72942033 Como o Casamento Afetam a Saude Psicologica e Fisica

8/3/2019 72942033 Como o Casamento Afetam a Saude Psicologica e Fisica

1/38

Page 1Pgina 1

How Does Marriage Affect Physical and Psychological Health?Como o casamento afetam a sade psicolgica e fsica?A Survey of the Longitudinal EvidenceUm exame da evidncia Longitudinal

Chris M. WilsonChris M. WilsonUniversity of East Anglia UK Universidade de East Anglia no Reino UnidoAndrew J. OswaldAndrew J. OswaldUniversity of Warwick UK and Harvard University USA Universidade de Warwick e do ReinoUnido E.U.A. Universidade de HarvardMay 2005Maio 2005Abstract ResumoThis paper examines an accumulating modern literature on the health benefits of relationships likeEste artigo analisa uma moderna literatura sobre acumulao de benefcios para a sade dasrelaes comomarriage. casamento. Although much remains to be understood about the physiological channels,we draw the Embora ainda haja muito a ser entendido sobre os canais fisiolgico, chamar a

judgment, after looking across many journals and disciplines, that there is persuasive longitudinaljulgamento, depois de olhar em muitas revistas e disciplinas, que no convincente longitudinalevidence for such effects. provas para tais efeitos. The size of the health gain from marriage isremarkable. O tamanho do ganho de sade do casamento extraordinrio. It may be as large asPode ser to grande quanto

the benefit from giving up smoking. o benefcio de deixar de fumar.Key words: mortality, health, marriage, happiness, longitudinalPalavras-chave: mortalidade,sade, casamento, felicidade, longitudinalCorresponding author: tel.Endereo para correspondncia: tel.02476 523510,[email protected] 523510, andrew.oswald @ warwick.ac.uk

Acknowledgments: work on this paper was done while the second author was a Wertheim Fellow

at Harvard UniversityAgradecimentos: o trabalho sobre este documento foi feito enquanto osegundo autor foi um companheiro Wertheim na Universidade de Harvard

Page 2Pgina 22 2How Does Marriage Affect Physical and Psychological Health?Como o casamento afetam asade psicolgica e fsica?A SurveyUm exameof the Longitudinal Evidencedas provas LongitudinalIntroduction IntroduoEconomists know that human beings reap financial benefits from marriage. Os economistas sabemque seres humanos colhem os benefcios financeiros do casamento. Even after controlling forMesmo aps ajuste paraother factors, married individuals earn much more than single people (eg Chun and Lee (2001),Daniel outros fatores, os indivduos casados ganham muito mais do que os solteiros (por exemplo,e Chun Lee (2001), Daniel(1995), Loh (1996), Reed and Harford (1989)). (1995), Loh (1996), Reed e Harford (1989)).There are gains, too, from economies of scale within the H ganhos, tambm, das economias deescala nafamily (Becker 1981). famlia (Becker, 1981).


2/38

Yet economists are less aware of an emerging body of research that finds wider benefits frommarriage -- No entanto, os economistas esto menos conscientes de um organismo de investigaoemergentes, que encontra maiores benefcios do casamento -one that suggests the mind and the body are intertwined in ways not fully understood. que sugereque a mente eo corpo esto entrelaadas de maneira no totalmente compreendida. Sociologists,

Socilogos,psychologists and epidemiologists have recently documented evidence of married people s betterphysical psiclogos e epidemiologistas recentemente evidncia documentada de pessoas casadas melhor fsicohealth, longevity, psychological health, and reported happiness. , A longevidade, psicolgicos, desade e felicidade de sade relatados. Married individuals fare better in these indivduos casadostarifa melhor nestesterms than the never married, who in turn do better than the divorced, separated and widowed.termos do que nunca se casou, que por sua vez, fazer melhor do que os divorciados, separados evivos. Often the Muitas vezes, ocoefficients in such research imply strikingly large consequences from relationships. coeficientes

em tais pesquisas implicam consequncias surpreendentemente grande de relaes. For example,using Por exemplo, usando

pooled cross-sections for the US and the UK, Blanchflower and Oswald (2004a) estimate the sizeof the pooled cross-sections para os E.U. e Reino Unido, Blanchflower e Oswald (2004a) estimamo tamanho damarriage effect upon mental well-being to be equal to that from an extra $100,000 dollars a year.efeito sobre o casamento bem-estar mental para ser igual ao que, a partir de US $ 100.000 dlaresextra por ano. Formal Formalmarriage itself seems to matter. casamento em si parece ter importncia. In the few studies thatcompare married and non-married cohabiters, the Os poucos estudos que comparam e no-casadoscohabiters casado,results typically show an extra beneficial effect from being married. resultados mostramtipicamente um efeito benfico adicional de ser casado. A common conclusion also asserts Aconcluso comum tambm afirmathat these benefits are larger for men than for women; pioneering work here was done by Gove etal que estes benefcios so maiores para homens do que para as mulheres; trabalho pioneiro aquifoi feito por Gove et al(1983). (1983). Finally, some argue that the benefits to happiness may be declining over time,relative to the non- Finalmente, alguns argumentam que os benefcios para a felicidade pode estarem declnio ao longo do tempo, em relao ao no-married (Lee et al 1991, Glenn and Weaver 1988). casado (Lee et al 1991, Glenn e Weaver,

1988).This paper tries to assess these ideas. Este trabalho pretende avaliar estas idias. It is not acomprehensive survey -- the literature is already too large No um estudo abrangente - aliteratura j muito grandefor that -- but it attempts, by looking across diverse literatures, to draw general messages thatseem of para isso - mas ele tenta, olhando atravs de literaturas diversas, para chamar a mensagensgerais que parecem deinterest to economists and other social scientists. interesse para economistas e outros cientistassociais. Past reviews include Ross et al (1990), Burman and Avaliaes anteriores incluem Ross etal (1990), Burman eMargolin (1992) and Kiecolt-Glaser (2001) on physiological evidence, Coombs (1991) on

wellbeing, and Margolin (1992) e Kiecolt-Glaser (2001) sobre a evidncia fisiolgica, Coombs(1991) sobre bem-estar, e


3/38

Waite (1995) on evidence of health, financial and other benefits. Waite (1995) nas provas desade, financeiros e outros benefcios. Waite and Lehrer (2003) compare the Waite e Lehrer(2003) comparam o

benefits from marriage with the benefits associated with being religious. benefcios do casamentocom os benefcios associados com ser religioso.

In contrast to previous reviews, we concentrate on longitudinal evidence. Em contraste comcomentrios anteriores, ns nos concentramos em evidncia longitudinal. The reason is that toeconomists A razo que os economistas(and many other kinds of investigators) this type of evidence is more persuasive than that fromcross- (E muitos outros tipos de pesquisadores), este tipo de provas mais persuasivo do que a decross-sections. sees. In cross-sectional patterns, causality is difficult to unearth. Em padres de seotransversal, a causalidade difcil de descobrir. Although panel data do not always Embora osdados do painel nem sempre

provide a solution, the ability to look at events through time is an advantage. fornecer umasoluo, a capacidade de olhar para os acontecimentos ao longo do tempo uma vantagem. First,

panel data offer Primeiro, os dados do painel ofertaPage 3Pgina 3

3 3before-and-after evidence. antes e depois da prova. Second, panel data make it possible todifference out -- in the econometric Por outro lado, dados em painel fazer diferena para a

possvel sada - no economtricossense -- the underlying characteristics of the individuals being studied. sentido - as caractersticassubjacentes dos indivduos estudados.We start in section I by looking at the possible mechanisms at work behind a correlation betweenhealth Partimos do ponto que eu olhando os mecanismos possveis de trabalho por detrs de uma

correlao entre a sade(mental or physical) and marriage. (Mental ou fsica) e do casamento. The section discussesmethodological problems in this area, and how A seo aborda os problemas metodolgicos nestarea, e comoresearchers have tried to resolve them. pesquisadores tm tentado resolv-los. Section II reviewsthe effects of marriage on mental health. Seo II analisa os efeitos do casamento sobre a sademental. In Emthese research studies, the benefits of marriage are commonly measured by studying the effect ofestes estudos, os benefcios do casamento so comumente medida pela estudar o efeito damatrimony on depression and alcohol abuse. matrimnio em depresso e abuso de lcool. SectionIII examines mortality. Seo III analisa mortalidade. It shows that, after Isso mostra que, aps

controlling for other factors, married people are far less likely to die in any given period than thenon- controle de outros fatores, as pessoas casadas tm muito menos probabilidade de morrer emum determinado perodo do que os no-married. casado. Section IV reviews the correlations between marriage and physical health.Seco IV revisa as correlaes entre o casamento ea sade fsica. Finally, section V Finalmente,a seo Vdiscusses tests of the channels from marriage to wellbeing. discute os testes dos canais decasamento ao bem-estar.Section I: Methodological Issues Seco I: questes metodolgicasHow are well-being and marital status interconnected? Como esto bem-estar e status maritalinterligados? First, marriage may itself lead to improved Em primeiro lugar, o casamento pode-se

levar a melhor


4/38

physical and mental health. mental e sade fsica. This has been called the protective effect ofmarriage. Isto tem sido chamado o efeito protetor do casamento. Alternatively, simple Emalternativa, simplescorrelations could be interpreted as telling us merely that certain types of people can persuade a

partner to correlaes poderia ser interpretado como dizendo apenas que certos tipos de pessoas

podem convencer um parceiromarry them. cas-los. This is the so-called selection effect. Este o chamado efeito de seleo-lo.According to evolutionary principles, it would be no surprise if physically and mentally healthierDe acordo com os princpios evolutivos, no seria surpresa se fsica e mentalmente saudvelindividuals are more attractive to mates, and thus more likely to show up in data sets as married(and to indivduos so mais atraentes para os companheiros e, portanto, mais susceptveis deaparecer em conjuntos de dados como casados (estay married). permanecer casado). Hence better health could be a cause of marriage. Assim,melhor a sade pode ser uma causa do casamento. If panels are not sufficiently long -- Se os

painis no esto suficientemente longo -stretching back before the date of marriage -- it may be impossible to correct fully for selection.

remonta antes da data do casamento - que pode ser impossvel corrigir integralmente para aseleo. In this Nestaresearch area, Lillard and Panis (1996) and Brockmann and Klein (2004) contain valuablediscussions of rea de pesquisa, Lillard e Panis (1996) e Brockmann e Klein (2004), contmvaliosas discusses dehow to control for selection. como controle para a seleo. Cohen (2004) considers the possiblechannels from marriage to health; Ren Cohen (2004) considera que as vias possveis de casamento

para a sade; Ren(1997) argues that marital quality and marital history should be held constant in testing; theintriguing (1997) argumenta que a qualidade conjugal e histria conjugal deve ser mantidoconstante nos ensaios, o intriganteidea of using twins data is explored in Kohler et al (2004). idia de usar dados gmeos exploradoem Kohler et al (2004).Selection effects need not work in a simple way. efeitos de seleco no precisa trabalhar de umaforma simples. Joung et al (1998) explain with reference to Collins and Joung et al (1998)explicam com referncia Collins eColtrane (1992) that, if assortative mating occurs, unhealthy people with unhealthy partners aremore Coltrane (1992) que, se acasalamento ocorre, as pessoas saudveis com parceiros so maissaudveislikely to be widowed. susceptvel de ser viva. This could lead to a spurious relationshipsuggesting that widowhood causes ill Isto poderia levar a um relacionamento esprio, sugerindo

que a viuvez causas malhealth. de sade. Alternatively, Lillard and Panis (1996) suggest that if marriage does givebeneficial effects, then Alternativamente, Lillard e Panis (1996) sugerem que, se o casamento nodar efeitos benficos, em seguida,

physically and mentally unhealthy individuals will face the greatest incentive to marry, giving an ementalmente saudvel indivduos fisicamente ir enfrentar o maior incentivo para se casar, darumaadverse selection effect. efeito de seleo adversa. As we shall see, the existence of adverseselection effects rather than the Como veremos, a existncia de efeitos seleo adversa e no a

Page 4Pgina 44 4


5/38

standard selection effect is supported by the evidence, but only for males. efeito de seleo padro apoiada pelas evidncias, mas apenas para os homens. The mechanisms leading to Osmecanismos conducentes

possible benefits from marriage have been explored. possveis benefcios do casamento foramexploradas. In section V, we look more closely at the research Na seo V, olhamos mais de perto

a investigaoevidence. provas. Here we review some suggested theoretical explanations. Aqui ns revemosalguns sugeriram explicaes tericas. Further discussion is available in Uma discusso maisaprofundada est disponvel emRoss et al (1990) and Wyke and Ford (1992). Ross et al (1990) e Wyke e Ford (1992).On the grounds that two can live almost as cheaply as one, marriage may work simply because it

provides Com o pretexto de que os dois podem viver quase to barato quanto um, o casamentopode funcionar simplesmente porque fornecehigher real income per partner. maior renda real per parceiro. Poorer standards of living arecorrelated with mental health problems padres mais baixos de vida esto correlacionados com

problemas de sade mental

(Ross et al 1990). (Ross et al 1990). Ross (1995) also shows that the married have the lowestincidence of economic Ross (1995) tambm mostra que os casados tm a menor incidncia deeconomiahardship, while Smock et al (1999), who study switchers , find clear financial benefits for marrieddificuldades, enquanto Smock et al (1999), que switchers estudo, encontrar claros benefciosfinanceiros para casarwomen. mulheres. Second, marriage is a source of emotional and instrumental support. Emsegundo lugar, o casamento uma fonte de apoio emocional e instrumental. Emotional help seemsto ajuda emocional parecereduce the incidence of depression and mental illness (Ross et al 1990), and may provide animportant reduzir a incidncia de depresso e doena mental (Ross et al, 1990), e pode fornecerum importante

buffer against stress (Kessler and Essex 1982, Berkman 1988). tampo contra o estresse (Kessler eEssex 1982, Berkman, 1988). Marriage can also enhance feelings of O casamento tambm podereforar sentimentos deattachment and belonging, which are thought to affect mental health (House et al 1988), whilereleasing penhora e de pertena, que so pensados para afetar a sade mental (House et al, 1988),enquanto libera

people from any possible social stigma associated with being unmarried. pessoas de qualquerestigma social associado com possvel ser solteiro. Single people are more likely to Os solteirosso mais propensos a

live alone, which is known to be associated with depression (Ross et al 1990). viver sozinho, que conhecido por estar associado com depresso (Ross et al 1990). Emotional help can, emocionalpode ajudar,nevertheless, come from sources outside marriage. no entanto, vm de fontes fora do casamento.Social networks thus may be important, especially for As redes sociais, portanto, pode serimportante, especialmente parathe unmarried. os solteiros. A common conjecture about why the benefits from marriage are largerfor males than A conjectura comuns sobre os benefcios do casamento so maiores para o sexomasculino do quefemales is that women have more developed social networks outside marriage, and these alreadygive fmeas que as mulheres tm mais desenvolvidas redes sociais fora do casamento, e esses j

do


6/38

valuable support (Schumaker and Hill 1991). valioso apoio (Schumaker e Hill, 1991). A thirdmechanism is the so-called guardian role. Um terceiro mecanismo o chamado guardio papel-lo.Married Casadoindividuals act differently from single people. os indivduos agem de forma diferente das pessoasindividuais. They tend to engage less in risky activities and more in Eles tendem a se engajar em

atividades menos arriscadas e maishealthy ones - perhaps for the sake of their partner (Umberson 1987, Ross et al 1990, Power et al1999). sadias - talvez por causa do seu parceiro (Umberson 1987, Ross et al 1990, Power et al1999).For example, married people smoke and drink less. Por exemplo, as pessoas casadas fumar e

beber menos. Moreover, partners probably unconsciously monitor Alm disso, os parceirosprovavelmente inconscientemente monitoreach other for early symptoms of illness. uns dos outros para os primeiros sintomas da doena.Early cross-sectional evidence could not distinguish between selection and protection effects.Early-seccional prova cross no pode distinguir entre a seleo e os efeitos de proteco. One lineof Uma linha de

attack since then has been to try econometrically to estimate the size of the selection effect. ataquedesde ento tem sido tentar estimar econometricamente o tamanho do efeito de seleo. In theory,Em teoria,this can be done by using pre-marital characteristics to predict later marital status (eg Joung et al1998, Isso pode ser feito usando-marital caractersticas pr prever mais tarde, estado marital (egJoung et al 1998,Masterkaasa 1992), and then, after subtracting the selection effect, viewing any correlation

between Masterkaasa 1992) e, em seguida, depois de subtrair o efeito de seleo, visualizao dequalquer correlao entre ahealth and marriage that remains as the true size of the protection from being married. sade eunio que permanece como a verdadeira dimenso da proteco de ser casado. Longitudinal dataOs dados longitudinaisshould allow this. deveria permitir isso.In calculating the effects on mental health, for instance, the following approach is common. Noclculo dos efeitos sobre a sade mental, por exemplo, a seguinte abordagem comum. Measuresare As medidas soobtained of mental health at time T1, before marriage, and at time T2, when marital status haschanged obtidos da sade mental no tempo T1, antes do casamento, e em T2 tempo, quando oestado civil mudoufor some of the individuals in the sample. para alguns dos indivduos da amostra. OLS regressionsare then performed to explain mental health at regresses OLS so ento realizados para explicar a

sade mental na Page 5Pgina 55 5time T2, while including both a set of explanatory variables and the mental health measure at timeT1. T2 tempo, enquanto incluindo um conjunto de variveis explicativas e as medidas de sademental no tempo T1.The inclusion of mental health at T1 acts as a simple control for pre-marital psychological well-

being. A incluso da sade mental no T1 atua como um controle simples para pr-marital bem-estar psicolgico. In Em

principle, any relationship then found between mental health and marriage should not be aspurious princpio, qualquer relao, em seguida, encontrado entre a sade mental eo casamento

no deve ser um falso


7/38

selection effect. efeito de seleo. Most evidence of this sort suggests that a protective effectexists. A maioria das evidncias desse tipo sugere que existe um efeito protetor.These regressions must be specified, however, with all the potential explanatory variables thataffect Estas regresses deve ser especificado, no entanto, com todas as variveis explicativas

potenciais que afetam

mental health at T2. sade mental em T2. Often-included variables are measures of socialintegration outside marriage, Muitas vezes, as variveis foram includas medidas de integraosocial fora do casamento,employment status, income and frequency of financial problems, age, gender, race, and parentalstatus. situao de emprego, renda e freqncia de problemas financeiros, idade, sexo, raa estatus parental.Unfortunately, only some of these variables are typically included in any one study, and theinfluence of Infelizmente, apenas algumas dessas variveis so normalmente includos emqualquer estudo de um lado, ea influncia da

physical health is often neglected. sade fsica muitas vezes negligenciada. This is a potentiallyserious omission, because health is an important Esta uma omisso grave, potencialmente,

porque a sade um importantedeterminant of happiness (Stack and Eshleman 1998). determinantes da felicidade (Stack eEshleman 1998). If health is positively correlated with the likelihood Se a sade est

positivamente correlacionada com a probabilidadeof marriage, that will produce an upward bias on the marriage coefficient, and lead to anoverstatement of do casamento, que ir produzir um vis ascendente sobre o coeficiente decasamento, e levar a um exagero dethe benefits of marriage. os benefcios do casamento.Section II: Mental Health Benefits from Marriage Seco II: Sade Mental Benefcios doCasamentoWe start with the mind. Comeamos com a mente.Gove et al (1983) is an early cross-section study. Gove et al (1983) um estudo transversal cedo.This uses US data, and finds marriage to be the best Este utiliza dados E.U., e encontra ocasamento para ser o melhor

predictor of happiness after controls are added for education, age, gender and race. preditor dafelicidade aps os controles so adicionados para a educao, idade, gnero e raa. The authorstake steps Os autores tomar as medidasto try to counter the selection problem by including a bad childhood experiences variable, whichthey para tentar combater o problema da seleo, incluindo uma varivel experincias da infnciaruim, quethink will predict pre-marital wellbeing. acha que vai prever o bem-estar pr-marital. The paper

finds the beneficial effects of marriage to be larger for O documento considera os efeitosbenficos do casamento deve ser maior paramen than for women. homens do que para as mulheres. Also, using the married sample only,marriage quality is the single best predictor of Alm disso, usando o exemplo casado s, aqualidade do casamento o melhor preditor dewellbeing. bem-estar.More recent cross-section evidence, by Stack and Eshleman (1998), considers 17 developednations. Mais provas cross-section recente, e Eshleman Stack (1998), considera 17 naesdesenvolvidas.Controls are included for gender, age, health, financial situation, children, education, religion,national Os controles so includos para sexo, idade, sade, situao financeira, crianas,

educao, religio, nacional


8/38

marriage and divorce rates, GDP/capita, and income distribution measures. casamento e as taxasde divrcio, o PIB per capita, e as medidas de distribuio de renda. This research finds financialEsta pesquisa encontra financeirasituation to be the best predictor of happiness -- followed by health, followed by marital status.situao de ser o melhor preditor da felicidade - seguido por sade, seguido pelo status marital.

The Omarried turn out to be happier than those who cohabit, who are themselves happier than singlecasou-se com revelar-se mais felizes do que aqueles que convivem, que so eles prprios maisfelizes do que simplesindividuals, ceteris paribus. indivduos, ceteris paribus. In regressions on each nation separately,for sixteen of seventeen nations the Em regresses de cada nao separadamente, dezesseis paradezessete pases damarriage coefficient is positive in a wellbeing equation. casamento coeficiente positivo em umaequao bem-estar. The existence of this cross-section pattern is now A existncia deste padrocross-section agoraquite firmly established. muito firmemente estabelecida. Similar patterns are reported, for

instance, for European countries in modern Padres similares so relatados, por exemplo, para ospases europeus nas modernasdata in Di Tella, MacCulloch and Oswald (2003). dados em Di Tella, MacCulloch e Oswald(2003).

Page 6Pgina 66 6Could this be an illusion within the data? Isto poderia ser uma iluso dentro dos dados? Johnsonand Wu (2002) argue that there are three potential Johnson e Wu (2002) argumentam que existemtrs potenciaisreasons why married people might incorrectly appear to have better psychological health. razes

pelas quais as pessoas casadas podem incorretamente parecem ter melhor sade psicolgica. First,married Primeiro, casou-se compeople were actually happier before they married; this is the selection effect. as pessoas estavamrealmente felizes antes de se casarem, este o efeito de seleo. Second, in a cross-section Emsegundo lugar, em uma seco transversalof individuals, some of the married people have only recently been married, and this initiallymakes them das pessoas, algumas das pessoas casadas s recentemente foram casados, e isso f-los inicialmentehappy, but does not do so in the long-run, because they adapt back to baseline . feliz, mas nofaz-lo no longo prazo, porque elas se adaptam de volta base. Third, marriage and Em terceirolugar, o casamento ea

divorce are correlated with important kinds of social roles and networks, and it may be these thatadd to or divrcio so correlacionados com importantes tipos de papis sociais e redes, e pode serque esses adicionar oureduce well-being; this is a kind of omitted-variable argument. reduzir o bem-estar, que umaespcie de argumento de varivel omitida.Several papers have used longitudinal data to estimate the benefits of marriage for mental health,and to Diversos trabalhos tm utilizado dados longitudinais para estimar os benefcios docasamento para a sade mental, eadjust for the selection problem. ajustar para o problema da seleo. Horwitz et al (1996) findsignificant benefits from marriage -- both Horwitz et al (1996) constatam benefcios significativosdo casamento - tanto


9/38

reductions in depression for married women, and reductions in alcohol abuse for married men.redues na depresso para as mulheres casadas, e as redues no consumo abusivo de lcool parahomens casados. Simon Simonand Marcussen (1999) demonstrate that marriage cuts the probability of depression. e Marcussen(1999) demonstram que o casamento corta a probabilidade de depresso. In Horwitz et al Em

Horwitz et al(1996), psychological health gains are measured by using a standard 10-item depression index,and (1996), os ganhos de sade psicolgica so medidos atravs de um item, depresso ndice 10standard, ealcohol abuse by a similar index measuring the frequency of various drinking problems. abuso delcool por um ndice similar medir a frequncia de vrios problemas com a bebida. This data set isEste conjunto de dados the Rutgers Health and Human Development Project in the US: it follows approximately 1400subjects, Sade da Rutgers e Desenvolvimento Humano Projeto de os E.U.: segue-se cerca de1400 indivduos,aged 12,15 and 18 years old, for seven years from sixteen of the twenty-one New Jersey counties.

idade 12,15 e 18 anos, durante sete anos a partir de dezesseis dos 21 condados de Nova Jersey.The Osubjects are tested at four points across the seven years. os sujeitos so testados em quatro pontosao longo dos sete anos. Seven years may be too short a follow-up period Sete anos pode ser muitocurto um perodo de follow-upto be ideal. de ser ideal. A response rate of less than 50% is obtained. A taxa de resposta de menosde 50% obtido. Only the respondents who never married over Apenas os entrevistados quenunca se casou maisthe whole period (482), and the respondents who were consistently married between the 3 todo o

perodo (482) e os entrevistados que eram casados consistentemente entre os trsrd rdand 4 e 4th th

period perodo(347), are used. (347), so utilizados. Rather interestingly, the research finds that depression andalcohol abuse fall over the interessante, a pesquisa conclui que a depresso e cair sobre o abusode lcoolseven years for all groups, but it declines most steeply among the married sample. sete anos paratodos os grupos, mas os declnios mais forte entre a amostra casados. Independent variables Asvariveis independentesinclude age, gender, individual income, and a 22-item index of social integration; the last of these

is used incluem a idade, sexo individual de renda, e um ndice de 22 itens de integrao social, altima delas utilizadato account for the apparently anti-depressive effects of having extensive networks of friends. paraexplicar o efeito anti-depressivo, aparentemente, de ter uma extensa rede de amigos.The data reveal a large and significant marital effect upon depression by the 4 Os dados revelamuma importante e conjugal grande efeito sobre a depresso, a 4th th

period. perodo. Higher incomes O aumento dos rendimentosand greater levels of social support also significantly reduce the probability of depression, whileage and e maiores nveis de apoio social tambm reduzir significativamente a probabilidade dedepresso, enquanto a idade e


10/38

gender have no statistically significant effect. gnero tm qualquer efeito estatisticamentesignificativo. Even after the controls are added, being married produces a Mesmo depois que oscontroles so adicionados, sendo casado produz umsignificantly reduced score in the depression index (p


11/38

What people think of marriage also turns out to matter. O que as pessoas pensam do casamentotambm despeja matria. Simon and Marcussen (1999) add to the literature Simon e Marcussen(1999) acrescentam literatura

by showing that beliefs about marriage are correlated with the mental-wellbeing benefits gainedfrom , mostrando que as crenas sobre o casamento so correlacionados com o bem-estar,

benefcios mental adquiridabeing married. ser casado. The authors demonstrate that individuals who value the permanenceand importance of Os autores demonstram que os indivduos que valorizam a permanncia eaimportncia domarriage have a larger reduction in depression, and suffer more from marriage dissolution.casamento tm uma maior reduo da depresso e sofrem mais com a dissoluo do casamento.Panel data dados em painelfrom the US National Survey of Families and Households (1987-88 and 1992-1994) are studied.de os E.U. Pesquisa Nacional por Amostra de famlias e domiclios (1987-1988 e 1992-1994) soestudados. Knowing Conhecimentoindividuals beliefs and initial status at 1987/8 (T1) allows an examination of the effect of beliefs

on indivduos crenas e estado inicial em 1987 / 8 (T1) permite uma anlise do efeito das crenassobremarital status. estado civil. In total, 10,005 adults are interviewed, with 74% and 82% responserates. No total, 10.005 so adultos entrevistados, com 74% e 82% as taxas de resposta. Attrition isa Atrito uma

problem, especially among those unmarried at T1, so the results here may have to be evaluatedwith care. problema, principalmente entre os solteiros em T1, por isso os resultados aquiapresentados podem ter que ser avaliado com cuidado.A standard depression scale is employed by the authors -- based on responses to questions on thenumber A escala de depresso padro utilizado pelos autores - com base em respostas a

perguntas sobre o nmeroof days they felt lonely, not having an appetite, etc. Two dummies are constructed to show maritalgain de dias que se sentia sozinho, no ter apetite, etc Dois bonecos so construdos para mostrarganho maritaland loss in the time intervals. e perda de intervalos de tempo. Marital beliefs are measured byquestions about the perceived importance crenas conjugal so medidos por meio de perguntassobre a importncia percebidaand desirability of marriage. ea convenincia de casamento. Control variables include age, yearsof education, gender, race, household As variveis de controle incluem a idade, anos deescolaridade, sexo, raa, famliaincome and parental status. renda e status parental. Noticeably, measures of social integration and

health are missing. Visivelmente, as medidas de integrao social e sade esto em falta.Large effects are found. grandes efeitos so encontrados. Depression at T1 is significantly higheramong the individuals about to Depresso em T1 significativamente maior entre os indivduossobre aexperience a marital loss. experimentar uma perda marital. Selection effects are not supported,however, because depression at T1 is not efeitos de seleco no so suportados, no entanto,

porque a depresso no est em T1significantly different for those about to experience a marital gain, or for the never married.significativamente diferente para aqueles prestes a experimentar um ganho civil, ou para oscasados nunca. Those who Aqueles quefeel strongly about marriage experience the largest gains from marrying. sentir fortemente sobre a

experincia do casamento os maiores ganhos de se casar.Page 8Pgina 8


12/38

8 8The paper then investigates the effect of changes in marital status upon depression. A seguir,investiga os efeitos das mudanas no estado civil em depresso. It uses T1 depression Ele usa T1depressomeasures to control for selection effects. medidas de controlo para efeitos de seleco. Older,

more educated, higher income, employed individuals are Mais velhos, mais instrudos, de maiorrenda, os indivduos empregados sosignificantly less depressed. significativamente menos deprimidas. Parenthood has no significanteffect, which is consistent with results A maternidade tem um efeito significativo, o que consistente com os resultadosreviewed in McLananhan and Adams (1987) and Ross et al (1990). revista McLananhan e Adams(1987) e Ross et al (1990). Initial depression is a strong predictor depresso inicial um forte

preditorof depression at T2, as would be expected. de depresso em T2, como seria de esperar. Even withother factors held constant, marital break-up has a Mesmo com outros fatores mantidos constantes,estado civil break-up tem um

large depressive effect (p


13/38

in mental well-being. no bem-estar mental.Striking results on the power of marriage are provided by Marks and Lambert (1998), who use atwo- resultados interessantes sobre o poder da unio so fornecidos pela Marks e Lambert (1998),que usam um de doiswave panel from the US National Survey of Families and Households (for the periods (1987-88)

onda painel do National Survey E.U. de famlias e domiclios (para os perodos (1987-88)and(1992-93)). e (1992-1993)). The authors analyse the effects of marital history on well-beingand have a particular Os autores analisam os efeitos da histria conjugal no bem-estar e ter um

particularinterest in whether effects differ by age and gender. interesse em saber se os efeitos diferem poridade e sexo. Marital history is captured by a comprehensive set of histria conjugal capturado

por um conjunto abrangente deten marital states. dez estados conjugais. Well-being is measured with a large set of measures thatrange from depression, Bem-estar medido com um grande conjunto de medidas que vo desde adepresso,happiness, and self esteem, to irritability and purpose in life. felicidade e auto-estima,

irritabilidade e propsito na vida. An individual s well-being in the later period Um indivduo obem-estar no perodo posterioris regressed on marital history, controlling for the individuals previous well-being values in 1987-88, and regredida sobre a histria conjugal, controlando os indivduos anteriores, sendo osvalores assim em 1987-88, efor other personal characteristics. para outras caractersticas pessoais. Compared to thosecontinually married, the continually Comparado aos continuamente casado, continuamenteseparated/divorced show significantly lower levels of mental health. separados / divorciadosmostram nveis significativamente mais baixos de sade mental. Similarly, the widowed have Domesmo modo, a viva tersignificantly increased chances of depression; the continually single are less happy and moredepressed, aumentou significativamente as chances de depresso, o nico continuamente somenos felizes e mais deprimidas,

but do exhibit higher autonomy and personal growth. mas apresentam maior autonomia do ecrescimento pessoal. Transition to marriage from being never-married Transio para o casamentode ser que nunca se casaramstrongly increases almost all measures of psychological wellbeing, and remarriage also works butoffers fortemente aumenta quase todas as medidas de bem-estar psicolgico, e recasamentotambm funciona, mas oferecereduced benefits. benefcios reduzidos. A transition from marriage to separation or widowhoodincreases depression and A transio do casamento separao ou viuvez aumenta depresso e

Page 9Pgina 99 9unhappiness. infelicidade. In contrast to some evidence, this effect is most pronounced for women.Em contraste com algumas evidncias, este efeito mais pronunciado para as mulheres. Notably,adverse Notavelmente, adversaseffects seem to be reliably tempered by middle age. efeitos parecem ser confivel temperada pelameia-idade.With the same data, Kim and McKenry (2002) address similar questions but look at cohabitation.Com os mesmos dados, Kim e McKenry (2002) abordam questes semelhantes, mas olhar para acoabitao. The Oauthors are thorough. autores so completas. They combine extensive controls for the quality of

the relationship with survey Eles combinam controles abrangentes para a qualidade dorelacionamento com o levantamento


14/38

responses on fairness, disagreements, satisfaction and conflict (in addition to the usual otherrespostas sobre a justia, os desacordos, satisfao e conflito (para alm dos habituais,characteristics, and initial well-being). caractersticas iniciais e bem-estar). As expected, thequality variables help to explain individuals Como esperado, as variveis de qualidade ajuda aexplicar os indivduos

well-being, but the pure marital effect is still significant in the data. bem-estar, mas o casamentoainda puro efeito significativo nos dados. Intriguingly, unlike marriage, Curiosamente, aocontrrio do casamento,cohabiting here produces few benefits. coabitao aqui produz poucos benefcios. This study alsofinds against strong selection effects. Este estudo tambm verifica em relao aos efeitos deseleco forte. Initial Inicialdepression has negligible consequences for observed marital transitions. depresso temconsequncias insignificantes para observar transies conjugais. This evidence reinforces theEsta evidncia refora ameasured protection effects from marital transitions, but leaves the comparisons betweencontinual medidos os efeitos de transies defesa civil, mas deixa as comparaes entre contnua

marital states still open to the selection explanation. estado civil ainda em aberto para a explicaode seleo.These NSFH data are also used in Simon (2002). Estes dados NSFH tambm so usados emSimon (2002). Compared to the continuously married, the author Comparado com o contnuocasado, o autordiscovers that a loss from separation or divorce is associated with an increase in depression andalcohol descobre que uma perda de separao ou divrcio est associado com um aumento nadepresso e lcoolabuse, and these effects are larger for women than men (as also observed by Marks and Lambert(1998)). abuso, e esses efeitos so maiores para mulheres do que homens (como observadotambm por Marcas e Lambert (1998)).Yet there are no statistically significant effects from widowhood. Ainda no h efeitosestatisticamente significativos da viuvez. Marital gains led to benefits in both ganhos conjugaltrazido benefcios tantowellbeing measures -- equally for men and women -- but these only seem to exist for the firstmarriage. medidas bem-estar - igualmente para homens e mulheres -, mas estes parecem existirapenas para o primeiro casamento.There are a number of differences between marriages and non-married partnerships. H uma sriede diferenas entre casamentos e parcerias no-casados. Obviously the level Obviamente, o nvelof commitment may be different, but more subtle differences are also present. de compromisso

podem ser diferentes, mas diferenas mais sutis tambm esto presentes. Cohabiters seem more

Cohabiters parecem maislikely to have lower quality relationships (Brown and Booth 1996) and unstable ones (Brown2000), and probabilidade de ter relaes de qualidade inferior (Brown e Booth 1996) e instvel(Brown 2000), emore often have lower socio-economic status (Rindfuss and VandenHeuvel 1990). maisfreqentemente tm menor status scio-econmico (Rindfuss VandenHeuvel e 1990). Theunmarried living- Os solteiros que vivem,together group is particularly of interest because in Western society its size is growing. juntamentegrupo particularmente interessante porque, na sociedade ocidental, seu tamanho decrescimento. Presumably many Presumivelmente, muitoscohabiters are testing their relationship to assess its long-term suitability. cohabiters est testando

sua relao com a avaliar a sua adequao a longo prazo. Within the US, more than half Dentro osE.U., mais da metade


15/38

of marriages are preceded by cohabitation (Bumpass and Sweet 1989). dos casamentos soprecedidos por coabitao (Bumpass e Sweet, 1989). In 1970, half a million cohabiters Em 1970,meio milho de cohabitersexisted; today, the figure exceeds four million. existia, hoje, o valor superior a quatro milhes.Looking across the literature, is cohabitation as good as having the band of gold itself? Olhando

atravs da literatura, to bom como a coabitao com o anel de ouro em si? It seems not. Pareceque no.Two papers find less depression among married people than among cohabiters. Dois papisencontrar menos depresso entre pessoas casadas do que entre cohabiters. Brown (2000) arguesthat Brown (2000) argumenta quethis is likely to result from the extra stability of a marriage. provvel que o resultado de maiorestabilidade de um casamento. Horwitz and White (1998) establish that Horwitz e White (1998)estabelecem quesingle people are similar to cohabiters in having more depression than the married. pessoassolteiras so semelhantes aos cohabiters em ter mais depresso do que os casados. Cohabiters areshown Cohabiters so mostrados

to be the group with the highest alcohol abuse. ser o grupo com o abuso de lcool mais elevada.Brown (2000) investigates depression with panel data Brown (2000) investiga a depresso comdados em painelfrom the US National Survey of Families and Households (1987-88 and 1992-1994). de os E.U.Pesquisa Nacional por Amostra de famlias e domiclios (1987-1988 e 1992-1994). This includes646 Isso inclui 646

Page 10Pgina 1010 10cohabiters and 3086 married respondents. cohabiters e 3.086 entrevistados casados. The data setallows the entire sample to be single at the first O conjunto de dados permite que toda a amostra a

ser nica no primeiropoint of measurement (to control for initial characteristics). ponto de medio (para controlar ascaractersticas iniciais). To assess depression, Brown draws upon the Para avaliar a depresso,Brown baseia-se nasame scale as Simon and Marcussen (1999), and shows that the average cohabiter score is 16.3,against an mesma escala e Marcussen (1999), Simon e mostra que o escore cohabiter mdia de16,3, contra umaverage married person s score of 13.0. pessoa casada s pontuao mdia de 13,0. Two variablesfor socio-economic status are used -- years of Duas variveis de status scio-econmico sousados - anos deeducation and total couple s earnings. educao e da s receitas Casal total. Cohabiters average

years of education are 12.4, compared to 13.2 mdia de anos de educao so Cohabiters 12,4,comparado a 13,2for the married (significantly different at


16/38

these are used. estes so utilizados. Interestingly, respondents provide information, on a scale of 1-5, on their perception of the Curiosamente, os inquiridos fornecer informaes, em uma escala de1-5, sobre a sua percepo dostability of the relationship. estabilidade da relao. Cohabiters record significantly (


17/38

Although marital status itself seems to be causing the change, this is checked more fully with aHeckman- Apesar de o estado civil em si, parece estar causando a mudana, isso verificado deforma mais completa com um Heckman-correction estimator. estimador de correo. An underlying equation for the decision to marry isestimated, and this is used to Uma equao bsica para a deciso de casar estimado, e isso

usado paramodel depression in Period 2, while including the explanatory variables and Period-1 depressionscores. modelo de depresso no Perodo 2, enquanto incluindo as variveis explicativas e 1 Perodo, escores de depresso.Cohabitation emerges with a 2.8-point effect upon depression (p


18/38

married. casado. To test this, a Cox regression is used to estimate the duration before marriage; aprobability Para testar isso, uma regresso de Cox usado para estimar a durao antes docasamento, com uma probabilidadedensity function is estimated with maximum likelihood methods, in terms of time and explanatoryfuno de densidade estimada com os mtodos de mxima verossimilhana, em termos de tempo

e explicativosvariables. variveis. The duration density function is transformed into a hazard function,representing the probability A funo densidade de durao transformada em uma funo derisco, o que representa a probabilidadeof transition, which is straightforward to interpret. de transio, que simples de interpretar. Theeffects of the explanatory variables on the Os efeitos das variveis explicativas sobre a

probability of transition can then be calculated. probabilidade de transio pode ser calculada.In a different tradition altogether, Kohler et al (2004) is an outstanding paper. Em uma tradiocompletamente diferente, Kohler et al (2004) um papel proeminente. These authors study the Osautores do estudorelationship between subjective well-being and partner and fertility status. relao entre bem-estar

subjetivo e scio e estado de fertilidade. The novelty and attraction of A novidade e atraco dethis study is that it uses variations within pairs of twins in order to control for any unobservablegenetic or deste estudo que ele usa as variaes dentro pares de gmeos, a fim de controlarqualquer gentico ou no observveischildhood factors that may influence well-being and status. fatores da infncia que podeminfluenciar o bem-estar e status. This can provide a way to control for Isso pode fornecer umamaneira para controlarselection effects and can ensure a more direct test for causality from marital status to well-being.efeitos de seleco e pode garantir um controlo mais directo para a causalidade do estado civil do

bem-estar. To fully Plenamentecontrol for the selection effect, individual panel data sets ideally need measurements onindividuals well- controle para o efeito de seleo, os dados conjuntos idealmente painelindividual necessita de medies de indivduos bem

being before they get married, which is usually hard to obtain, whereas variation in twinsoutcomes is a sendo antes de se casar, que normalmente difcil de obter, ao passo que a variaodos resultados em gmeos umdifferent way to proxy for this. maneira diferente de proxy para isso. The data comes from theDanish Twin Registry, in 2002, which surveys Os dados vem do dinamarqus Twin Registry, em2002, que pesquisasnon-identical and identical sets of twins. Idnticos e no idnticos conjuntos de gmeos. The dataare split into age groups (25-45 and 50-70) to test for Os dados so divididos em grupos etrios

(25-45 e 50-70) para testarany difference in effects. qualquer diferena nos efeitos. The results show significant positiveeffects upon well-being for individuals in Os resultados mostraram efeitos positivos significativossobre o bem-estar para os indivduos emcurrent cohabitation or marriage relationships relative to being single (or, more precisely, to thecoabitao ou casamento relacionamentos atuais em relao a ser nico (ou, mais precisamente, aocorresponding twin who is single). twin correspondente que nico). This effect is somewhatlarger for males. Este efeito um pouco maior para os homens. Perhaps surprisingly, past Talvezsurpreendentemente, passadoseparations appear to have statistically insignificant consequences for subjective well-being;having a separaes parecem ter consequncias estatisticamente insignificante de bem-estar

subjetivo, tendo umPage 12Pgina 12


19/38

12 12current partner is what matters. parceiro atual o que importa. By contrasting their within-twinestimates to standard Ordinary Least Contrastando suas twin-estima dentro de padres mnimosordinriosSquares estimates, the authors show that OLS tends to overestimate the gains for women and

estimativas Squares, os autores mostram que OLS tende a superestimar os ganhos para asmulheres eunderestimate the effects for men. subestimar os efeitos para os homens. Nevertheless, the size ofthe estimated beneficial effect on wellbeing No entanto, o tamanho do efeito benfico estima no

bem-estarfrom marriage is fairly well captured by OLS. do casamento muito bem capturado por OLS.That is encouraging, in a sense, for the previous literature. Isso encorajador, em certo sentido,

para a literatura.While Ordinary Least Squares estimates are biased, the size of the bias, according to the twins-Enquanto Quadrados Ordinrios estimativas so menos inclinados, o tamanho do preconceito, deacordo com os gmeos,

methodology, appears to be no more than approximately 10%. metodologia, parece ser no maisdo que cerca de 10%. Even so, for women, unobservable Mesmo assim, para as mulheres, noobservveischaracteristics that increase well-being are correlated with the chance of being married confirmingthat caractersticas que aumentam o bem-estar esto correlacionados com a chance de ser casado,confirmando quethere is a selection effect. h um efeito de seleo. However, although the authors do not discuss itas such, this finding is No entanto, embora os autores no discuti-la como tal, este resultado consistent with evidence of adverse selection effects among men. consistente com a evidncia dosefeitos da seleo adversa entre os homens. Here, men with lower well-being are Aqui, os homenscom menor bem-estar soMoney is a useful metric. O dinheiro uma mtrica til. Clark and Oswald (2002) suggest a way,in panel data, to put (large) dollar Clark e Oswald (2002) sugerem uma forma de dados em painel,

para colocar (grande) do dlarvalues on life events like marriage. valores em eventos da vida como o casamento. By using a 17-wave German panel with 15,000 individuals, Stutzer Usando uma alem onda painel 17 com15.000 indivduos, Stutzerand Frey, in a forthcoming paper, go further and analyse the two-way links between subjectivewell-being e Frey, em um artigo futuro, ir mais longe e analisar as ligaes entre os dois sentidos

bem-estar subjetivoand marital status. e estado civil. After controlling for initial status and other characteristics, the

authors find that the Depois de controlar para o estado inicial e outras caractersticas, os autoresconcluem que agains from marriage increase from lowest to highest across the following four marital states:separated ganhos de casamento aumento menor para o maior entre os seguintes quatro estadosconjugais: separadoswith no partner, single, unmarried with partner, married. They provide no direct account for theselectioneffect but do confirm its existence. Instead, the authors demonstrate that the life satisfaction ofsingleindividuals who will later remain single is significantly lower than the life satisfaction of singleindividuals who will marry 4 years hence. Perhaps curiously, the effect is most pronounced for

people intheir twenties and for those who marry later in life, with a smaller effect at around thirty years old.


20/38

Although of a different type, the evidence of Blanchflower and Oswald (2004b), based on 14,000American adults, is consistent with the hypothesis that marriage makes people happier. The studysuffersfrom the limitations of pooled cross-sectional data, but it provides unusual evidence for theinclusion of a

sexual activity variable in empirical analyses of happiness, and finds that married people havemuch moresex than other groups. The authors demonstrate that sexual activity is strongly and monotonicallycorrelated with happiness. So more sex may be one reason why marriage raises psychologicalwellbeing.However, being single or divorced continues in this data set to have an adverse impact onhappiness overand above the effects of reduced sexual activity.The quality of marriage itself is likely to come increasingly under the scrutiny of researchers. ForParainstance, Gardner and Oswald s (2005) results on British panel data reveal that, after some time,

couplesdo recover and benefit from marital dissolution, and that on average it is those with low mentalwell-beingscores who divorce. Finally, modern research is turning, more broadly, to the idea that adaptationmay

Page 13Pgina 1313 13wipe out some of the psychological benefits from marriage and other life events. Lucas et al(2003) andClark et al (2003) use German longitudinal data on 24,000 individuals to explore this, and do find

someevidence of adaptation towards the set point or baseline. More research is needed of this importantkind.In general, the papers reviewed here are consistent with the idea that marriage improves

psychologicalhealth. de sade. This is true even when comparing married couples to those who cohabit, andafter controlling forselection effects. Moreover, marriage has large effects not small ones.Section III: Longevity Benefits from MarriageWe turn now to the body.There is remarkable evidence that marriage helps to keep human beings alive. People who are

divorcedor separated or widowed are at a particularly high risk of dying prematurely. Those never marriedfacesomewhat lower risks of death in any given period, but the married have easily the lowest risk ofall thegroups. grupos. Evidence for this is now widespread across the world. For example, in data fromBangladesh,Rahman (1993) shows that married individuals have significantly lower mortality rates than thenevermarried or divorced, while Mete (2005) finds evidence that being married increases the chance of

being

alive seven years later by 5% in longitudinal data from Taiwan. Other non-longitudinal studies,that do


21/38

not attempt to control for selection effects, include a 16-nation study (Hu and Goldman 1995),work onIsraeli data (Manor et al 2000), and much evidence for the US (Johnson et al 2000, among others).As ever, causality is the issue. The longitudinal studies of marriage and mortality have typicallyadopted

the same methodology as the mental health studies. Some initial health measures are taken andused withother explanatory variables, including marital status, to predict mortality. Initial health measuresareusually self-reported, which is not ideal, because the presumption is that measurement error isgreater insuch data. However, it has been shown that self-reported health measures are reasonably good

predictorsof actual mortality (Idler and Kasl 1991).In this literature, Cox regressions are often used to estimate duration to death, or probability ofdeath

within a given period. Results are then typically described in Relative Risk ratios (so-called RRs).These Estesdescribe the relative risk of each marital-status group when compared to people who are inconventionalmarriage. casamento. Unlike in the mental health papers, the selection effect in this literature isoften estimatedseparately. separadamente.

Page 14Pgina 1414 14An obvious difficulty is that initial health measures may not be detailed enough. If not, then there

remainsomitted-variable bias, and estimates of the causal benefits from marriage are correspondinglycontaminated by some selection. Another problem is that marital status is often measured someyears

before death, so full marital history is not, and often cannot be, included in the regression. Tuckeret al(1996) and Cheung (2000) try to use marital history, but in doing so struggle to account for theselectioneffect or control for other variables.A fascinating paper by Murray (2000) attempts to overcome this by using a complete historicaldata set.

The paper follows a sample of subjects from an unmarried state at 18 years of age through todeath. The Osample are graduates from Amherst College, Massachusetts, born between 1832-79. Those whodid notgraduate are not followed. The sample gives date of first-marriage (only), so full marital history isunknown. desconhecido. The date of wife s death, however, is known. At 18, these men had theirheight and weightmeasured. medidos. Height -- although affected genetically -- depends on a person s nutrition andhealth. A BMI(Body Mass Index = height/mass squared) measure can be calculated. Research has found a non-linear

relationship between both height and BMI with mortality (Costa 1993), although the heightrelationship is


22/38

thought by some investigators to be spurious. In this historical Amherst study, height is dividedinto 5groups, from very short [>2sd from mean] up to very tall, and BMI into 4 groups.Regressions are estimated to explore the selection question: can initial health account for marital

behaviour? First, a logit model is constructed to estimate the consequences of marriage versus

lifelongbachelorhood. Other models are used to estimate the duration until first marriage. Underweightmen are4% less likely ever to marry and 12% less likely to marry in a given year (p =


23/38

formal organisation involvement, and leisure activities (passive and active). These are used topredictmortality in the usual fashion, within a binary logit equation. For men, being retired and smokingaresignificant predictors of death, ceteris paribus. The social measures prove especially interesting.

Frequency of attending voluntary associations, spectator events and classes all seem to reducemortality(p


24/38

the longer the duration of a marriage the greater the gains. Lund et al also find that the durationsince adivorce increases the risks of mortality, and that having higher numbers of marital dissolutions canincrease mortality.There are other persuasive papers in the literature on the health and mortality effects from

marriage.Lillard and Panis (1996) make a fundamental contribution to this literature and show clear benefitsfrommarriage, while Brockmann and Klein (2004) replicate and extend their findings. These papers areimportant because they simultaneously separate and measure both the selection and protectioneffects.Lillard and Panis use the same panel as Lillard and Waite (1995), but take a longer period, 1968-1990,and only include male individuals. First, the authors estimate a standard hazard equation, as would

becommonly used in the literature, while controlling for initial health and personal characteristics.

They Elesfind that the never married and the widowed have significantly higher mortality risks than themarried andseparated/divorced. Furthermore, the authors show that the health of never married and divorcedmenhealth deteriorates approximately 15% faster than that of married men.However, Lillard and Panis paper s greatest innovation comes from specifying these modelsalongsideequations that model the risks of marriage formation or dissolution. Unobserved variables maylead anindividual to have better health, lower mortality risk and higher marriage formation. By estimatingtheseequations simultaneously, it is possible to control for these effects and provide measurements ofselectionand protection effects. The procedure uses an instrumental variable technique and assumes thatseveralvariables concerning the individuals childhoods and parents affect only the individuals health butnoneof the other dependent variables. This assumption, however, is insensitive to the exactspecification. The Oauthors results show that, having controlled for the selection effect, large benefits of marriage

persist,with the never married and widowed men having larger risks of mortality. However, with thecontrols, thehealth effects differ slightly: now there is no significant difference between the health of the nevermarried and married, but the benefits of marriage remain when compared to the other maritalgroups.Most notably, in the selection equation they find that less healthy men are more likely to marry,which is

Page 17Pgina 1717 17again suggestive of adverse selection effects into marriage. Brockmann and Klein (2004) extend

Lillard


25/38

and Panis (1996) by incorporating the wider effects of marital history rather than only currentmaritalstatus. status. This allows them to see if the timing of a martial transition, the duration of atransition or thesequencing of marital transitions affects mortality. To do this they use German panel data from

1984 to1998. 1998. Compared to married men, widowed men have a 44% higher risk of mortality, anddivorced men a60% higher risk (after controlling for selection), while widowed women have 36% and singlewomen50% greater risks. These are enormous effects from relationships.Hibbard and Pope (1993) focus on finding out how the quality of a marriage affects mortality andmorbidity. morbidade. Marital quality is measured with some responses to questions designed to

provide index valueson of marital satisfaction, equality in decision-making and marriage companionship. Whilecontrolling for

baseline health and other personal factors, the authors analyse mortality over a fifteen-year followup.They find that marriage quality affects mortality risk for females but not males. Ben-Shlomo et al(1993)use a sample of middle-aged males from the UK civil service, over an 18-year period. In 1967-69,theirsample is of approximately 18,000 men, aged between 40 and 64 years old. After an 18-yearfollow up

period, those who had died (3433) are categorised by cause of death. The figures per 1000persons,adjusted for age, show that married individuals have a 13.9 mortality rate, while the widowed,single andseparated have rates of 20.6, 16.9 and 21.0 respectively. The main difference between married andsinglemen seems to come from differences in rates of cardiovascular disease and lung cancer. This maytell ussomething about how a protective effect operates.Somewhat different conclusions are drawn by Ebrahim et al (1995). Only the single have excessmortalityrisk, and not the divorced or widowed. The research monitors middle-aged British men from1978/80 to

1983/85. Explanatory variables in the statistical work include employment and occupation status,medicalhistory with medical assessments including activity measures, and smoking/drinkingconsumption.A 10-year panel of British households is used by Gardner and Oswald (2004) to compare themortalityeffects of marriage relative to income. They also find large benefits of marriage enough to offsettherisk of smoking for men, and enough to offset approximately half the smoking risk for women. Anovel Um romanceaspect in this study is the inclusion of mental stress measures in the controls for health. In

principle, thisnot only provides a better control for the estimation, but also allows the authors to test whether the


26/38

beneficial effect from marriage arises through reductions in stress. The data suggest it does not.After Depoiscontrolling for other factors, the coefficient on marriage in the authors mortality equation isunaffected

Page 18Pgina 18

18 18by the inclusion of a mental health measure (the so-called GHQ score). This suggests that thereare otherimportant, and as yet incompletely understood, channels at work behind the correlation betweenmarriageand mortality.In Tucker (1996), marital history is monitored up to the middle of life. Cox regressions are used to

predictmortality. mortalidade. Marital turbulence matters. Adjusting for self-reported health,inconsistently married men havea 40% greater risk of dying, the divorced a 250% greater risk, and the single no significant riskabove theconsistently married males. When social-ties measures are allowed for, the risk for single males iseliminated, and the risk for divorced males is somewhat reduced. For females, whether or notsocial tiesare included in the equation, the divorced have a dramatically higher risk of mortality than themarried.Cheung (2000) explores a longitudinal data set on British women, and has a time-varyingindependentvariable to allow for marital status durations. An initial assessment in 1984-5 in which the womenwere

35 and older was followed up twelve years later. Self-reported health, height, education andsmoking/drinking habits are included, but social networks and financial status are not. The results findexcessmortality in the unmarried but not the divorced or widowed. The single face an odds ratio as highas 1.45(p


27/38

Summarizing this evidence, marriage has enormous effects on people s longevity.Section IV: Health Benefits from MarriageSince a much-cited study by Verbrugge (1979), many investigators have tried to work out whethermarriage makes people generally healthier.

Page 19Pgina 19

19 19Joung et al (1998) takes longitudinal data from the Dutch GLOBE Project. Here the aim is a littledifferent. The data are used to test whether health measures can predict marital status ie if there isahealth-driven selection effect. The GLOBE study is longitudinal and follows individuals aged

between15-74; the subjects come from around Eindhoven; the survey response rate was 70%. Here allthose livingin Eindhoven are used (10811). Between 1991 and 1995, marital status changes, deaths andmigrationsare all monitored. Perceived general health, subjective health complaints and chronic conditionsare usedas health measures. Age, sex, education level, religion and employment status are controlled for.Regressions are used to establish relative risks (RRs) of the various changes in marital status. TheOtransition from marriage to divorce reveals evidence that health determines the probability ofmarriage.Waldron et al (1996) demonstrate that jobs provide valuable social networks outside marriage, andthatthese improve mental and physical health. The unemployed lose this channel of emotionalsupport,

(Waldron and Jacobs 1988,1989). The authors use the National Longitudinal Surveys of LabourMarketExperience, covering US women in 1968,1978,1983,1988. The cohort start aged 14-24yrs, andend at 34-44yrs. A health scale is constructed from 11 questions on physical difficulties and 6 questions onmoregeneral difficulties, eg tiredness. Working hours are measured. Other controls include age, genderand

parental status. No measures for income or social networks are available. For 1978 and 1983-88,OLS isused to predict health status

to test for protective effects from being married. Initial marital statusseemed to be a good predictor of later health. This varies, however, as interaction terms areincluded.Part-time and full-time married women appear to gain no protection; only non-working marriedwomendo. fazer. No other significant differences in the protective effect is found. Next, selection effectsare tested by

predicting marital status over the 2 periods. In the 1st r

period, initial health is a strong indicator, but not inthe 2 os dois

nd nd. . Again the effect varies by employment. No selection effect is discovered for full-timers.


28/38

The possibility of adverse selection into marriage is supported by studies based on a British cohort(Cheung 1998, Cheung and Sloggett 1998). The National Child Development study followschildren bornin a particular week of 1958 with follow-ups at 7,11,16,23, and 33. Variables included are socialclass and

housing tenure, education, unemployment record, alcohol / smoking consumption, healthmeasuresheight, weight, self-reported measures and limiting illnesses. Also included is a variable aimed atestimating temperament. Cheung (1998) uses those married or divorced at 33 years of age to try tounderstand the probability of divorce. Logit models for females indicate that council renters andheavydrinkers are more likely to divorce. An adverse selection effect in health is found. Those with poorself-rated health are less likely to divorce, with an odds ratio of 0.57 p


29/38

marriage quality reducing distress, which in turn reduces risky behaviours, which brings healthbenefits.In similar work, Prigerson et al (1999) link marital harmony with measures of health for a set ofmarriedwomen over a three year follow up. The health measures used are more robust -- including not

only self-reports of specific conditions but health service use. Results show that marital harmony isassociated withmuch better sleep, less depression and fewer visits to the doctor. Williams and Umberson (2004)analysethe effects of marital history and the timing of marital transitions on health. They use a short US

panel.Remarriages confer benefits that weaken with age. The latest work continues to suggest that a

person swhole marital history matters (Hughes and Waite, 2005). Much more research is needed here.Health, also, is improved by marriage especially a good one.

Page 21Pgina 2121 21Section V: Investigations into the Channels Providing the Benefits from MarriageIf we accept from all this statistical evidence that human beings gain some kind of protectiveeffect frommarriage, how are those benefits actually transmitted? The answer is not yet known.Married couples may gain financially. Marriage may bring increased emotional and instrumentalsupport.Marriage may change lifestyles because of some kind of guardian effect, where healthy activitiesare

increased and risky behaviours reduced. We have already seen some possible evidence for suchchannels;but again, selection effects may distort these results. Higher socio-economic status and financialadvantages for the married have been shown (Brown 2000, Ross 1995, Rindfuss and VandenHeuvel1990). , 1990). Ross (1995) establishes that the married have higher levels of emotional support.Evidence ofmarried individuals drinking and smoking less has been found, perhaps suggestive of a guardianeffect(Horwitz et al 1996, Ben-Shlomo 1993). Rogers (1995) prove that married people are less likelyto die

from accidents, suicide and cirrhosis of the liver.We should mention three papers that explore these explanations. All use cross sectional data, sohavelimitations, but the patterns are of interest. The papers are as follows. Joung et al (1997) have datathatcome from the GLOBE project in Eindhoven and its surroundings. The response rate was 70%.Wyke andFord (1992) review the explanations, and test them. A sample of 55-year old women from aroundGlasgow is used. Hahn (1993) examines 14,000 households in the US. All three papers point tofinancialfactors as being important.

Umberson (1987, 1992) investigates the guardian effect further. A US sample of 1826 married or


30/38

divorced/separated individuals in the mid 1970s is used to investigate this channel. Six indicesmeasuringorderliness of lifestyle, risk taking behaviour, drinking problems, drinking and driving, marijuanause andsubstance abuse related to stress are all constructed. Parental status (kids at home, kids not at

home, nokids), age, education, gender and race are also all controlled for statistically. The divorced arefound to bemore likely to have adverse (higher) scores in the indices -- showing a lack of healthy behaviourand highrisk-taking behaviours.By looking at changes in drinking patterns, Power (1999) attempts to factor out the selectioneffect. The Owell-known British birth cohort of 3-9 March 1958 had data collected about them at ages7,11,16,23 and33. 33. At age 33, 69% of the sample (11045) remain. Getting married per se seems to reduce

male drinking;but parenting is a stronger force in reducing married women s drinking. The recently divorcedchange

Page 22Pgina 2222 22their drinking behaviour after remarriage. They show a large rise in drinking but the continuouslymarriedand not recently divorced show small declines in drinking. This paper demonstrates the absence ofstrongselection effects in terms of alcohol consumption, but strong effects for those experiencing

changes inmarital status. estado civil. Becoming married makes alcohol consumption fall, and recent divorcecauses it to rise.This would seem to confirm the existence of at least some guardian effect. Wickrama et al (1995)findthat positive interaction within a marriage increases health by reducing the propensity to live anunhealthylifestyle. The study uses a 4-wave panel of 320 men who had been married over 14 years with atleast 2children. crianas. Burke et al (2004) uncover mixed evidence of the effect of living with a

partner. The authors use

a panel that follows a set of nine-year old, Australian children from 1985. The study tracks theconsequences of living with a partner for health, but no distinction is made between marital andnon-marital cohabitation. The individuals in 2000 are only 25 years old, so not only are the measuredeffects

possibly short-term but they are also likely to capture young partnerships that may not be entirelyrepresentative of the population. Despite these shortcomings, the results are interesting and showthat, forwomen especially, there is a notable increase in body mass index associated with living with a

partner,which has been noted before (eg Sobal et al (2003)). In contrast, men have a reduced fat and

energyintake in their diet, perhaps as an effect of the partner changing their eating habits.


31/38

The papers covered in this section do not offer overwhelming evidence for any of the explanationssuggested about why marriage is good for people. Power et al (1999) is the only study that canreject theselection explanation, and Power et al s study is limited by possible misspecification.In our judgment, a more promising way to think about all these phenomena is to look at laboratory

work.Cohen et al (1997, 2003) show that, in randomised trials, people who have stronger social ties arebetterable to withstand cold viruses. In each of these two studies, some hundreds of volunteers areexposed tonasal drops containing 1 of 2 rhinoviruses, and monitored to see who develops the common cold.The Osubjects are interviewed about their social networks and relationships, and studied in actual socialsettings. configuraes. Increased sociability is found to be associated in a monotonicallydecreasing way with the

probability of developing a cold. The statistical association remains after allowance for many

controls,including baseline immunity, demographic factors, smoking, alcohol consumption, education,

body massindex, and diet. This research is important.How such protective channels work remains an open question. The results in Gardner and Oswald(2004)appear to indicate that it is not merely that marriage keeps people alive by reducing stress levels.In Em

Page 23Pgina 2323 23

mortality equations on longitudinal British data, the authors find that their large negativecoefficient onmarriage is unaffected by the inclusion of a mental stress variable.Conclusions ConclusesMarried people live longer and are healthier. What is less clear is whether this pattern in the datatells ussomething reliable about what marriage does to health.This paper reviews the modern longitudinal evidence. By studying people over time, it is possibletodraw more persuasive judgements than by looking at cross-sectional evidence. The same personcan be

followed through the years, and that makes it feasible to factor out influences that are specific totheperson. pessoa. Longitudinal data also make it possible for people s early health to be heldconstant within anequation that explains their health and mortality later in life. It is important to be clear,nevertheless, thatlongitudinal data do not solve all problems in this or other fields. First, as the philosopher DavidHume

pointed out, just because something moves first does not mean that it causes what goes second, nomatterhow reliable the correlation. Hume s example was the crowing of the cock followed by the

invariable


32/38

rising of the sun. Second, human beings are not, of course, randomly assigned to marriage. Theychooseit. ele. Because of that, there are likely always to be difficulties about how to interpret the deeplinks betweenmarriage and wellbeing.

Nevertheless, our reading of the longitudinal evidence, after looking across studies in a variety ofliteratures, suggests that:Marriage makes people far less likely to suffer psychological illnessMarriage makes people live much longerMarriage makes people healthier and happierBoth men and women benefit, though some investigators have found that men gain moreThese gains are not merely because married people engage in less risky activitiesMarriage quality and prior beliefs can influence the size of the gains.In the terminology of this research field, there is a genuine protection effect from marriage. The

pattern O padroin the data is not a cross-section illusion. Moreover, it is large: for males the longevity effect of

marriagemay even offset the consequences of smoking. What, in a deep sense, are the processes at work? Itdoes Ele faz

Page 24Pgina 2424 24not seem merely to be that having a wife or husband lowers a person s stress and that this makesthe

person physically healthier.Marriage is arguably the most fundamental of all social ties. Hence the Cohen laboratory studies(1997,

2003), which show that human beings with good friendship networks can repel the simplecommon cold,are particularly suggestive. Our empirical instinct is that intimate friendships provide a kind ofdeep

protection from attacks such as infections and medical disorders. Perhaps, figuratively speaking,marriage is like having a reserve army of volunteers, rather than simply a large army of regulars.How Comomarriage works its magic remains mysterious. If an answer can be found, it may unlock the puzzleofexactly how the mind and body are intertwined.

Page 25Pgina 25

25 25References RefernciasBarrett, AE (2000), Marital Transitions and Mental Health , Journal of Health and SocialBehaviour,45, 451-464.Becker, GS (1981), A Treatise on the Family, Chicago: University of Chicago Press.Ben-Shlomo, Y., GD Smith and M. Shipley et al (1993), Magnitude and Causes of MortalityDifferences Between Married and Unmarried Men , Journal of Epidemiology and CommunityHealth, 47,200-205.Berkman, LF (1988), The Changing and Heterogeneous Nature of Aging and Longevity: A SocialandBiomedical Perspective , Annual Reviews in Gerontology and Geriatrics, 8, 37-68.


33/38

Berkman, LF and SL Syme (1979), Social Networks, Host Resistance and Mortality: A Nine YearFollow-Up Study of Alameda County Residents , American Journal of Epidemiology, 109, 186-204.Blanchflower DG and AJ Oswald (2004a), Well-Being Over Time in Britain and the USA ,Journal of

Public Economics, 88, 1359-1386.Blanchflower DG and AJ Oswald (2004b), Money, Sex and Happiness: An Empirical Study ,Scandinavian Journal of Economics, 106, 393-415.Brockman H. and T. Klein (2004), Love and Death in Germany: The Marital Biography and ItsEffect onMortality , Journal of Marriage and the Family, 66, 567-581.Brown, SL (2000), The Effect of Union Type on Psychological Wellbeing: DepressionCohabitantsVersus Marrieds , Journal of Health and Social Behaviour, 41, 241-255.Brown, SL and A. Booth (1996), Cohabitation Versus Marriage: A Comparison of RelationshipQuality , Journal of Marriage and the Family, 58, 668-678.

Bumpass, LL and JA Sweet (1989), National Estimates of Cohabitation , Demography, 26, 615-25.Burke V., LJ Beilin, and D. Dunbar et al. (2004), Changes in Health Related Behaviours andCardiovascular Risk Factors in Young Adults: Associations with Living with a Partner ,PreventiveMedicine, 39, 722-730.Burman, B. and G. Margolin (1992), Analysis of the Association Between Marital RelationshipsandHealth Problems: An Interactional Perspective , Psychological Bulletin, 112, 39-63.Cheung, YB (1998), Can Marital Selection Explain the Differences in Health Between MarriedandDivorced People? From a Longitudinal Study of a British Birth Cohort , Public Health, 112, 113-117.Cheung, YB (2000), Marital Status and Mortality in British Women: A Longitudinal Study ,International Journal of Epidemiology, 29, 93-99.Cheung, YB and A. Sloggett (1998), Health and Adverse Selection into Marriage: Evidence fromaStudy of the 1958 British Birth Cohort Study , Public Health, 112, 309-311.

Page 26Pgina 2626 26Chun, H. and I. Lee (2001), Why Do Married Men Earn More: Productivity or Marriage

Selection? , ,Economic Inquiry, 39, 7-19.Clark, AE, E. Diener, Y. Georgellis, and RE Lucas (2003), Lags and Leads in Life Satisfaction: ATest of the Baseline Hypothesis , working paper, DELTA, Paris.Clark, AE and AJ Oswald (2002), A Simple Statistical Method for Measuring How Life EventsaffectHappiness, International Journal of Epidemiology, 31, 1139-1144.Cohen, S. (2004), Social Relationships and Health , American Psychologist, 59, 676-684.Cohen, S., WJ Doyle, DP Skoner, BS Rabin, and JM Gwaltney (1997), Social Ties andSusceptibility to the Common Cold , Journal of the American Medical Association, 277, 1940-1944.

Cohen, S., WJ Doyle, R. Turner, CM Alper (2003), and DP Skoner Sociability and Susceptibilityto


34/38

the Common Cold , Journal of the American Medical Association, 14, 389-395.Collins, R. and S. Coltrane (1992), Sociology of Marriage and the Family: Gender, Love andProperty 3rd rded. ed. Nelson Hall, Chicago

Coombs, R. (1991), Marital Status and Personal Wellbeing: A Literature Review , FamilyRelations, 40,97-102.Costa, DL (1993) Height, Weight, Wartime Stress and Older Age Mortality: Evidence from UnionArmy Records , Explorations in Economic History, 30, 424-49.Daniel, K. (1995) The Marriage Premium Ch.7 from The New Economics of Human Behaviour(eds)M. Tommasi and K. Ierulli, Cambridge University Press.Di Tella, R., MacCulloch, R., and Oswald, AJ (2003), The Macroeconomics of Happiness ,Review ofEconomics and Statistics, 85, 809-827.

Ebrahim, S., G. Wannamethee and A. Mccallum et al. (1995), Marital Status, Change in MaritalStatusand Mortality in Middle Aged British Men , American Journal of Epidemiology, 142, 834-842.Gardner J. and A. Oswald (2004), How is Mortality Affected by Money, Marriage and Stress? ,Journalof Health Economics, 23, 1181-1207.Gardner J. and A. Oswald (2005), Do Divorcing Couples Become Happier by Splitting Up? ,working

paper, Warwick University. papel, Warwick University.Glenn, ND and CN Weaver (1988), The Changing Relationship of Marital Status to ReportedHappiness , Journal of Marriage and the Family, 50, 317-324.Goldman, N., S. Korenman and R. Weinstein (1995), Marital Status and Health Among theElderly ,Social Science and Medicine, 40, 1717-1730.Gove, WR, M. Hughes and CB Style (1983), Does Marriage Have Positive Effects on thePsychological Well-being of the Individual? , Journal of Health and Social Behaviour, 24, 122-131.

Page 27Pgina 2727 27Hahn, BA (1993), Marital Status and Women s Health: The Effect of Marital Economic MaritalAcquisitions , Journal of Marriage and the Family, 55, 495-504.

Helmer, C., D. Damon and L. Letenneur et al. (1999), Marital Status and Risk of Alzheimer sDisease ,Neurology, 53, 1953-1958.Hibbard JH and CR Pope (1993), The Quality of Social Roles as Predictors of Morbidity andMortality , Social Science and Medicine, 36, 217-225.Horwitz, AV, HR White and S. HowellWhite (1996), Becoming Married and Mental Health: ALongitudinal Study of a Cohort of Young Adults , Journal of Marriage and the Family, 58, 895-907.Horwitz, AV and HR White (1998), The Relationship of Cohabitation and Mental Health: AStudy ofa Young Adult Cohort , Journal of Marriage and the Family, 60, 505-14.

House, JS, C. Robbins and H. Metzner (1982), The Association of Social Relationships andActivities


35/38

with Mortality: Prospective Evidence from the Tecumseh Community Health Study , AmericanJournalof Epidemiology, 116, 567-567.House, JS, KR Landis and D. Umberson (1988), Social Relationships and Health , Science, 241,540-

545.Hu, Y. and N. Goldman (1995), Mortality Differentials by Marital Status: An InternationalComparison , Demography, 32, 485-507.Hughes, ME and LJ Waite (2005) Marital Biography and Health at Midlife , working paper,Department of Sociology, Duke University.Idler, EL and S. Kasl (1991) Health Perceptions and Survival: Do Global Evaluations of HealthStatusReally Predict Mortality? , Journal of Gerontology, 46, S55-65.Johnson, DR and J. Wu (2002), An Empirical Test of Crisis, Social Selection, and RoleExplanations ofthe Relationships Between Marital Disruption and Psychological Distress: A Pooled Time Series

Analysisof Four-Wave Panel Data , Journal of Marriage and the Family, 64, 211-224.Johnson, NJ, E. Backlund and PD Sorlie et al (2000), Marital Status and Mortality: The NationalLongitudinal Mortality Study , Annals of Epidemiology, 10, 224-238.Joung, IMA, K. Stronks and HD Vandemheen et al. (1997), The Contribution of IntermediaryFactorsto Marital Status Differences in Self Reported Health , Journal of Marriage and the Family, 59,476-490.Joung, IMA, HD Vandemheen and K. Stronks et al. (1998), A Longitudinal Study of HealthSelectionin Marital Transitions , Social Science and Medicine, 46, 425-35.Kessler, RC and M. Essex (1982), Marital Status and Depression: The Importance of CopingResources , Social Forces, 61, 484-507.Kiecolt-Gla

72942033 Como o Casamento Afetam a Saude Psicologica e Fisica

Documents

Transcript of 72942033 Como o Casamento Afetam a Saude Psicologica e Fisica