Verified by 
Here are some simple questions: If you get married, will you get healthier? Will you get happier?
Researchers have been studying marriage for more than a half-century, and we have come to believe that they’ve answered those questions with a resounding yes. In an op-ed that was published in the New York Times, I challenged our cherished beliefs about the transformative power of marriage. The best possible studies don’t support the simple fortune-cookie message, “Get married, get healthy.” And the reason we think there is a vast body of research supporting claims that marriage makes us healthier and happier is that there is a vast body of research that has been interpreted as supporting those claims. In fact, though, that research relies heavily on what I will call the “cheater technique” that gives marriage an enormous unfair advantage. When studies are analyzed and reported in such a biased way, it doesn’t matter if there are 10 such studies or 10,000: We can’t rely on their results.
Without using the phrase “cheater technique” in my op-ed, I described the problem very succinctly. Many readers immediately understood the problem. But many others did not, including some who have been doing marriage research and marriage advocacy for many years. They have been critiquing my op-ed, and one of those critiques was just republished in full in the National Review.
That motivated me to step back and try to explain, once again, what I mean by the cheater technique. I’ve also explained some of the other enduring problems in the study of marriage. In addition, I’ve addressed the specific critique touted by the National Review.
What you will find below is quite long for a blog post. Think of it as a little book that you are getting for free.
I will give you a preview of the various sections in case you want to just skip to the ones that interest you most. But don’t skip the first one, that explains the cheater technique, unless you already understand it from my op-ed or many of my previous writings.
PREVIEW
#1 The Cheater Technique in Studies of the Supposed Benefits of Getting Married
The cheater technique is the most important thing to understand when trying to figure out what we really do know about the supposed benefits of getting married. It is easy to make marriage look good by using this technique.
#2 The Problem with Marriage Research that Can Never Be Fully Resolved
Even when the cheater technique is not used, it can be dicey to try to make the case that marriage causes people to be healthier or happier.
#3 Longitudinal Studies Are Better than Cross-Sectional Ones, But You Still Don’t Get to Cheat
Most of the very best studies are still biased in ways that favor marriage.
#4 Institute for Family Studies Responds by Continuing to Use the Cheater Technique, and the National Review Says, “Me, Too!”
This is the heart of my response to the critiques of my op-ed. But it is important to read the section on the cheater technique (#1) before reading this section.
Brief summary of my response: The National Review’s critique of my op-ed leans heavily on studies using the cheater technique, makes unwarranted generalizations, includes mistakes of logic, and misleads about the results of previous research.
#5 Beware of Sloppiness in Descriptions and Reviews of Previous Research
When you see claims about the benefits of marriage, check them against the original research reports.
#6 About that “Harvard Professor”
Just a light interlude.
#7 My Critics Are Protesting but Also Making Some Concessions
In pushing back against my op-ed, critics made some concessions I do not often see in writings by marriage advocates.
#8 Want to Continue Studying and Discussing Marriage? There Are Lots of Important, Unanswered Questions
#9 Resources and Further Reading
**************************************
Okay, back to the beginning.
Here are the questions I am addressing: “If you get married, will you get healthier? Will you get happier?”
In an op-ed that was published in the New York Times, I challenged the claims that getting married makes people healthier and happier. I know you think you have heard those claims a million times, and you have. You also think they are true and they are based on science. The “science,” though, is massively flawed.
The comparisons that are most often made to try to show that marriage makes people healthier and happier just do not pass methodological muster. Here, I will call the use of those inappropriate comparisons the “cheater technique.” It is a way of giving marriage a massive, unfair advantage. By using this methodological sleight-of-hand that would never be accepted by any reputable academic journal if the topic were anything but marriage, marriage advocates have succeeded in convincing us that fairy tales really are true. Get married, they promise, and you will live happily ever after. And you will be healthier, too.
If studies of marriage were analyzed properly, the supposed benefits of marrying would be greatly diminished. Sometimes they would disappear entirely. Sometimes we would find just the opposite of what we’ve been told all along about how people who marry do better. In fact, even when studies use that cheater technique that puts a big, fat thumb on the scale, weighing it down on the side of marriage, sometimes it is the lifelong single people are doing the best. (I discussed some of those studies in Marriage vs. Single Life: How Science and the Media Got It So Wrong.) I don’t see those studies mentioned very often in writings about marriage.
#1 The Cheater Technique in Studies of the Supposed Benefits of Getting Married
I’m going to give you a few examples of the cheater technique as it would apply to hypothetical examples. Then I will come back to its relevance to answering the questions, “If you get married, will you get healthier? Will you get happier?”
Hypothetical Example #1
Imagine that members of a financial consulting company make money by persuading clients to invest in startups. To make their case, they tabulate data only from successful startups, and ignore all the others. Facebook, for example, would be included in their data; Pets.com would be set aside. Their results look amazing! The startups are doing great.
That would be cheating. Of course the startups look good – the consultants only counted the ones that are still in business. They are not giving their clients a fair and accurate prediction of whether money put into a startup would be a good investment.
In an even more egregious practice, some members of the consulting firm compare the money made by the startups that are still in business with the money made by the startups that are no longer in business. The currently operating startups look better, and on that basis, the consultants tell their clients to invest in startups.
You would fire those consultants. So would I. This is not hard to understand.
Hypothetical Example #2
Imagine that you are a drug company testing a new drug. You let people decide for themselves whether to take the drug. More than 40 percent of the people who take the drug refuse to keep taking it, usually because they had such terrible experiences with it. After quitting the drug, they felt worse than they did before they started taking it. The drug company decides not to include the data from those people when they make claims about the effectiveness of their new drug. They only count the people who are currently taking the drug.
The people currently taking the drug are doing better than the people not taking the drug, so the drug company advertises that their drug is effective.
Sometimes the drug company does something even more outrageous. They compare the people currently taking the drug to the people who started taking it but refused to continue because it made them feel so terrible. (Or, they put the people who refused to continue taking the drug in with the people who never did take the drug.)
They find that the people currently taking the drug are doing better than the people not currently taking the drug. On that basis, they tell consumers that their drug is amazing and everyone should take it.
Their drug made many people feel sick – worse than they felt before they tried the drug. The drug company is counting that sickness as evidence that everyone should take their drug! After all, the people currently taking the drug feel better than the people who are no longer taking it, because it made them feel worse.
Would you take that drug? What opinion would you have of the drug company that tried to reel you in with that argument?
Hypothetical Example #3
You are a researcher. You come up with a program designed to improve people’s lives. You let people decide whether they want to try the program. More than 40 percent of the people who try the program refuse to stay in it, usually because it made them feel so terrible.
When you analyze your data, you include in your key group only the results for the people who are still in your program. You set aside the people who refused to continue.
Or, you compare the people who are still in your program to those who refused to stay in it. Then, if the people currently in the program are more satisfied with their lives than those who refused to stay in the program, you declare your program successful. You boast about your findings, and enthusiastically encourage everyone to sign up for your program.
Would you sign up for that program? Do you think the editor of a prestigious journal (or even a mediocre journal) would accept that study, with those conclusions, for publication?
Example #4: The real one, about getting married and getting healthier or happier.
The question is: If you get married, will you get healthier? Will you get happier?
In real life, people get to decide whether to get married. More than 40 percent of them refuse to stay married, usually because their marriage turned out to be such a bad experience. People who divorce often end up feeling worse (less healthy, less satisfied with their lives) than they did when they were single.
Researchers set aside those 40 percent. They do not include them in the marriage group.
Or, even worse: They compare the people who are currently married to the people who are divorced. Or they compare the people who are currently married to everyone who is not currently married – including both the previously married and the people who never married.
Sometimes they find that the currently married people are doing better. Then they say: See, marriage makes people healthier and happier. Single people should get married. If they do, they will get healthier and happier.
This is the cheater technique in research on marriage. It is what thousands of studies of marriage look like. It is easy to make marriage look good if you ignore all the people whose experiences were bad enough to warrant getting divorced. That’s what is happening nearly all the time when you hear about the results of the latest research. It is why you think that marriage makes people healthier and happier and better off in other ways, too.
(If it is still not clear why the cheater technique is wrong, find an introductory-level book on research methods and look up the section on attrition.)
What researchers really need to do is to compare everyone who ever got married (whether they stayed married or not) to people who never married. That almost never happens.
By using the term “cheater technique,” I am not making a statement about whether the people who use this technique are knowingly cheating. It is possible that they really think that skimming the currently-married people off the top of the bigger group of people who ever got married is a fine way to reason and do research.
#2 The Problem with Marriage Research that Can Never Be Fully Resolved
Suppose researchers get a large, nationally representative sample of adults and they compare all the people who ever got married (those who are currently married, as well as those who were previously married) to those who never got married. That’s not the cheater technique. That’s an appropriate comparison.
Now suppose they find that the marriage group is doing better than the never-married group. Does that mean that if the single people get married, they will do better, too?
Not necessarily. Do you see why?
The people who choose to marry are different people than the people who stay single. Some of the people who stay single have chosen single life. They embrace what it has to offer. They have different values and preferences than the people who choose to marry. If they were badgered into marrying, they might not do very well at all. In fact, they might do worse than if they just got to live the life they chose.
Researchers typically address this kind of problem by using random assignment. By assigning people at random to different conditions (for example, a drug condition vs. a placebo condition), chances are that the people in the different groups will be about the same in every way except that some are taking the real drug and the others are taking the placebo. For example, on the average, the people in the two groups will have similar personality characteristics, similar preferences, and similar overall health. If one group ends up doing better than the other, it is probably because of the kind of drug they were assigned to take.
We cannot randomly assign people to get married or stay single. So when we find, for example, that the people who get married have different life outcomes than the people who stay single (whether better or worse), we can never really know for sure that their outcomes were different because of their marital status. Maybe the differences can be explained by some other way that the two groups differ. (Single people, for example, are more likely to be stereotyped and stigmatized than people who marry. They are also targets of discrimination and end up with less money because of that.)
#3 Longitudinal Studies Are Better than Cross-Sectional Ones, But You Still Don’t Get to Cheat
The kinds of studies I have been describing are “cross-sectional”: They compare different groups of people (such as people who are currently married and people who are not currently married) at one point in time.
A different kind of study is “longitudinal,” in which the same people are followed over time. That way, you can see if people’s health or happiness changes as they go from being single to getting married, or from being married to getting divorced. Longitudinal studies are better than cross-sectional studies, because the same people are being compared at different points in time. Some things about them, such as their personality and their preferences, are likely to stay relatively constant over time. So if they become more or less healthy or happy when they marry, or when they divorce, maybe it is because they married or got divorced (though we still can’t be completely sure).
Most longitudinal studies, though, use the same sort of cheater technique that is commonplace in cross-sectional research. Specifically, researchers usually look only at the people who get married and stay married. If the people who get married and stay married are healthier or happier than they were when they were single, then sometimes single people are told: See, if you get married, you will be healthier and happier, too.
By now, you probably understand why you should be skeptical about that conclusion.
Sometimes researchers look separately at the people who get married and stay married, and the people who get married and then divorce. The results are telling.
In Singled Out, I showed the results of one such study that has been ongoing for many years. The people who got married and stayed married became a little more satisfied with their lives when they first got married. But it was just a “honeymoon effect” – their life satisfaction slipped over time, and they ended up about as satisfied or dissatisfied as they were when they were single.
That was the best possible outcome. The results were different for the people who got married and then divorced. They were already becoming less satisfied with their lives as the day of their wedding drew near. Then they continued to become less and less satisfied over the course of their marriage, and ended up less satisfied than they were when they were single.
In the majority (probably the vast majority) of longitudinal studies, only the people who got married and stayed married are included in the marriage group. But remember the questions we are addressing: “If you get married, will you get healthier? Will you get happier?” You can’t answer those questions by looking only at the people who get married and stay married. That’s cheating.
In my op-ed, I referred to a review of 18 longitudinal studies of well-being (happiness, life satisfaction, and relationship satisfaction). Those studies showed that people who got married did not generally become happier or more satisfied than when they were single. At best, they just enjoyed that brief honeymoon effect.
What I didn’t mention in my op-ed is that most of those 18 studies were biased by including only those people who got married and stayed married. Even with that bias that gave marriage a giant unfair advantage, people who got married did not end up lastingly better off than when they were single. In most cases, they did not even enjoy a brief honeymoon effect. Getting married just did not do anything for their happiness or relationship satisfaction.
When you read a claim that says, “Look, we are basing our conclusions on longitudinal research, so you have to believe us,” beware. Check out the studies to see if they are using the cheater technique. The cheater technique makes marriage look better than it really is. If studies using the cheater technique still don’t show impressive results for people who marry, there’s probably something really wrong with the claim that if you get married, you will get healthier and happier.
#4 Institute for Family Studies Responds by Continuing to Use the Cheater Technique, and the National Review Says, “Me, Too!”
A few days after my op-ed appeared in the New York Times, the Institute for Family Studies, a group “dedicated to strengthening marriage and family life,” published a critique by Tyler VanderWeele.
That critique has been highlighted by articles in other publications such as the Christian Post and PJ Media (a company that describes itself as having “a special connection to the values which make America special, as well as a dedication to keeping America great”). It has also been reprinted in full in the National Review.
Are There Lots of Studies Supposedly Showing the Magic of Marriage? Yes, If You Use the Cheater Technique
The author of the critique cited one study after another that used the cheater technique. He found lots of studies in which marriage seems to result in benefits to psychological and physical health – as long as you include in the marriage group only those people who are currently married. He even included graphs of the results of some of them.
There are even more studies like the ones the author cited. I’ve been writing about them for two decades. But it doesn’t matter if you have a thousand studies or ten thousand studies that seem to show that marriage makes people healthier or happier if all those studies are based on the cheater technique.
When Even the Cheater Technique Does Not Produce Clear Evidence of the Supposed Benefits of Marriage
The VanderWeele critique features some sweeping statements about the health benefits of marriage. In support of the supposedly “strong evidence for the protective effects of marriage on health…,” the author references two reviews, both from a decade or so ago, and not published in peer-reviewed journals. I looked at the sections on physical health in the more recent one. All the studies used the cheater technique. Nonetheless, the results are qualified rather than strong or sweeping. For example, one of the studies finds results for men but not women, and another finds results for women but not men. Other studies, using the most egregious version of the cheater technique, compared married people to divorced or widowed people. Results were again qualified – for example, showing differences only for some of the years of the study and not others.
The review does not mention a study showing that, even using the cheater technique, lifelong single people are healthier than everyone else. Also, as I noted in Marriage vs. Single Life: How Science and the Media Got It So Wrong, in research on marriage, even “when the married group appears to do best, often that edge is qualified by factors such as whether the adults are men or women, Black or White, younger or older, and whether the married people married recently or not so recently.”
The National Review is selling the fortune cookie version of marriage: “Get married, get healthy.” The reality is not so simple, even when studies give marriage an unfair advantage by using the cheater technique.
Did I “Sidestep” Findings About Divorce? No, I Emphasized Them
In my op-ed, I described the results of a 16-year longitudinal study of more than 11,000 Swiss adults. (Read the whole thing – it’s short.) The study, by Professor Matthijs Kalmijn, showed that people who got married rated their overall health as slightly worse than it was when they were single. And, their health continued to deteriorate a bit over the course of their marriage. As for the number of illnesses they had, that did not change much as people got married or stayed married, though people did have somewhat more illnesses after they divorced than they did before.
The participants were also asked every year about their satisfaction with their lives. People who married did become a bit more satisfied with their lives when they first got married (a small effect); then, over the course of their marriage, they became less and less satisfied, though their satisfaction decreased more slowly than in most previous studies of marriage.
My op-ed mentioned a review of 18 previous longitudinal studies of well-being (happiness, life satisfaction, and relationship satisfaction). As I noted previously, those studies showed that people who got married did not generally become happier or more satisfied than when they were single. At best, they just enjoyed a brief honeymoon effect.
I also said this about the results of the Swiss study:
“Dr. Kalmijn also examined the implications of divorce and found that people who divorced became significantly less satisfied with their lives. In fact, the negative implications of divorce for life satisfaction were more than three times greater than the positive implications of marrying.”
Look at what VanderWeele said about the Swiss study:
“The study also notes even larger effects of moving into marriage on increasing life satisfaction, as well as larger effects of divorce on decreasing life satisfaction. The New York Times article seems to try to sidestep these results…”
I didn’t sidestep those results; I emphasized them.
The Mindset of Marriage Advocates
The author seems to have the same mindset when he points to his graph showing that people who get divorced don’t live as long as people who stay married. He is proud of those findings. So are other people, such as Brad Wilcox, the Director of the National Marriage Project, who tweeted them out and tagged me on his tweet.
Here’s what’s going on. Tyler VanderWeele and Brad Wilcox think that the worse life satisfaction of people who got divorced, compared to when they were married, is evidence for how great marriage is. They think that if people who get divorced don’t live as long as people who stay married, that just shows the wonders of marriage. They are the consultants telling you to invest in start-ups because the startups still in business made more money than the startups that went out of business. They are the representatives from the drug company telling you to take their drug because the people currently taking the drug are doing better than all those people who took the drug and felt worse than they did before they tried it.
The implication seems to be that people should just stay married instead of getting divorced. Here’s an example of VanderWeele’s reasoning:
One study indicated that among those who were married and rated their marriage as “very unhappy” but stayed married, 77 percent said that five years later the same marriage was either “very happy” or “quite happy.”
If you have no training in research methods and you think that reasoning is convincing, that’s fine. But VanderWeele is a PhD and – as people touting his critique cannot say often enough (see below) – a professor at Harvard (the School of Public Health).
Go ahead and give it a try, even if you don’t have any special research training. See if you can figure out what’s wrong with taking that study (showing that many people who started out unhappy but stayed married ended up rating their marriage as happy five years later) and using it to imply that people should just stay married.
How about this? The unhappy married people who stayed married are different people than the unhappy married people who got divorced. Those who got divorced may have had different experiences in their marriages. To take an extreme – but real – example, maybe some of them were being beaten to a bloody pulp. If they stayed married, they would be less likely to end up happy than to end up dead.
There were problems with VanderWeele’s critique even beyond his reliance on the cheater technique. One problem – not unique to VanderWeele – is that some of what he said was exactly wrong. I’m not talking about interpretations, which can be subjective. He says that a study found one thing, when in fact it found exactly the opposite.
#5 Beware of Sloppiness in Descriptions and Reviews of Previous Research
For the first article I ever published about single life and the science of marriage, Wendy Morris and I looked at claims made by some of the most eminent scholars in the fields of marriage and positive psychology. They all made bold pronouncements about the supposed happiness advantage of marriage. We looked at the original research reports of the specific studies they referenced, and found again and again that the evidence was not what readers were led to believe.
I learned from that exercise not to take claims about marriage at face value. Always check them against the original research reports.
When I wrote Singled Out, I did careful checking of the stark claims made by Waite and Gallagher in their much-cited book, The Case for Marriage. The case they made was embarrassingly sloppy. What they said about specific studies all too often departed notably from what the studies really did show – always, of course, in the direction of making marriage look better than it really was. Take a look at Singled Out for some line-by-line, data-point by data-point debunking of the claims made in The Case for Marriage. VanderWeele leans heavily on that book in trying to refute my op-ed. But no scholar should ever cite The Case for Marriage without noting that it is filled with errors and misleading claims. Better still, don’t cite it at all. Study the original research reports.
VanderWeele’s critique also includes misrepresentations of what the studies he is reviewing actually did show. For example, he claims that marriage is “linked to a higher level of personal growth.” At best, there is a honeymoon effect. But then, over a five-year period, it is the lifelong people who experience more personal growth than the people who stayed married. And remember, that comparison of never-married people only to those who are currently married is one that unfairly favors marriage.
He also claims that marriage is linked to “higher levels of meaning and purpose,” but he got that wrong, too. Again, at best, the results show that there is a honeymoon effect when people first get married. Once lifelong single people and continuously married people settle into their lives, there is no significant difference in their experiences of meaning and purpose (even though such a comparison favors the married people). The biggest effects are for comparisons of currently married people to divorced people, or comparisons of people before and after they are widowed. The divorced and the widowed experience less meaning and purpose than the currently married people. (Again, VanderWeele seems to want to say that if divorced people do worse than married people, that shows how great marriage is. I think he is implying that they would have found more meaning by staying in their miserable marriages.)
There are other problems, too. For example, VanderWeele confounds marriage with parenting, but the best studies do not. If married people end up less healthy than they were when they were single, it is not because they are like VanderWeele in having an 18-month-old son who wakes them up at night.
#6 About that “Harvard Professor”
The writers who picked up VanderWeele’s critique and ran with it seem mighty impressed with the fact that he is a Harvard Professor. PJ Media mentions Harvard at least four times. The Christian Post mentions Harvard and “the Ivy League school.”
I think we should judge arguments on their merits and not on the credentials of the people making the arguments.
But okay, if you want to play that game, here’s something not mentioned in my New York Times article: My PhD was from…Harvard! My advisor was one of the most eminent methodologists in my field of social psychology. I have taught graduate courses in research methods for decades. I’ve published more than 100 scholarly papers, some of them in the most competitive and prestigious journals in psychology.
Still, judge my arguments by their merits.
#7 My Critics Are Protesting But Also Making Some Concessions
In saying that claims about the transformative power of marriage are greatly exaggerated or just plain false, I stepped on the toes of decades of social scientists, pundits, and marriage advocates who have made their careers out of insisting that people who marry are better people. The myth of marital superiority is supported by an entire industry of well-funded think tanks with clout in politics and the media. Scholars who study marriage can get big-time funding for their work. (Not all of them study or write about marriage in the ways I’ve described here.) They have many journals in which to publish their findings, and conferences where they can spread the word. They can get jobs in universities where they can teach courses using textbooks describing their research on marriage and get research assistants to help them with their work.
All those people were not just going to read my op-ed and say, “Hmm, I guess she is right.” I knew that.
The key questions, remember, are: “If you get married, will you get healthier? Will you get happier?” Over time, as new generations of thinkers and researchers arrive on the scene who are less steeped in the ideology of marriage, we are likely to start seeing more accurate answers.
In the meantime, I’m happy to see some baby steps, such as a few concessions. And already, in the critiques that have been written about my op-ed, there are some. For example:
“Single people have immense worth, and they unquestionably contribute to society in myriad ways.”
And from VanderWeele:
“…some of the earlier studies were indeed methodologically weak…”
“…the effect of marriage is not huge…”
“…research practices could certainly be improved…”
“…single people occupy an important place in society…”
It’s a start.
#8 Want to Continue Studying and Discussing Marriage? There Are Lots of Important, Unanswered Questions
Marriage researchers and marriage advocates have been telling us for decades that people who get married have many advantages over people who stay single. They include, for example, all that spousal love and support and buffering against whatever life throws at you. They also include the monitoring that couples do, to make sure their partners don’t do risky things like smoking or eating badly.
In the U.S., there are other very significant advantages we hear little about, such as the 1,000+ federal laws that benefit and protect only people who are legally married. Those laws include many that greatly favor married people financially. In addition, married people’s lives are respected and celebrated, whereas single people are more often stereotyped and stigmatized.
With all that spousal love and support and monitoring that only married people get, and with all their legal protections and cultural adulation, people who marry should do great. They really should end up much happier and healthier than they were when they were single, and their happiness and health should only increase over the course of their marriages, instead of what actually happens – their well-being decreases.
The misleading claims about the transformative power of marriage should be true. So why aren’t they?
I’m more interested in the positive version of that question, as it pertains to single people. How is it possible that so many single people -- who are stereotyped and stigmatized, who are living without the love and support of a spouse, and who do not get those 1,000+ federal benefits and protections earmarked solely for married people – are doing so well?
That’s a question worth answering.
Here’s one more. Research studies usually report results averaged across all participants. But the implications of getting married are not the same for everyone. I think that some people really do lead their best lives by getting married. There’s nothing new about that – it’s the conventional wisdom of our time. What I am adding that is new is the flip side of that: I also think there are people who lead their best, most meaningful, and most fulfilling lives by living single. How do we know who will do best by marrying and who will do best by living single? Researchers should address that.
#9 Resources and Further Reading
Marriage vs. Single Life: How Science and the Media Got It So Wrong
My TEDx talk: “What no one ever told you about people who are single”
A collection of articles on marital status and longevity
A collection of articles on marital status and health (including mental health, such as depression)
A collection of articles on marital status and happiness
Collections on other topics relevant to marriage and single life
