Marriage, cohabitation and mental health

My research shows that marriage is indeed followed by significant improvements in people's mental health. Following marriage, people report fewer symptoms of ... Skiptomaincontent Home Research FamilyMatters FamilyMattersNo.96,2015 Marriage,cohabitationandmentalhealth Contenttype FamilyMattersarticle Published June2015 DownloadFamilyMattersarticle Marriage,cohabitationandmentalhealth 965.08KB Abstract Researchconsistentlyshowsthatmarriedpeoplehavebettermentalhealththansinglepeopledo.However,theresearchisunclearonwhethermarriagecausesimprovementsinmentalhealthorwhetherpeoplewithbettermentalhealtharemorelikelytomarry,andwhetherthebenefitsofmarriageextendequallytowivesandhusbandsandalsotonon-maritalrelationshipssuchascohabitation.ThisarticlelooksatfindingsfromanewU.S.studythatseekstoexplorethesequestions:theNationalLongitudinalStudyofAdolescenttoAdultHealth(AddHealth). Researchconsistentlyshowsthatmarriedpeoplehavebettermentalhealth,onaverage,thandosinglepeople.Thisgeneralconclusionappliestoarangeofoutcomes,includingdepression(Brown,2000;Ross,1995),happiness(Zimmerman&Easterlin,2006),lifesatisfaction(Williams,2003),psychologicalwellbeing(KampDush&Amato,2005)andmortalityfromsuicide(Rogers,1995).Moreover,themarriageadvantagehasbeendemonstratedinavarietyofcountriesandregions,includingtheUnitedStates,Australia,NewZealand,EuropeandAsia(Diener,Gohm,Suh,&Oishi,2000;Lee&Ono,2012;Soons&Kalmijn,2009).Compellingreasonsexistforwhymarriagemightbegoodforpeople'smentalhealth.First,marriageisanimportantsourceofcompanionship,intimacyandsocialsupport(Waite&Gallagher,2000).Marriagealsoconnectsspouseswithoneanother'ssocialnetworks,thusexpandingthenumberofpeoplewhocanbedrawnonforassistance.Second,peoplebenefitfromtheinstitutionalnatureofmarriage(Cherlin,2004).Marriageinvolvessocialnormsandexpectationsthatclarifyspouses'rightsandresponsibilitiestowardoneanotherandreducerelationshipambiguity.Moreover,throughmarriage,peopleachieveapositivelyvaluedsocialstatusthatotherpeoplerespectandsupport.Andbecausemarriageisinstitutionalised,spousesacquiremanylegalbenefits.IntheUnitedStates,forexample,thesebenefitsincludeaccesstothespouse'shealthinsurance,taxdeductionsforone'sspouse,theoptiontofilejointtaxreturnsandtherighttomakemedicaldecisionsforone'sspouse.Third,thelong-termcommitmentimpliedbymarriagereducesrelationshipinsecurity,andthegradualaccumulationofasharedhistorywithone'sspouseisasourceofmeaningandidentitytomanypeople.Fourth,marriageprovidesfinancialadvantagesoversinglehood,includingeconomiesofscaleandtheabilitytopoolincomeandaccumulatewealthmorerapidly.Despitetheconsistencyofresearchfindings,severalpointsofambiguityremaininthisresearchliterature.Itisnotclearwhether:theassociationbetweenmarriageandmentalhealthiscausalorduetotheself-selectionofhealthierpeopleintomarriage;thebenefitsofmarriagepersistindefinitelyorfadeovertime;thebenefitsofmarriageextendequallytowivesandhusbands;andthebenefitsofmarriageapplytoothertypesofromanticrelationships,suchasnon-maritalcohabitation.Theseambiguitiesintheresearchliteratureledmetoinitiateaprogramofresearchontheeffectsofmarriageandotherrelationshiptransitionsonpeople'smentalandphysicalhealth.Theresearchdescribedinthisarticleinvolvesonepartofthatlargerprogram.Thecurrentreportdrawsonalarge,longitudinaldatasetintheUnitedStates-theNationalSurveyofAdolescenttoAdultHealth-andaddresseshowthetransitiontocohabitationandmarriageaffectsmen'sandwomen'sreportsofdepressivesymptomsandthoughtsofsuicide.Doesmarriagecausechangesinmentalhealth?Becausepeopledonotmarryatrandom,itisdifficulttodeterminewhethermarriageimprovespeople'smentalhealth(acausalhypothesis),orwhetherpeoplewithbettermentalhealtharemorelikelytomarry(aselectionhypothesis).Giventheimpossibilityofconductingexperiments,fixedeffectsmodelsarearguablythebestavailablemethodtocontrolforselectioneffectswhenusingcorrelational,longitudinaldata(Allison,2009).Anadvantageoffixedeffectsmodelsisthattheycontrolforallunmeasured,time-invariantfeaturesofpeople,suchasraceandethnicity,stablepersonalitytraits,cognitiveability,familyoforigincharacteristicsandmanygeneticfactors.Becausefixedeffectsmodelsinvolveonlywithin-personvariation,eachpersonservesashisorherown"control".Appliedtothecurrenttopic,thismethodanswersthequestion:Doespeople'smentalhealthimproveaftertheymarry?Fourstudieshaveusedfixedeffectsmodelstoestimatetheeffectsofmarriageonhealth.ZimmermanandEasterlin(2006)foundthatmarriagewasfollowedbyanincreaseinlifesatisfactionina20-yearlongitudinalGermandataset.Theyobservedasimilarbutweakereffectfornon-maritalcohabitation.Althoughlifesatisfactiondeclinedmodestlyafterthefirstyearofmarriage,itremainedhigherthanithadbeenduringthesingleyears.Soons,Liefbroer,andKalmijn(2009)reachednearlyidenticalconclusionsusingan18-yearlongitudinaldatasetfromtheNetherlands.IntheUnitedStates,MusickandBumpass(2012)foundthattransitionsintomarriagebetweenthefirsttwowavesoftheNationalSurveyofFamiliesandHouseholdswereassociatedwithincreasesinhappinessanddeclinesindepression,providedthatcouplesdidnotdivorce.Thesechangesweremodestinmagnitude,however,andtendedtodissipateovertime.Incontrast,WuandHart(2002)didnotfindthatmarriagebetweenthefirsttwowavesoftheCanadianNationalPopulationHealthSurveywasassociatedwithchangesindepression.Theydidnote,however,thatthelongerpeoplestayedmarried,themoredepressedtheybecame.Insummary,threeofthefourstudiesthathaveusedfixedeffectsmodelssuggestthatmarriagehasapositive,causaleffectonseveraldimensionsofmentalhealth,includingsatisfactionwithlife,self-reportedhappinessanddepressivesymptoms,althoughtheCanadianstudyprovidesacontraryresult.Moreover,allfourstudiessuggestthatwellbeingdeclinesafterthefirstyearofmarriage,althoughtheamountofdeclinevariesacrossstudies.Noneofthefourstudiesfoundconsistentevidenceofgenderdifferences.Giventhesmallnumberofstudiesthathaveusedfixedeffectsmodels,however,additionalresearchisnecessarytoestablishthebroadergeneralityofthesefindings.Marriageversusnon-maritalcohabitationNon-maritalcohabitation,likemarriage,providespeoplewithcompanionship,intimacyandeverydayassistance.Andlikemarriedcouples,cohabitingcouplesbenefitfinanciallyfromeconomiesofscale.Forthesereasons,livingtogethermayprotectpeople'smentalhealthinwayscomparabletomarriage.Ofcourse,cohabitationislessinstitutionalisedthanmarriage,andcohabitingpartnersdonothavethesamelong-termtimehorizonasmarriedcouples.Moreover,cohabitingindividualsreportlessrelationshiphappinessandcommitmenttotheirpartnersthandomarriedindividuals(Brown,2000;Brown&Booth,1996;Nock,1995).Consistentwiththesetrends,cohabitingrelationships(thatdonottransitiontomarriage)arelessstablethanmarriages(Bramlett&Mosher,2002;Kennedy&Bumpass,2008).ThisistruenotonlyintheUnitedStates,butalsoinothercountries,includingtheScandinaviancountries(Kennedy&Thomson,2010)andAustralia(Wilkins&Warren,2012).Theseconsiderationssuggestthatwithrespecttomentalhealth,cohabitingindividualsmaybebetteroffthansingleindividuals,onaverage,butnotaswelloffasmarriedindividuals-aconclusionsupportedbycross-sectional(Brown&Booth,1996;KampDush&Amato,2005)aswellaslongitudinalstudies(Soonsetal.,2009;Zimmerman&Easterlin,2006).Comparativestudiesalsoshow,however,thatthemagnitudeofthemarriageadvantagedependsonavarietyofcontextualcharacteristics.Inparticular,thegapinwellbeingbetweenmarriedandcohabitingindividualsislargerincountrieswherenon-maritalcohabitationisrelativelyuncommonandlowinsocialacceptance(Soons&Kalmijn,2009),traditionalgenderrolespersistandreligiosityishigh(Lee&Ono,2012),andthecultureemphasisescollectivismratherthanindividualism(Dieneretal.,2000).Giventhattheadvantageofmarriageovercohabitationmaybesociallyandhistoricallyspecific,additionalresearchthatcomparestheeffectsofcohabitationandmarriageonmentalhealthindifferenttimesandplacesiswarranted.Wivesaswellashusbands?Scholarswritinginthe1970s(e.g.,Bernard,1972;Gove&Tudor,1973)oftenassumedthatmarriageislessadvantageousforwivesthanforhusbands.Inthepast,manywomenquittheirpaidjobswhentheymarried(orbecamemothers)andsacrificedtheirautonomyfortheirrelationships.Moreover,wiveswhoworkedathome,comparedwithhusbands,hadlesspower,engagedinlesssatisfyingwork,andperformedworkthatwasvaluedlessbysociety.Andwiveswhodidcontinueinthelabourforceoftenfoundthattheywereforcedtodoasecondshiftwhentheygothome(Hochschild&Machung,1989).Accordingtothisview,marriageinthepastimprovedthementalhealthofhusbandsbuthadnoeffect,orevenadetrimentaleffect,onthementalhealthofwives.Ofcourse,therolesofhusbandsandwiveshavechangedagreatdealinthelast40years.Andalthoughmarriagesinanearliereramayhavebeenlessbeneficialforwivesthanhusbands,morerecentstudieshavefoundfewgenderdifferencesintheestimatedeffectofmarriageonmentalhealthandgeneralpsychologicalwellbeing(Musick&Bumpass,2012;Ross,1995;Soonsetal.,2009;Williams,2003;Zimmerman&Easterlin,2006).Despitethelackofsupportforgenderdifferencesinthisarea,persistentconcernsaboutgenderinequalitysuggestthatitwouldbeprematuretoabandonthesearchforgenderdifferencesinmarriageandothercloserelationships.Moreover,cohabitingrelationshipstendtobemoreequalthanmarriages,atleastwithrespecttothehouseholddivisionoflabor(Baxter,2001).Thisobservationraisesthepossibilitythatcohabitationmaybeabetterarrangementthanmarriageforsomewomen'smentalhealth.ThecurrentstudyMycurrentresearchdrawsontheNationalLongitudinalStudyofAdolescenttoAdultHealth(AddHealth)tounderstandhowthetransitiontomarriageaffectsmentalhealth.AddHealthiswellsuitedtoassessingtheeffectsofmarriagebecauseitfollowsrespondentsfromadolescencethroughtotheirearly30s-theperiodwhenmostpeoplemarryforthefirsttime.Thedatasetalsoincludestwoexcellentindicatorsofmentalhealth:depressionandsuicideideation.Thecurrentstudyreliesonfixedeffectsmodels,whichcontrolforallstable(time-invariant)personalcharacteristicsofindividuals.Althoughfixedeffectsmodelscontrolforallstableindividualtraits,theydonotcontrolfortraitsthatchangeovertime.Forthisreason,thecurrentanalysisincludescontrolsforage(andagesquaredtoaccountfornon-linearity),yearsofeducation,hoursofemploymentandwhetherrespondentshavebiologicalchildren.Myresearchalsoconsiderswhethertheestimatedeffectsofmarriagedeclineovertime.Althoughseveralstudiessuggestthatthebenefitsofmarriagefadeafterseveralyears,long-termmarriagesmaystillbepreferabletosinglehood.Thecurrentstudyalsoconsiderswhethernon-maritalcohabitationisfollowedbyimprovementsinmentalhealth.Despitethefactthatcohabitatingrelationshipstendtobelesshappyandstablethanmarriages,priorresearchsuggeststhatbothtypesofrelationshipsaregoodforpeople'slifesatisfactionandmentalhealth.Finally,thepresentstudyfocusesongenderdifferencesinhowmarriageandcohabitationaffectmentalhealth.Althoughsomepriorresearchsuggeststhatbothwomenandmenbenefitfrombeingincloserelationships,thepersistenceofmaritalinequalitysuggeststhatmarriagemaystillbeamorebeneficialarrangementforhusbandsthanwives.MethodDatacamefromWaves1,3and4oftheAddHealthdataset.AddHealthstartedasanationallyrepresentativesurveyof20,745adolescentsinGrades7through12intheUnitedStatesin1994-95.Inthefirstwave,datawerecollectedthroughin-homeinterviewswithadolescentsandoneoftheirparents.Youthwereinterviewedasecondtimein1996,athirdtimein2001-02,andafourthtimein2007-08.(Ididnotusethesecondwavebecauseitfollowedthefirstwavebyonlyoneyear.)Thefinalsampleincluded18,924respondentsforwhomatleasttwowavesofdatawereavailable;thatis,Waves1and3,Waves1and4,orWaves1,3and4.(SomerespondentswhomissedWave3weretrackedandinterviewedinWave4.)Table1includesdescriptiveinformationaboutthesamplebysurveywave(weighted).Themeanage(inyears)increasedfromalittleover16inWave1,to29inWave4.(Theoldestrespondentswere34atthefinalwave.)Atthetimeofthefirstinterview,theaverageadolescenthadcompletedbetween9and10yearsofeducation.ThecorrespondingfigureforWave4wasabout14years.Meanweeklyhoursofemploymentincreasedfromabouteightduringadolescencetoover41inWave4.Only2%ofadolescentshadbecomebiologicalparentsbythetimeofthefirstinterview,althoughthisfigureincreasedtooverhalfofthesampleinWave4.Table1:Descriptiveinformationofsampleshowingvariablemeansandproportions,bysurveywave Wave1Wave3Wave4Source:AddHealthAge16.1922.4829.01Yearseducation9.7013.2214.28Hoursemployed7.7528.2641.59Havechildren.02.26.51Marriage.00.18.43Cohabitation.00.15.18Depression-.05-.06.08Suicideideation.13.07.07Table1alsoshowsinformationonrelationshipstatusatthetimeofthesurvey.ByWave3,18%weremarried,and43%byWave4.Correspondingly,15%ofyouthswerecohabiting(butunmarried)byWave3and18%byWave4.Notethatthesefiguresunderestimatethepercentageofrespondentswhoevermarriedorcohabited,becauserelationshipstatusiscodedatthetimeofthesurvey.Respondentswhomarriedorcohabitedandbrokeupbetweensurveysarenotincludedinthesefigures.Thisqualificationisespeciallyrelevanttocohabitations,whichgenerallylastonlyayearortwointheUnitedStatesbeforebreakinguportransitioningintomarriage.Althoughnotshowninthetable,themeandurationofmarriageformarriedrespondentswas2.2yearsinWave3and4.7yearsinWave4.Correspondingly,themeandurationofcohabitationwas1.9yearsinWave3and3.1yearsinWave4.(AhandfulofadolescentsweremarriedorcohabitinginWave1;thesecasesweredroppedfromtheanalysisbecausenoinformationonmentalhealthwasavailablepriortorelationshipformation.)AmeasureofdepressionwasbasedonnineitemsfromtheCenterforEpidemiologicalStudiesDepressionScale.Theseitemswerewordedidenticallyinallsurveywaves,andthemeanalphareliabilityacrossallwaveswas.81.Itemswereaddedandthetotalscorewasstandardisedacrosswavestoimproveinterpretability(mean=0,standarddeviation=1).Table1showsthatthemeanlevelofdepressionwasrelativelystableacrossallthreewaves.Thesecondmeasureofmentalhealthwasaquestionaboutsuicide:"Duringthelastsixmonths,didyoueverthinkseriouslyaboutcommittingsuicide?"Theproportionofpeoplerespondingpositivelytothisquestiondeclinedslightlybetweenadolescenceandearlyadulthood.ResultsTable2showstheresultsfromregressingdepressionscoresonmarriageandcohabitation.Althoughnotshown,age,agesquared,yearsofeducation,weeklyworkhoursandhavingbiologicalchildrenwereincludedascontrolvariablesinthemodels.Separatemodelswereestimatedformenandwomen.Forbothgenders,thetransitiontomarriagewasassociatedwithadeclineinsymptomsofdepression.Moreover,themagnitudeofthedeclinewasthesameformen(-.15)andwomen(-.15).Table2:Fixedeffectsregressionofdepressivesymptoms(zscores)onrelationshipstatusandduration,bygender MenWomenNotes:Modelsincludecontrolsforage,agesquared,yearsofeducation,weeklyworkhoursandhavingchildren.Significancetestsarebasedonrobust,clusteredstandarderrors.*p<.05 figure2showstheoveralltrajectoryofdepressivesymptomsforwomen.comparingfigure2withfigure1revealsthatd epressionscores table3showstheresultsofregressingsuicideideationonmarriageandcohabitation.forbothgenders figure4showstheprobabilityofsuicideideationforwomen.thesolidlineshowstheprobabilityofsuicideideation scoresforsinglewomen.thetrendlineforsinglewomenrevealsaprominentdeclinebetweenadolescenceandthemid-2 discussionmanystudieshavefoundthatmarriedpeoplearehappierandhealthierthansinglepeople.butdoesthismea nthatmarriageisgoodforpeople>.Williams,K.(2003).Hasthefutureofmarriagearrived?Acontemporaryexaminationofgender,marriage,andpsychologicalwell-being.JournalofHealthandSocialBehavior,44(4),470-487.Wu,Z.,&Hart,R.(2002).Theeffectsofmaritalandnon-maritaluniontransitiononhealth.JournalofMarriageandFamily,64(2),420-432.Zimmerman,A.C.,&Easterlin,R.A.(2006).Happilyeverafter?Cohabitation,marriage,divorce,andhappinessinGermany.PopulationandDevelopmentReview,32(3),511-528.ProfessorPaulR.AmatoisArnoldandBetteHoffmanProfessorofFamilySociologyandDemographyatthePennsylvaniaStateUniversity,USA.ThisarticleisbasedonthekeynoteaddressgivenbyPaulAmatoatthe13thAustralianInstituteofFamilyStudiesConference:FamiliesinaRapidlyChangingWorld,Melbourne,1August2014.Acknowledgements:ThispublicationusesdatafromAddHealth,aprogramprojectdirectedbyKathleenMullanHarrisanddesignedbyJ.RichardUdry,PeterS.Bearman,andKathleenMullanHarrisattheUniversityofNorthCarolinaatChapelHill,andfundedbygrantP01HD31921fromtheEuniceKennedyShriverNationalInstituteofChildHealthandHumanDevelopment,withcooperativefundingfrom23otherUSfederalagenciesandfoundations.SpecialacknowledgementisduetoRonaldR.RindfussandBarbaraEntwisleforassistanceintheoriginaldesign.NodirectsupportwasreceivedfromgrantP01HD31921forthecurrentanalysis.ThePennStatePopulationResearchInstitutesupportedtheresearchdescribedinthispublicationunderawardnumberR24HD041025fromtheNationalInstitutesofHealth. Citation Amato,P.R.(2015).Marriage,cohabitationandmentalhealth.FamilyMatters,96,5-13. Relatedtopics Couples Couples Share