Getting Married Makes You Happier? Again, No, Part 1

We believe in marriage more than data

Posted Jan 09, 2015

We are a nation of matrimaniacs, and it seems like too many of us just can't get enough of studies claiming that married people win. The latest study claiming that marriage makes you happier hit the New York Times (and many other media outlets) and immediately climbed to the number 1 spot on the most-read articles.

The article in question is bolder than most. The authors, Shawn Grover and John Helliwell, know it is challenging to claim that marriage causes people to become happier, but they think they can make the case anyway. (By the way, the article is just a working paper—it has not been published in a professional journal, meaning it has not been reviewed by other scholars and deemed worthy of publication.)

I have a lot to say about this unpublished working paper in Part 2, so let me start here in Part 1 by telling you my bottom line about this study and every other study making the same claim, including every study that will ever be done: No study has ever shown definitively that getting married causes people to become happier, and no study ever will.

I will explain my general case in this post. Then in my next post, Part 2, I will critique the Grover and Helliwell research in detail.

Why no study can ever show definitively that getting married will make you happier

The current authors—and all the other people who try to make the same kinds of claims—want to compare married people to single people, and if they find that the married people look better in some way (say, they are happier), then conclude that if single people would only get married, they would get happier, too. But there is a big problem that can never be fully surmounted: The married people and the single people are different people. Just because the people who chose to get married got happier (if they really did—a key question I will address in Part 2) does not mean that people who like being single and choose that life would be happier if they instead chose to get married.

Think of it this way: Suppose you found, hypothetically, that people who become accountants are happier than those who become poets. Should all the poets then set their imaginations aside and set up shop as accountants—and expect to become happier as a result? It's ridiculous.

Because married people and single people are different people, any way that they are different could account for any differences in happiness (again, if such differences really do exist), meaning that marriage might have nothing whatsoever to do with the happiness differences. For example, maybe the people who get married are already happier even before they marry—in which case marriage may be irrelevant. Or maybe they differ in other ways associated with happiness. For example, because legal marriage comes with a treasure trove of legal and financial benefits, as well as discount rates on just about everything, married people may be less likely to be impoverished and that could make them happier. In that case, it's not the marriage that matters but being able to afford food and clothes and shelter.

The smarter researchers realize these things, and they try to take them into account in this way or that, usually using statistical fixes. But the statistical "controls" are just a way of trying to make the best of an impossible situation. Plus, importantly, no one can think of all of the possible ways in which married and single people might differ that could account for any differences in happiness (or anything else). If they can't think of them, they can't control for them at all. And—this will not surprise readers of this blog—marriage researchers don't consider ways that married and single people differ that reflect positively on single people. I have been reading the research on marital status for nearly two decades, and I don't think I have ever seen a researcher acknowledge that some people like their single lives and choose to live single, and for them, getting married would not make them happier no matter how many married people have tried marriage and liked it. (Considering the divorce rate, it is not as many as the matrimaniacs would have us believe.)

Here's why the goal of showing that getting married causes people to be happier is an impossible one to achieve: The gold standard for determining causal relationships is a true experiment, in which people are randomly assigned to different conditions. So if you really want to know if getting married causes people to become happier, you need to assign people at random (literally—like picking marital status out of a hat) to get married or stay single, and then randomly assign the married people to stay married or get divorced, or heck, become a widow. If you have enough people in your study, then the people who, by chance, end up married are not going to differ, on the average, in any way from the people assigned to stay single. They are not going to be happier or wealthier or healthier or more sociable or more interested in being married. On the average, they will be the same, and so any differences in their happiness will be due to marriage. And those differences won't necessarily favor those assigned to be married, as is so often assumed by matrimaniacs.

(It's more complicated than that. The groups would be the same at the start of the study, but marriage could still benefit the married people in ways that have nothing to do with the marital relationship itself if, for example, married people got more benefits and protections and more money, and more validation and celebration.)

The two main kinds of studies used to address the question of whether getting married makes you happier are cross-sectional studies, in which people are compared at just one point in time, and longitudinal studies, in which the same people are followed over time. Cross-sectional studies are by far the weaker of the two—it is really a stretch to compare people at just one point in time and then try to declare that the people in one of the groups (say, the married people) are happier because they are married. No matter how impressive your statistical gymnastics, you are still dancing around the fact that you just don't know why the two groups differ.

Longitudinal research is better because you can follow the same people over time and see whether, for example, people generally become happier after they get married than they were when they were single. In theory, you don't need to worry about ways in which the single people differ from the married people because they are the same people—you are watching how their happiness changes as they transition from single to married.

In fact, though, it's more complicated than that. When single people get married, they do change in ways that could benefit them but have nothing to do with the marital relationship. For example, they get tons more money in federal benefits, they get lots of consumer deals and discounts, and (if they are going from living alone to living together) they benefit from the economies of scale (e.g., one rent and one set of utilities is split two ways). Maybe if people who get married get happier, it is because they have more disposable income, and not because of, say, their relationship with their spouse.

Or maybe they are happier (if they are happier, which I will address in Part 2) because marriage and married people are celebrated, valued, and validated, and single life and single people are stereotyped, stigmatized, discriminated against, marginalized, and often just plain ignored. Maybe if the single life and single people were not targets of all that singlism and had just as much respect as married people and marriage, then it would not matter who did or did not get married. Everyone could choose their own path and happiness would follow. (That's the bottom line of what I really think, though there is no good research that tests it. Some people really are better off married; other people really are better off single—that's how they live their most authentic and meaningful life, despite all the ways in which the deck is stacked against them for living single.)

The other problem with longitudinal studies is that if you want to compare people who get married to people who stay single, you are back to all the same problems you have with cross-sectional research. The people who get married are different people than the people who stay single. Just because the people who choose to marry become happier (if they do) does not mean that people who choose to live single would be happier if they were forced into marrying, or bludgeoned by a posse of deluded and matrimaniacal pundits, reporters, and social scientists into thinking that marriage would make them happier.

In my next post, I'll get into the details of the particular unpublished study that is getting so much attention. The authors did take some steps that got them closer to a causal conclusion than many of their predecessors, but the proper conclusions from their research are not at all what their most enthusiastic media cheerleaders would lead you to believe.

Stay tuned for Part 2. (Here it is.) In the meantime, if you want to read other critiques of other claims about getting married and getting happy, click here. Or if you want to read critiques of claims that getting married makes you healthier or live longer or have better sex or have more social connections, you can find those critiques here. Also, chapter 2 of Singled Out is always a good start for thinking about how to assess claims about marital status and other life outcomes.

Notes: Thanks to Erin Albert, Kim Calvert, Carol Hynson, and Elizabeth Saenger for the heads-up about this latest bout of matrimania. Image is from Google Images, labeled for reuse.