Here’s a fun research project for you.
Part 1
David Weakliem wrote:
On March 30, Nicholas Kristof had wrote:
Survey data indicates that married couples on average report more happiness, build more wealth, live longer and raise more successful children than single parents or cohabiting couples, though there are plenty of exceptions.
The most popular reader comment, with over 2000 likes, said
I notice that you didn’t talk about the research that shows differences in gender in marriage happiness levels. Marriage is generally GREAT for men, who report being far happier in marriage than being single. Much research indicates the reverse is true for women. Single women report being happier, in general, than married women are.
Most of the other leading comments were along the same lines.
Weakliem then looked up some data:
Since the 1970s, the General Social Survey has asked whether “taken all together, how would you say things are these days.” 41% of married women say that they are very happy, and 8% say they are “not too happy” (the rest chose “pretty happy”); among single (never married) women, 23% say they are very happy and 17% not too happy. This is just one survey, but it’s one that puts a lot of effort into selecting a representative sample and gets a high response rate, so it can be considered to be pretty definitive. Of course, the difference in happiness between married and never-married women is not necessarily caused by marriage, but it’s there.
And he shares this graph:
The smug married of both sexes seem to be doing just fine. Weakliem continues:
I looked to see if the gap varied by other characteristics—for example, education, race, political views. To make a short story even shorter, I found nothing worth mentioning.
But he did find that men were more likely than women to say they thought that married people were generally happier than unmarried people. So that’s something.
This is all old data so it’s just male/female marriages. Same-sex marriage is another story; you can google *happiness same sex couples* and find some references.
To summarize the story so far, we’ve seen four perspectives:
Kristof: Married people are happier than single people.
Commenters to Kristof’s article: Married men are happier than single men; married women are unhappier than single women.
General Social Survey (via Weakliem): Married people are happier than single people, with roughly the same difference for men and women.
General Social Survey (via Weakliem): Men are more likely than women to think that married people are happier.
Part 2
This got me curious so I did some googling and found all sorts of things:
– This article in Psychology Today from 2022 states, “Married men are happier than married women, and unmarried women are happier than unmarried men.” But it doesn’t offer any citation for this claim, except a 2015 article in Psychology Today by a different author, which states:
Marriage, we have been led to believe, is a natural habitat for women, but a stifling cage for men. Thus goes the popular fantasy. However, in the real world of data, things shake out quite a bit differently. . . . confounding the view of marriage as the female heaven and haven is the fact that marriage actually appears to benefit men more than it does women. Research has shown that the “marriage benefits”—the increases in health, wealth, and happiness that are often associated with the status—go disproportionately to men. Married men are better off than single men. Married women, on the other hand, are not better off than unmarried women.
Interesting. This is basically the opposite of Weakliem’s take. Weakliem wrote that there’s a popular view that marriage benefits men but not women, but that actually the gap in happiness between married and unmarried is the same for the two sexes. This Psychology Today article takes the opposite position, stating that there’s a popular view that marriage benefits women more than men, but that actually the opposite is the case.
We know where Weakliem’s data come from—the General Social Survey from 1972 onward. What’s the source of the claims in the Psychology Today article? The “benefit men more than it does women” link sends us to an article in the journal Family Relations from 1991, “Marital Status and Personal Well-Being: A Literature Review,” which has a lot of things, but I’ll focus on the part on Happiness:
Self-report studies of happiness indicate that the married are happier than the unmarried (Bradburn, 1969; Bradburn & Caplovitz, 1965; Campbell, 1981; Glenn, 1975; Gurin, Veroff, & Feld, 1960; Schmoldt, Pope, & Hibbard, 1989) . . . Studies also indicate that married men are happier than married women. . . . Radloff (1975) found that men benefit more than women from marriage. . . . Virtually all data show that unmarried men have lower levels of happiness than their female counterparts. This holds true for never-married, widowed, and divorced males (Campbell, 1981; Gurin et al., 1960). . . . Studies also show that single women are often happier than single men. Gove (1972a, 1972b, 1973) explains that single women are more likely to develop strong social ties, such as close relationships with family and friends. Buffered by emotional support of others, these women, compared to unmarried men, report greater happiness. . . . However, they are not happier than married women (Tcheng-Laroche & Prince, 1983). . . . Popular folklore has it that marriage is a blessed state for women and a burdensome trap for men. In reality, however, men more than women, receive marriage’s mental health benefits because women, more than men, provide emotional aid and other support in marriage. . . .
The Psychology Today article also has a “happiness” link, which sends us to an article from the Australian Daily Telegraph newspaper in 2015. The link no longer works, but I was able to find the article on the Internet Archive. Here’s the relevant bit:
According to the [Household, Income and Labour Dynamics in Australia] study of 20,000 people between 2001 and 2012, on a scale of zero to 10, married men in 2012 gave their relationship an 8.5, while women gave it an 8.2. . . . being married improves men’s sense of wellbeing, but doesn’t offer the same boost for women (though neither was it a negative). . . . The study also showed men’s sense of wellbeing and health benefited from being married, while women’s was not affected either positively or negatively by being married.
And then this:
The finding that marriage makes men happier but doesn’t affect women in the same way echoes other international studies.
Unfortunately, no links here, either to the Australian study or to the international studies. Some googling turns up this 2021 report from Australia, “Social determinants of subjective wellbeing” which states:
Married individuals (both men and women) experience higher levels of life satisfaction than those in other family arrangements (Evans & Kelley 2004). As well, transitions into relationships, marriage or cohabitation have been shown to significantly increase wellbeing, while transitions out of relationships due to separation or widowhood negatively affect wellbeing (Baxter & Hewitt 2014).
Following up on those references, Evans & Kelley 2004 write:
Men and men in formal marriages experience higher levels of life satisfaction than do people in other family arrangements. . . . In general, the average levels of life satisfaction of men do not differ from those of women within each marital status . . .
And Baxter & Hewitt 2014 write:
Men and women who are married have higher levels of wellbeing than those who are not married, although there are no significant differences in wellbeing between people cohabiting intending to marry and married. We find that transitions into relationships, marriage or cohabitation, significantly increase wellbeing while transitions out of relationships because of separation, or widowhood, negatively impact on wellbeing. We find no gender differences in these patterns and no significant differences between cohabitation and marriage.
That’s from 2014 and it’s based on that same Households, Income and Labour Dynamics in Australia. I’m not sure what happened between 2014 and 2015 for the story to have changed so much.
I also found this 2019 report, “Are Married People Happier?,” from some organization called Population Europe, which states:
Numerous studies have been published that have examined subjective wellbeing (SWB) and marriage status, finding that married people tend to have a higher SWB. But in today’s society, more couples are opting for cohabitation, which includes many benefits associated with marriage. This then leaves the question of whether individuals who cohabit have similar levels of SWB as married people. Brienna Perelli-Harris, Stefanie Hoherz, Trude Lappegård and Ann Evans look more closely at this question by exploring the situation in Australia, Germany, Norway and the United Kingdom. . . . The authors used harmonized data sets from the UK Household Longitudinal Study (UKHLS), the Household, Income and Labour Dynamics in Australia (HILDA) survey, the Socio-Economic Panel (SOEP), and the Norwegian Generations and Gender Survey (GGS) to analyse partnership status in midlife. They found that differences between marriage and cohabitation disappear in all countries once selection and relationship satisfaction are considered, which means marriage does not lead to higher SWB.
Just to clarify, that last statement (“marriage does not lead to higher subjective wellbeing”) is marriage compared to cohabitation, not marriage compared to being single. This indicates one of the challenges of following this literature: in some analyses, cohabitation is considered to be equivalent to marriage; in other analyses, cohabitation counts as single.
For the purposes of our investigation here, I’m most interested in married vs. single, and I’m inclined to think of cohabitation as just one form of marriage; this would be consistent with the above-cited article.
Getting back to the U.S., here’s a recent report of a recent Gallup poll stating:
Both married men and married women see a 20-percentage-point advantage compared to their same-sex peers who never married.
The report is called, “Married People Are Living Their Best Lives,” which is vivid, but, as a statistician, I wish they’d added “On Average” somewhere in that title.
But then there’s this article from 2019 which states:
We may have suspected it already, but now the science backs it up: unmarried and childless women are the happiest subgroup in the population. . . . Paul Dolan, a professor of behavioural science at the London School of Economics, said the latest evidence showed that the traditional markers used to measure success did not correlate with happiness – particularly marriage and raising children.
“We do have some good longitudinal data following the same people over time, but I am going to do a massive disservice to that science and just say: if you’re a man, you should probably get married; if you’re a woman, don’t bother.” . . .
Dolan’s latest book, Happy Ever After, cites evidence from the American Time Use Survey (ATUS), which compared levels of pleasure and misery in unmarried, married, divorced, separated and widowed individuals.
But then a google of *paul dolan american time use survey* turns up this news article entitled, “A new book says married women are miserable. Don’t believe it,” which points to this series of posts by economist Gray Kimbrough, who wrote, regarding the claim, “healthiest and happiest population subgroup are women who never married or had children,” that “the ATUS lacks data on *ever* having children, but I can compare never/ever married with and without children in the household. This doesn’t back up his claim.” But I guess the erroneous reports will stay on the internet forever, for example here, here, and here.
Part 3
What to think here? What to do? I’m not sure. Here’s where we are:
– There are varying claims regarding the difference in happiness comparing married to unmarried men or women. Starkly different claims! Various claims are based on different surveys in different countries, but each claim is made in a different way, so it’s hard to know how to put all this together. Weakliem’s analysis of the General Social Survey seems pretty convincing, but there are lots of other studies out there.
– There are also varying sources of evidence regarding popular belief. Weakliem cites a mainstream newspaper columnist writing about the general benefits of marriage, but also a comment with thousands of likes expressing the belief that the research shows marriage to be associated with greater happiness for men and lower levels happiness for women.
– Muddying the waters are statistical complexities such as how the cohabitants are classified, and whether we’re talking about association or causation. There’s also the question of whether divorced people and never-marrieds are lumped into a single “unmarried” category; from the causal direction, there’s no reason to think that the effects of marriage would be the opposite of the effects of divorce.
For my purposes here, I’d just like to nail down the descriptive questions: (1) what are the average differences between married and unmarried men and women in happiness (and how does this depend on the measurement of happiness/wellbeing/satisfaction/etc. and the classification of cohabitants and never-married/widowed/divorced, and what’s the story with same-sex couples), and (2) what do people believe about that first question. I’m not so concerned right now about the causal questions; ultimately, yes, I do care about causal effects of marriage, but it would seem enough as a starting point to resolve the descriptive questions.
This seems like a fun research project. God is in every leaf of every tree and all that.
I’m not so concerned right now about the causal questions; ultimately, yes,
I don’t see how this is meaningfully separable. I kept waiting to read the part where you talk about whether married people (men or women) are happier, or whether happier people (men or women) are more likely to get married. I don’t really get how how it’s meaningful to look at the one aspect without the other.
Maybe if I turn it around. Are happier people more likely to get married? Nah. Still don’t see a whole lot of point in asking that question either.
I’m still looking for the control group. Since we are measuring (crudely at that) self-reported well-being over time and across demographic groups, I keep wondering about the overall level of well-being – on average – and how that may have changed over time. And, if it has changed, how does that play out in the disaggregated groups – those groups may exhibit differential adjustments when overall happiness is rising or falling over time. I feel like there should be some kind of underlying dynamic theory about how feeling of well-being are established and change over time. Included in this should be something about the ages of the people, time since marriage, time since divorce, etc. So many variables and so little structure, it seems to me. I’d be happy with a clear descriptive picture.
Dale –
Yea, I also thought about the question of cross-sectional versus longitudinal data. I didn’t mention it because I think there would be some value in cross-sectional data here, but ultimately, really, how much value would there be, especially if there isn’t anything about so many possible interaction effects like age, economic status, time since marriage, and in particular, whether there are children (presumably, having children would have a differential moderating effect between the happiness of women, who shoulder the heaviest burden of caregiving, versus men)?
I just don’t get that there’s really any value in looking at these data without so much needed information to place it into a meaningful context.
I agree that the longitudinal/time aspect is crucial to the discussion. In the blog above we read that there is a group of researchers who have found that *getting into* relationships tends to increase happiness, and *getting out* of them tends to decrease it, on average. Since most people enter relationships voluntarily, the entering part is entirely credible (if it made them unhappy, they would not enter the relationship, unless they were forced into the relationship). The leaving part is more difficult. Leaving usually makes people unhappy (evidence: all those sad love songs mourning the loss of a relationship), but in some cases the opposite is true (e.g. fleeing a situation of domestic violence, other evidence: comparatively few songs about being happy about a break-up).
I think the stage of the relationship, with its changing dynamics, complicates cross-sectional analyses. If the population includes a lot of people who have recently started a relationship, this would certainly make marriage and cohabitation look attractive. If the population contains a lot of people who are about to break up, it would look like marriage reduces happiness. The distribution of relationship stages would probably be massively influenced by cohort effects… It is an exciting area of research, but really tricky, IMO.
“comparativ that celebrate breakupsely few songs about being happy about a break-up”
I am pretty certain that is true, but there are a lot of great songs about fortunate breakups, particularly in country music. Brandy Clark’s Hungover (a particular favorite of mine), The Dixie Chicks Goodbye Earl, Carrie Underwood’s Before He Cheats, half of Taylor Swift’s oeuvre (the other half feels bad about a breakup)
The fact that all of these songs are sung by women says something about the underlying theme of the post, I guess. Lots of Rap songs have men happy to leave particular women, but I’m not sure any of them were ever married beforehand.
I haven’t looked at data/literature in this space for over a decade, but my strong recollection from when I did (British Household Panel Survey) was that, longitudinally, people exiting marriages were on average happier afterwards than they were in them. (I assumed because it’s the unhappy marriages that tended to end.)
If that is the mechanism (unhappy marriages disproportionately end, leaving happy ones) would the effect be smaller where divorce is more difficult? Or would other counter effects (people thinking differently about entering marriage where divorce is harder, different cultural values, etc) disrupt that “logical” conclusion?
I also wonder about the differences between going from “multiple short/casual relationships -> married” vs “no relationship -> married” vs “long term stable committed relationship -> finally formally getting married”.
Right at the beginning was this:
“On March 30, Nicholas Kristof had wrote”
so I sort of gave up. However, because my wife is helping me recover from minor surgery, I thought I would press on.
“Happiness” is so vague such that the response is virtually meaningless. What is definitely true is that my employer at the time in the 1960s, UNESCO, paid me extra only when I, that is we, filed a marriage certificate.
Much less vague, and possibly easier to measure, is whether or not there is a name change upon marriage and why or why not it took place. And, does it correlate with the number of people at the wedding?
Another way to think about this is to just take the general social survey at face value and accept what it makes clear without reading anything else into it. Seems like the data are pretty straight forward and as sound as any survey can be. The questions are straight forward and no one is reading anything into the answers with unsupported assumptions. Weakliem notes:
“This is just one survey, but it’s one that puts a lot of effort into selecting a representative sample and gets a high response rate, so it can be considered to be pretty definitive.”
Chipmunk:
I also like the GSS. A couple issues do arise, though. First, their happiness question has only three levels (“very happy,” “pretty happy,” “not too happy”), and most of the responses are in just two of those categories. Other surveys use 1-10 scales etc. which can provide more information. Second, even when restricting oneself to the GSS there are many options depending on what years of data you use. This came up a few years ago in the context of a botched GSS analysis that appeared in the New York Times.
Here’s another post from a few years back on the challenges of studying happiness and well-being.
Andrew,
I recall that post. In the social sciences the caveats always seem infinite…
However, it occurred to me that there is possibly a strong bias towards happiness in marriage: if a person is unhappily married, it’s fully within their control to become unmarried. However, if a person is unhappily unmarried, they cannot simply decide to get married and be assured of success. So it’s hard for unahppy people to get married and easy for people who become unhappy in marriage to get unmarried.
chipmunk (the poster who won’t be accountable for making false accusations and shall henceforth be referred to as such) –
You say:
no one is reading anything into the answers with unsupported assumptions.
Christoff’s paper was titled “The Case for Saying I Do.” It clearly suggest that the act of getting married makes people happier. Outside of all the other questions about the data, they don’t support that conclusion. So actually, people are reading answers with unsupported conclusions. Without unsupported conclusions, one couldn’t determine whether getting married makes people happier, or if happier people are more likely to marry.
This may be a particularly important time to clarify some of those questions – since a certain politician running for very high office us strongly advocating that people who decide to not marry harm society and should have less voting power. Especially, apparently, if they have cats.
“no one is reading anything into the answers with unsupported assumptions. ”
As usual, Joshua, you misconstrue my statement to have much broader implications than are even remotely concievable from the context.
My statement applies – as I think any reasonable reading of it makes 100000% clear – specifically to the GSS survey, which does not interpret the answers to people’s questions, but simply reports the answers and is therefore reliable within general limitations of polls and surveys. My statement does not refer to the broader social implications of marriage or any social aspect thereof, or the veracity or non veracity of any statements made by any politicians about that. I don’t know who made such statements. I don’t know what statements they made.
Note what I said about the GSS survey: “without reading anything else into it. ” IOW, at the level of resoultion of the survey with all the caveats implied, married people report being happier. End of story. No causal relationship is indicated or claimed, no link to anything in society is construed. It’s just a report of what people say about how they feel. Andrew’s comment notwithstanding, I think it’s pretty clear that, for whatever reason, married people claim to be happier. However, as I suggested in my response to Andrew it seems likely that this is almost certain to always be true, as long as unhappy people can divorce at will.
Commenter who shirks accountability -.
Now watch closely:
I apologize for my misreading of your comment.
Try it. It feels good.
I have nothing in particular against the GSS, but in this space, I do have a strong bias towards panel data that allows you to track correlations between changes in reported satisfaction and changes in circumstances. This removes a lot of potential confounders that can influence cross-sectional results
This small section of the quoted 2015 Psychology Today really stood out to me:
> Marriage, we have been led to believe, is a natural habitat for women, but a stifling cage for men. Thus goes the popular fantasy. However, in the real world of data, things shake out quite a bit differently…
To me, without doing further evaluation, this feels so Freakonomics-coded. A matter-of-fact statement of popular belief/stereotype, with no critical evaluation of that position, sets up the straw man; and the wise sages who exist “in the real world of data” will tear that down.
It’s exactly the type of contrarian just-so story that I expect to seep into the popular consciousness via NPR stories, Ted talks, op-eds, social media, etc. while being wholly unsupported by whatever “real world of data” they claim to live in.
I’ve no evaluation of the actual claim, either direction seems plausible at first blush, but the framing primes me to skepticism.
Different:
I get what you’re saying, but this sounds more reverse-Freakonomics coded. The Freakonomics version would be something like, “It’s not politically correct to say so, but in the real world of data, it turns out that traditional stereotypes of men and women are correct.”
Exactly the same stuck out to me! What a shocking phrase, “the real world of data.” “The real world of the statistical model.” “The real world of the textbook.” “The real world of my interpretation.”
I must be above the 99th percentile of dataphilia in the world—at least upper half in the scientific world. I could never mistake it for the real world. I have whiled a long time in the field, experiencing and absorbing the sense of place. I have also abstracted the field to a few columns on a spreadsheet. Ain’t never confused the spreadsheet for the real world.
I think the following needs to be dealt with for these studies to be taken seriously (several of these have been mentioned by Andrew or others):
– divorced people has to be its own category
– how long has the couple been married and related to that, the age of the people
– the fact that every married person has passed through a state of being single
– in general, people are circulating between single, dating, married, divorced states over time
The comments and answers to your question have been thorough and well thought out. I would just add a few global observations, based on my readings and analysis of both cross-sectional and panel data over the years.
On average, people who are partnered feel better off and appear, by social indicators, better off than those who are not. Expressions of loneliness, for example, are highly associated with having a partner or not. Suicide, as a gross indicator, also points to it being better to be partnered. As to gender differences, I think the most miserable people are not-partnered males, particularly older ones.
On a topic like this, current marriage happiness, it is hard to resist this from https://forward.com/culture/576274/jewish-jokes-about-love-funny-marriage-dating/
———————————————
Preparing for the worst
Miriam and Yossi were married for 75 years when Yossi was on his deathbed. In the back bedroom, Yossi lay, listening to the sounds in the house around him. Soon, he started to smell something delicious coming from the kitchen. It was his favorite chocolate strudel.
His mouth began to water and he smiled, knowing that his beloved wife Miriam was making his favorite food for him to enjoy one last time.
It took all his energy to hobble from the bed into the kitchen, where he spied his wife working on the strudel. She stepped away for a moment to the refrigerator to retrieve something and Yossi reached out his hand to touch the delectable treat on the kitchen counter.
Suddenly, he felt a slap on his wrist.
“Yossi!” his wife cried out. “It’s for the shiva.”
——————————————————————————-
Is the above joke funny? Is it understandable to this audience? Is it offensive? Who is “allowed” to tell this joke? Would any of the answers have been different if the issue were raised two-three decades ago? Let me put it this way:
iz di aoybn vits modne? iz es farshteyik tsu dem oylem? iz es afensiv? ver iz “derloybt” tsu dertseyln dem vits? volt eyner fun di entfers geven andersh aoyb di aroysgebn iz aoyfgeshtanen tsvey-dray yortsendling tsurik?
It’s funny to me, but then I’m part of the “in” group for that joke. My wife would hate it, but that’s because she doesn’t find anything about death funny.
The question of who is “allowed” to tell this joke is a good question, and was being discussed as a question about this type of joke 20-30 years ago. I think it’s not a question of “who can tell the joke”, but “to what audience is the joke being told”?
Interaction effects!
I think Robin Morris raise a valid point about jokes. The speaker and the audience are involved in a delicate dance. I too am “part of the ‘in’ group” but I often overlook or am insensitive to my audience. Back in antiquity when I was employed in the classroom, I tended to “read the room” incorrectly. In fact, I was once accused of being antisemitic but was unanimously declared innocent by a faculty committee chaired by a dean; hence, unlike others so accused, I have actual written proof that I am not an antisemite.
A man walks into a bar and says, “I have proof that I’m not an antisemite.”
The bartender raises an eyebrow and asks, “Oh yeah? What’s that?”
The man pulls out a piece of paper and says, “See? It’s a letter from the Anti-Defamation League. They say I’m not one of them!”
The same joke was on Prairie Home Companion, only it was Ole instead of Yossi; does that matter?
Hypothesis: Men live happier life than women because they marry older and die younger.
The Vox article linked to above says: “Dolan had misinterpreted one of the categories in the survey, “spouse absent,” which refers to married people whose partner is no longer living in their household, as meaning the spouse stepped out of the room.”
This led him to say “”Married people are happier than other population subgroups, but only when their spouse is in the room when they’re asked how happy they are. When the spouse is not present: f***ing miserable,” Dolan said, according to The Guardian.”
I find this a hilarious mistake.
Jonathan:
People are happier when their spouse is in the room—as long as the room is big enough!
Happiness, causality, marriage (and other things), partly in response to Claude: Several articles in the Economist reminded me of a quip in a short paperback I’d read years earlier:
“Married people are happier than unmarried people, but the unhappiest of all are those in bad marriages.”
Unfortunately … the book has disappeared from my shelves, and its title from my memory – google gives too great a haystack of hits.
Lots of interaction effects – economics, politics, historical setting. Great discussion topic for students.
The thing I found most interesting was the change in women’s answers to the question about whether married people are generally happier than unmarried people–in 1988, 52% agreed and 17% disagreed; in 2012 (the last year it was asked) 30% agreed and 35% disagreed. Although I didn’t do a thorough analysis, as I recall that movement was pretty uniform among all groups. Some survey should ask that question again.
“…ultimately, yes, I do care about causal effects of marriage…”
1. Do you really feel that this question is answerable in any sort of reliable way, methodologically speaking? If so, how?
2. Why? If you could somehow answer this question, what *concrete* impact on the world would you expect to make?
3. Do you foresee any possibility that suboptimally reliable studies on this topic could be co-opted by people with political agendas to influence public policy (e.g., discrimination against unmarried people)?
Anon:
Good questions! Here are my answers:
1. I’m not sure. As is well known in statistics, it is hard to reliably answer a casual question when there is no process model and no possibility of a controlled experiment. Any statements we could make about the causal effects of marriage would be very dependent on assumptions and restricted in their generalizability. One thing I find refreshing about most of the reporting on this topic is that it’s framed descriptively—comparisons of average happiness levels among different groups—rather than causally.
2. I care about the causal effects of marriage even if I’d never be able to reliably estimate such effects. Lots of people around the world, including people close to me, make decisions about whether or when to marry. As to your question about my impact on the world here, I’m not quite sure what you mean by “concrete,” so let me just say: These surveys exist; researchers, journalists, and policy makers will be comparing happiness and other outcomes of married and unmarried people; with writings such as the above post I’d like to help them think more clearly about such studies; and to the extent that the work in this area translates into policies (for example, tax benefits for married couples, or laws on marriage and civil unions), I think it can be helpful to consider a causal framework, while recognizing that any causal estimates will depend strongly on assumptions.
3. I guess that all studies in this area will be “suboptimally reliable.” I expect that high-quality and low-quality studies alike can be used by people with political agendas. Sometimes I don’t mind, for example when it’s a political agenda that I agree with. More generally, this is an ongoing concern in research. Develop a new chemical technique, it can be used to build a bomb. Uncover a principle of biology, it can be used to design a killer virus. Improve Bayesian workflow, it can be used to more effectively spy on people. Etc. I don’t have any general answer to this one. I’d like to believe that a better understanding of social processes will ultimately result in social progress, but it could well be that the net effect in this world of my research and writings is to empower bad guys to do bad things more effectively.
Thanks for your detailed reply!
I’ll just poke a little harder on point #2 in your response:
“…and to the extent that the work in this area translates into policies (for example, tax benefits for married couples, or laws on marriage and civil unions), I think it can be helpful to consider a causal framework.”
Hypothetical scenario: I have managed (somehow) to design a reliable study showing that marriage causes “X” (e.g. happier people, longer life, greater per capita income, less violent crime…). Therefore, I will design a policy that does “Y” (fill in policy decision). Can you imagine any combination of variables “X” and “Y” that a large segment of the population wouldn’t find objectionable? I’m having trouble answering this question…
Anon:
Sure, many people might find it objectionable—but that doesn’t mean it shouldn’t be done. What if you’re doing research on how to make abortion safer? Many people might find that objectionable too. This relates to my point #3, that methods can be used for all sorts of purposes.
I think we’ll have to agree to disagree here- that’s okay. There are methodologically rigorous/reliable ways to study the safety of abortion techniques – arguably this is not the case for questions around the “causal” effects of marriage.
It’s not the subject matter that makes some types of research inadvisable. Rather, it’s the degree of mismatch between the rigour of the research method being applied and the societal consequences of getting the wrong answer. Poor rigour generates unreliable results. And unreliable results, in the hands of political ideologues (of any stripe), can do a lot of harm.
Forget about marriage, ask about happiness levels of people with children. And I mean children, not teenage or adult descendants. Children. Below 8 to 10 years old. Talk to the parents that leave the house for the sole reason of having to visit a playground 3 to 5 days a week. And definitely differentiate between the hobbyists(1 child) and the actually committed ones (2 or more).
A:
This is the sort of reason that researchers sometimes distinguish “happiness” from “life satisfaction.”
That makes sense.
And if I am reading “A poor soul”‘s comment correctly (parents are less happy) that is pretty bad news for society long term. If that became widely known…well birthrates are already below replacement in nearly all (perhaps all) developed nations and a growing number of middle income ones (including some with reputations for large families like Mexico and even India).
Immigration can only sustain populations so long if the source countries themselves have below replacement birthrates.
I doubt it’s not widely known already. The term “Terrible twos” should be a clue.
It might be one of those open secrets that everybody who knows doesn’t really talk about it openly because:
a) It wouldn’t really help them socially. In many cultures children are referred to as “joy/happiness/god’s gift”. Both literally or as an antiphrasis. It might backfire if one says publicly they are unhappy since they had children
b) It wouldn’t help them inside their family (which child/spouse would want to hear such a thing?)
c) There isn’t really much you can do beside accepting it. There are worse fates and the worst part of it lasts less than a decade in most cases (2 children or less)
Once you are in the club and others in the club can relate to you, they share experiences (mostly privately or in small groups). I once heard a parent of twins saying “For 4 years, we didn’t have a life.”. I didn’t dare ask about their life satisfaction level.
As for the falling birthrates, that’s a big discussion and not only related to happiness/life satisfaction. I ‘ll just anecdotally note whenever I hear the offspring of really rich people, I tend to count more than 2 children. It’s as if money makes this easier, go figure :P
True, although I ‘ll say I have some doubts about the efficacy of such distinctions with such subjective and self reported things such as “happiness” or “life satisfaction”.
On the causality question — the causality question is everything that matters about this question, IMO.
There is an old trick by marriage promoters, which goes like this:
1. There are many studies showing correlation between marriage and well-being.
2. Of course, there are complexities establishing causality so you can’t judge from one study.
3. But there are SO MANY correlational studies that there MUST be something there.
4. Thus, marriage is good and we need more of it.
The lie is that adding more bad studies strengthens the case. This is the strategy employed by Melissa Kearney in her recent book, even though her OWN STUDY showed the benefits of marriage were concentrated among those most likely to marry.
I once used the analogous example of “cars are good for children” in a post that should strong correlations between cars and child wellbeing across countries, states, and US households. The correlational evidence is overwhelming! https://familyinequality.wordpress.com/2014/06/17/data-stories/
I also wrote about this on Kearney: shorter, paywalled (https://journals.sagepub.com/doi/10.1177/00943061241240874d); longer, open (https://osf.io/preprints/socarxiv/jt9ad)
Thanks for linking to your paper- it’s incredibly incisive. I’m “Anonymous” from the Aug 28 post above. I’m a non-U.S. physician with a good friend who works as a trauma surgeon in the U.S. Her stories about the lack of a U.S. social safety net are mortifying. So I get really agitated when I hear researchers suggest that the solution to U.S. social woes is more marriage and church attendance. No! The solution is for U.S. governments to spend more to take care of the less fortunate, bringing them in line with other developed countries (Figure 1 in your paper says it all…).
You articulated so clearly the reason why I get the heebie-jeebies when I hear about the Templeton Foundation hurling vast sums of money at prominent Harvard University researchers to identify “causal” roles for marriage and religion in “human flourishing.” It’s blatantly obvious to anyone with two neurons to rub together that promotion of conservative ideology, not societal improvement, is at the root of these efforts. These initiatives can do real harm by distracting from the real problem- U.S. economic inequality.
I don’t think a European style social safety net is politically feasible in the US. In 2016 people were framing Trump’s unexpected victory as “a referendum on the Affordable Care Act”. You’d need a President with control of both houses of Congress and a willingness to burn all their political capital on the issue – and to accept the risk that it would tip the next election the other way.
The nature of the US federal system and the (I think uniquely among developed nations) strong individualist streak of much of the US makes it very hard here.
It’s possible that a potential Harris administration could surprise me, but that would probably require something more than a 50-50 Senate and narrow House control, which doesn’t seem the way to bet right now.
Dramatic changes of any kind seem very unlikely to me in the current very close political environment, whoever wins.
And I don’t know how broad the time window is. The social safety nets we do have (Social Security, Medicare, Medicaid, welfare) may become harder to maintain as the population ages and the worker/retiree ratio shifts.
Philip:
I have a post coming up at some point regarding your dispute with Kearney . . . Regarding your Jennifer-Hill-esque statement that “the causality question is everything that matters about this question”: sure, when it comes to policy, I agree with you. I also think that purely descriptive, non-causal work can be valuable as an attempt to establish the “facts on the ground” that motivates causal theorizing. So often I see the following sort of reasoning:
1. Descriptive study finds pattern X.
2. X is treated as a stylized fact, motivating causal theory Y.
3. Causal theory Y is used to support policy recommendation Z.
And then it turns out that step 1 is flawed: the descriptive evidence is not so clear.
Phillip –
the causality question is everything that matters about this question, IMO.
I agree. As I said above, I don’t see how the correlation and the causation can be meaningfully separated here.
As an example, I think that in this context the cross-sectional correlational data – particularly when it’s not adjusted for things like stage of life, or if there are children, or SES – are useless. What matters are the trajectories, and the story they tell given interacting variables.
What might it might mean if there’s strong correlational, cross sectional evidence that happier people are more likely to get married? But how happy were they before they got married? Are they less happy after getting married?
You can’t describe causality without some form of correlational data, but correlational data that stands on an island can’t tell a story. Yes, it’s important to ascertain if the correlational data are wrong, but it’s also important if they’re useless.
Self-reported “happiness” in these surveys can be pretty worthless, especially since categorization is always changing between singledom and partnership, but djacent data can tell quite the story. Like how there’s likely a major orgasm gap for straight/bi women with men. And how women initiate divorce more than men. Or how men are more likely to leave women when they have cancer.
The number one reason married people may report greater happiness, in my opinion, is that they’re least likely to be poor: https://www.ssa.gov/policy/docs/population-profiles/marital-status-poverty.html. At the very least, this supremely complicates self-reported happiness levels when categorized by partnership status.
What if there are heterogeneous treatment effects, so that people who benefit more from marriage are more likely to get married? Perhaps marriage is good for the married but not so good for the not married, because the two groups have sorted based on their treatment effect of marriage?
I make this argument (in a bit more detail) around page 12 of this paper
https://docs.iza.org/dp15151.pdf
in the context of a discussion of the Waite and Gallagher (2002) The Case for Marriage book.