectionwithdivorcebut thatadvan- threeyears),butpsychosomatic stresssymp-
tagedones do notappearto do so to thesame tomselevateonlyone yearbeforedivorce.It
extent,one could arguethatthecrisismodel appears thatas maritalseparationbecomes
is appropriateforthedisadvantaged groupbut imminent, signsof severestressshowup just
notfortheadvantaged one. Whilethismaybe a yearpriorto divorce,if at all. Menaghan's
the case for advantagedindividualson this pre-divorce assessmentof stress(comparable
one measureof stress,it is important to keep to ourmeasureof symptoms) was obtainedup
in mindthatour analysisof themeasuresof to four years prior to divorce. Had she
unhappiness and psychologicaldistress(with assessed depressiona few monthspriorto
one exception)are consistent withthe crisis divorcein a largernumberof people, she
model.To fullytestwhether thecrisismodel mighthave foundevidenceof highpsycho-
is appropriate for psychosomatic symptoms logical stressbeforedivorce. Methodologi-
among advantagedindividuals,a thirddata cally, this means that futureresearchon
pointtwoto fouryearshenceis needed.Thus stressful lifeeventsshouldincorporate several
whilewe cannotrejectthepossibility thatthe measuresof stresstappingdifferent levels of
crisismodelonlyfitsdisadvantaged individu- severity.
als, withrespectto twomeasuresof stressthe Our results also appear to contradict
datado notsupportthisinterpretation. McLanahan's(1983) studyin whichpsycho-
logicalproblemswereobservedto persistfor
yearsamongwomenin single-parent families.
SUMMARY AND DISCUSSION The discrepancy maybe due to thefactthat
McLanahan'ssamplewas drawnlargelyfrom
While researchon adjustment to divorce a lowersocioeconomicpopulation,and all of
has been extensive,the paucityof studies herrespondents weremothers.Higherstress
assessingstressbeforeand afterdivorcehas may indeedbe a chronicproblemforpoor,
kepttherelationbetweenpsychological stress singlemothers.However,ourresultssuggest
and maritaldissolutionunclear.Conflicting thatpsychologicalstressdecreasesovertime
findingscast some doubt on the utilityof in the general population regardless of
using a crisis model to study divorce. whether or notthepersonremarries.

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
DIVORCE AND PSYCHOLOGICAL STRESS 405
Our findingssupportthe hypothesisthat numberof cases to such a low level that
pre-divorce resourcesand outlooksinfluence meaningful analysisis precluded.
the amount of psychosomaticsymptoms 2. Giventheevidencethatstressmayincreasein
experiencedin the two years immediately the years prior to divorce, controllingfor
pre-divorce levelsof stresscouldbe viewedas
followingdivorce. Below median family
complicating the interpretation
of post-divorce
incomes,no post-high schoolexperience,and stressscores since individualswho differin
wife not in the labor force put divorcing post-divorce timesince divorcealso differin
individualsat a disadvantage.Individuals pre-divorce timebeforedistress.To checkout
reporting few maritaltroublesand beliefsin this possibility,the equations were re-run
the immoralityof divorce also appear to leavingout the pre-divorce level as a control
experienceheightened psychosomatic symp- and controlling forsex, maritalduration,and
tomsin thetwoyearsfollowingdivorce. educationinstead.The patternof means was
Our analysisseems to indicatethateven identical,but in some cases the time since
disadvantagedindividualsreturnto normal divorce variable was no longer statistically
levels of stress aftertwo years. In fact, significant.By leaving out the pre-divorce
stressmeasure,the standarderrorincreases.
disadvantagedpersonsmay do even better Since the patternof means is the same, we
than advantagedindividualsthreeor more elected to controlfor pre-divorcelevels of
years afterdivorce.We can only speculate stress in order to controlfor a varietyof
aboutthisunexpected tendency.Perhapsthere potentially biasingfactors.We did this only
is a greatersense of reliefand fewerstress afterexamining thestandard deviationsand the
symptomsfollowingextremestress(experi- case-by-casedataforoutliers.
enced by less educatedindividuals),thatis 3. Meanunhappiness forthenondivorcing popula-
not evidentamong those who reportlower tionis 1.6; forpsychological distress,.18; and
levels of stress(experiencedby those with forstresssymptoms, 4.6.
greatereducation).As peoplelookbackto the 4. Evidencethatphysicalsymptoms are regarded
as a moresevereformofstressmaybe foundin
divorce,thedisadvantaged maysee thattheir theAmericanPsychiatric AssociationDiagnos-
conditionshave improvedso muchthatin a ticand StatisticalManual ofMentalDisorders
relativesense theyfeel theyare muchbetter (1987). For example,for an individualto be
off.If thechangeis viewedas achievement, diagnosed as having a generalizedanxiety
the greateraccomplishments may resultin disorder,thepatientmustshow signssuch as
fewersignsof stress.The advantaged person, sweating hands and abdominal distress in
on the otherhand, may feel only marginal additionto excessiveanxiety.
improvement has occurredsince he or she
was notas stressedby thedivorcein thefirst
place. As a resulttheirstresssymptoms do REFERENCES
not change appreciably.A similarsort of
tendencywas foundby Ambert(1984) in a Ambert, Anne-Marie. 1984. "Longitudinal
longitudinalstudyof separatedmothers.She Changesin Children'sBehaviorTowardCusto-
discovered that two and one-half years dial Parents." Journalof Marriage and the
followingseparation,low income custodial Family46:463-67.
Amato,Paul and Sonia Partridge. 1987. "Widows
parentsreportedmuch greatersatisfaction and DivorceeswithDependentChildren:Mate-
with theirchildrenthan did high income rial,Personal,Family,and Social Well-being."
mothers.It seemed that the departureof FamilyRelations36:316-20.
disruptivechildrenaccountedforpartof this AmericanPsychiatric Association.1987. Diagnos-
difference. Understanding the exact mecha- ticand StatisticalManual ofMentalDisorders.
nismat workwill have to await subsequent Washington, DC: AmericanPsychiatric Associ-
research. ation.
Bloom, BernardL., ShirleyJ. Asherand Stephen
W. White. 1978. "Marital Disruptionas a
NOTES Stressor:A Reviewand Analysis."Psychologi-
cal Bulletin85:767-801.
1. We exploredthe option of consideringthe Bloom,BernardL., RobertL. Niles and AnnaM.
absenceof childrenas a resource.We elected Tatcher.1985. "Sourcesof MaritalDissatisfac-
not to include childrenbecause they are a tionamongNewlySeparatedPersons."Journal
whentheyare dependent
liabilityprimarily and ofFamilyIssues 6:359-73.
because theirpresenceseemsto affectwomen Booth, Alan and David R. Johnson. 1985.
morethanmen. These two factorsreducethe "TrackingRespondentsin a TelephoneInter-

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
406 JOURNALOF HEALTH AND SOCIAL BEHAVIOR
view Selected by Random Digit Dialing." and KathleenS. Placidi. 1989. "Adjustment
to
SociologicalMethodsand Research14:53-64. Widowhoodand Divorce." Journalof Family
Booth, Alan, David Johnson,Lynn White and Issues 10:5-32.
JohnEdwards. 1985. "PredictingDivorce and Kitson,Gay C. and MarvinB. Sussman. 1962.
PermanentSeparation." Journal of Family "MaritalComplaints, Demographic Characteris-
Issues 6:331-46. tics, and Symptomsof Mental Distress in
Bower, GordonH. 1981. "Mood and Memory." Divorce." Journalof Marriageand theFamily
AmericanPsychologist 36:129-48. 44:87-101.
Colletta,NancyDonohue. 1983. "StressfulLives: Langner,ThomasS. 1962. "A Twenty-two Item
The Situationof DivorcedMothersand Their ScreeningScale of PsychiatricSymptomsIndi-
Children."JournalofDivorce6:19-31. cating Impairment."Journalof Health and
Doherty,William J., Susan Su and Richard Social Behavior3:269-76.
Needle. 1989. "MaritalDisruption and Psycho- McCubbin,HamiltonI. and JoanM. Patterson.
logical Well-being."Journalof FamilyIssues 1982. "Family Adaptationsto Crisis." Pp.
10:72-85. 26-47 in Family Stress, Coping, and Social
Dohrenwend,Bruce P. and BarbaraS. Dohren- Support,editedby HamiltonI. McCubbin,A.
wend. 1976. "Sex Differences and Psychiatric Elizabeth Cauble, and Joan M. Patterson.
Disorders." AmericanJournal of Sociology Springfield,IL: CharlesC Thomas.
81:1447-54. McLanahan,Sara S. 1983. "FamilyStructure and
Duncan,GregJ. and Saul D. Hoffman.1985. "A Stress: A LongitudinalComparisonof Two-
Reconsiderationof theEconomicConsequences parentandFemale-headed Families."Journalof
of Marital Dissolution." Demography 22: Marriageand theFamily45:347-57.
487-95. Menaghan,ElizabethG. 1985. "DepressiveAffect
Gove, WalterR., CarolynB. Styleand Michael and SubsequentDivorce." Journalof Family
Hughes. 1990. "The Effectof Marriageon the
Issues 6:295-306.
Well-being ofAdults." JournalofFamilyIssues
Menaghan,ElizabethG. and MortonA. Lieber-
11:4-35.
man. 1986. "Changesin DepressionFollowing
Gove,WalterR. andHee-ChoonShin. 1989. "The
Divorce:A Panel Study."Journalof Marriage
Psychological Well-being of Divorced and
WidowedMen andWomen."JournalofFamily and theFamily48:319-28.
Issues 10:122-44. Pearlin,Leonard, MortonLieberman,Elizabeth
Groves,Robertand RobertKahn. 1979. Surveys Menaghan and Joseph Mullen. 1981. "The
by Telephone:A National Comparisonwith StressProcess." Journalof Healthand Social
Personal Interviews.New York: Academic Behavior22:337-56.
Press. Pearlin,LeonardI. and JoyceL. Johnson.1977.
Hetherington, E. Mavis, MarthaCox and Roger "MaritalStatus,Life Strainsand Depression."
Cox. 1982. "Effectsof Divorceon Parentsand AmericanSociologicalReview42:704-15.
Children."Pp. 233-88 in Nontraditional Fami- Tschann,JeanneM., JanetR. Johnston andJudith
lies: Parentingand Child Development, edited Wallerstein.1989. "Resources,Stressors,and
by MichaelLamb. Hillsdale,NJ:Erlbaum. Attachment as Predictorsof AdultAdjustment
Hill, Reuben. 1949. Families UnderStress.New AfterDivorce:A Longitudinal Study."Journal
York:Harperand Row. ofMarriageand theFamily51:1033-46.
Holmes,ThomasH. and RichardH. Rahe. 1967. Weiss, Robert.1979. GoingitAlone: TheFamily
"The Social Readjustment RatingScale." Jour- Lifeand Social Situationof theSingleParent.
nal ofPsychosomatic Research11:213-18. New York:Basic Books.
House, JamesS., Debra Umbersonand Karl R. Weitzman,LenoreJ. 1985. The DivorceRevolu-
Landis. 1988. "Structuresand Processes of tion: The UnexpectedSocial and Economic
Social Support."Pp. 293-318 inAnnualReview Consequencesfor Women and Children in
ofSociology,Volume14, editedby W. Richard America.New York:Free Press.
Scottand Judith Blake. Palo Alto,CA: Annual Wertlieb, Donald, Simon Budman, Annette
Reviews. Demby and Mary Randall. 1984. "Marital
Kitson,Gay C. 1982. "Attachment to theSpouse Separation andHealth:StressandIntervention."
in Divorce: A Scale and Its Application." JournalofHumanStress10:18-26.
JournalofMarriageand theFamily44:379-93. Wheaton,Blair. 1990. "Life Transitions,Role
Kitson,GayC. andLeslie A. Morgan.1990. "The Histories,andMentalHealth."AmericanSocio-
MultipleConsequencesof Divorce: A Decade logicalReview55:209-23.
Review." Journalof Marriageand theFamily White, Lynn and Alan Booth. 1991. "Divorce
52:913-24. Over the Life Course: The Role of Marital
Kitson,Gay C., KarenB. Babri,MaryJ. Roach Happiness."JournalofFamilyIssues 12:5-21.

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions
DIVORCE AND PSYCHOLOGICAL STRESS 407
ofSociologyatThe Pennsylvania
Alan Boothis Professor StateUniversity.
His researchinterests
include
over the life course, and the role of testosterone
factorsthatinfluencemaritalinstability human
competition.

Paul Amato is AssociateProfessorof Sociologyat the University

of Nebraska-Lincoln.Amonghis
are the consequencesof parentaldivorceand maritalqualityon the well-beingof
researchinterests
offspring.

This content downloaded from 132.203.227.62 on Sat, 29 Nov 2014 16:53:32 PM

All use subject to JSTOR Terms and Conditions