The_Science_of_Marri_Cover_for_Kindle, 2-28-15

[This post was originally published at Psychology Today. I just discovered that it disappeared! I have no idea why, but I thought I’d just republish it at my own site where I have control over what appears and disappears.]

In my previous post, I explained why no study has ever shown definitively that getting married causes people to become happier – and no study ever will. Here, I will critique the research (an unpublished working paper by Grover and Helliwell) that set off the latest round of matrimaniacal claims that we single people would be happier if only we would get married. The claims the authors are making are unapologetically causal: They think their research shows that getting married causes people to become happier. It doesn’t. The very premise of their claim (that married people are happier, and we just need to figure out if marriage is causing married people’s greater happiness) is undermined by some of their own findings – not that you would have read much about those results in any of the many media stories gleefully declaring a win for Team Marriage.

Places Where Married People Are Not Happier Than Single People

The authors argue that too many of the studies of the implications of marrying have been conducted in “western, educated, industrialized, and rich democracies” – or WEIRD places, for short. They are right that research in places such as North America, Western Europe, Australia, and New Zealand is far more plentiful than research in other parts of the world. But there are relevant data from other places. Between 2005 and 2013, the Gallup World Poll collected life satisfaction ratings from many nations around the world. (“Happiness” studies are very often studies of life satisfaction.) The data are cross-sectional: married people were compared to not-married people at just one point in time.

Here are 5 findings I bet you did read much about in any of the barrage of articles or opinion pieces or blog posts about Grover and Helliwell’s research:

  1. In Latin America, single people are more satisfied with their lives than married people are.
  2. In the Caribbean, single people are more satisfied with their lives than married people are.
  3. In Sub-Saharan Africa, single people are more satisfied with their lives than married people are.
  4. In Southeast Asia, single people are just as satisfied with their lives as married people.
  5. In South Asia, single people are just as satisfied with their lives as married people.

All of these findings that belie the conventional wisdom that married people are happier than single people. What is especially telling is that the results come from comparisons that were already biased to advantage married people. The married group, so far as I can tell, includes only those people who are currently married. (I’ve emailed both authors to confirm this and I’ll update this post if I ever hear back.) That means that the findings are based on the cheater technique, whereby all of the people who got married and hated it are removed from the married group, making it easier to pretend to have shown that getting married makes you happier. But even with that big, unjustifiable advantage given to the married group, they still aren’t any happier than the single people (and sometimes significantly less happy) in Latin America, the Caribbean, Sub-Saharan Africa, Southeast Asia, and South Asia.

What About the Research Showing that Any Increase in Happiness After Marrying Is Just a Honeymoon Effect?

If you have followed the research on getting married and getting happier even just casually, you may remember some findings that have gotten a fair amount of attention in the past. They are from a German longitudinal study, analyzed by Lucas and his colleagues, in which the same people were followed for years, as they stayed single or got married or unmarried. Probably the best known findings showed that those people who got married and stayed married (already a select group of all the people who ever got married) did show a brief and modest increase in happiness around the time of the wedding. Then within a year or maybe a few years, they went back to feeling as happy or as unhappy as they were when they were single. So, among people who got married and stayed married (but not those who later divorced), marriage resulted in a short period of feeling a bit happier – basically, a honeymoon effect. It didn’t last. (I discussed these results in Singled Out and elsewhere.)

Grover and Helliwell used the data from the British Household Panel Study to do the same kinds of analyses on a British sample. When they did the same analyses in the same way that Lucas and his colleagues did, they found the same thing: “…the long-term marriage effect for people who have been married at least six years is approximately zero.” Translation: By the time people have been married for six years, they are not any happier than they were when they were single.

Yes, this is from the working paper that got all that attention for declaring that getting married makes people happier. You see, the authors did not like the finding that any happiness boost after getting married is short-lived. They believe in marriage and its super powers. So they came up with a way to reanalyze the data to save the day for marriage. (If Lucas had found that marriage was lastingly wonderful, do you think the authors would have challenged the findings?)

Lucas included in his sample anyone who had gotten married and had been in the study for a number of years, even if only one or two of those years had been spent single. His analyses look at how much happier (if at all) people get after they marry, compared to when they were single. Grover and Helliwell argue that it’s not fair to include people who have only been in the study and single for a year or two before they married. In that short period of time before marrying, they reason, people are already becoming happier in anticipation of getting married. That anticipatory happiness, they think, is part of the benefit of getting married. If you compare how happy they are later in their marriage to how happy they were just a year or so before they married, then the increase in happiness (if there is one, and after a few years, there’s not) is going to be too small.

So the authors did new analyses in which they included in their sample only those people who were single and participating in the study for at least 5 years before they got married. Once they included only that subset of people, then they found that people who got married and stayed married were happier than they were when they were single, even six years into the marriage.

People who are marriage apologists often like to argue that marriage is more than just the relationship itself. Making that formal, legal commitment matters. It’s a piece of paper, and much more. People making this argument are usually saying that cohabitation isn’t good enough. Real, legal, official marriage is what’s special.

But with their five-years-single stipulation, Grover and Helliwell seem to want to be sure that the people they include in their analyses did not even have marriage in their vision when they first joined the study as single people. Then, once they get close to marrying, they might get happier, but that’s anticipatory happiness that belongs with the marriage effect – or at least should not be allowed to dampen the marriage effect (which is the supposed boost in happiness you get by marrying).

You can buy their argument or not. What’s clear, though, is that finding evidence to suggest that getting married makes people happier, is not a simple task. Even setting aside all of the methodological challenges I described in my previous post, researchers’ first attempts at demonstrating that marriage causes people to be lastingly happier have not been all that successful. And so researchers persist, trying this and that, including and excluding certain people, until they get results that seem to support their beliefs. You can be impressed if you want. I’m not.

In One Analysis of One Hypothesis with People from One Country, the Authors Did Not Use the Cheater Technique

One of the articles about the Grover and Helliwell research was titled, “Middle age is slightly less terrible when you’re married.” In addition to arguing that marriage causes people to be happier, Grover and Helliwell want to make the case that marriage is especially good for your well-being during middle age.

Maybe you know from reports of other research (I haven’t read it closely myself) that happiness tends to decrease over the early adult years, reaching a low point sometime around the late 40s, then gradually increasing again. It’s a “U-shaped” effect. Grover and Helliwell believe that the marriage advantage will be especially strong during that most miserable time of many adults’ lives. So if married people are happier than single people in the group being analyzed, then they will be especially happier during middle age. And if, as in places such as Latin America and the Caribbean, it is the single people who are happier than the married people, well then, the single people’s advantage will be smallest during middle age. Their evidence is mostly consistent with that, except for in Sub-Saharan Africa.

The authors’ first attempt to test the middle-age effect involved data from the UK Annual Population Survey. These data are cross-sectional: They compare people of different marital statuses at one point in time. As I explained in my previous post, these are the kinds of data that just cannot strongly support any causal statement. Initially, the authors compounded their cross-sectional problem by adding on top of it their use of the cheater technique. That’s the one where researchers who want to make the case for marriage do so by including in the married group only those people who are currently married, and setting aside all those people who got married, hated it, and got divorced. Then, if and when married people look better, they say, “See, getting married made people happier!” (It didn’t. The people who divorced also got married and it didn’t make them happier.)

If your big idea is that marriage is especially likely to make people happier during middle age, then the cheater technique becomes even more problematic than usual. By the time people reach middle age, many of them who once married have now divorced. So by middle age, the currently-married group is an even more select group than it was in the earlier adult years.

Now here’s the good news. The authors realized this. So they did something I have been urging researchers to do for about two decades: Compare everyone who ever got married to those who stayed single. If you want to talk about the implications of getting married, you need to include in your analyses everyone who ever got married.

When the authors did that more appropriate analysis, they found that the marriage advantage was smaller than it was when they used the cheater technique. But for their one sample (UK), there was still an advantage. For all the reasons I described previously, this result does not demonstrate causality; it does not show that getting married makes people happier. But it is a better approximation to a causal argument than arguments based on the cheater technique.

Because it is better in that way, it is worth looking at the results a bit more closely. The authors compared the life satisfaction of ever-married people and always-single people for 14 different age groups, starting with 25 or younger and continuing through 86 and older. On an 11-point scale of life satisfaction (0 through 10), the biggest difference in happiness between the two groups was just under 0.4 points. For five of the 14 age groups, the difference favoring the ever-married people was 0.2 points or less. For a sixth group, the oldest group, the always-single people were happier than the ever-married people. I don’t think these results support a simple fortune cookie type message, “Get married, be happier.”

The authors realized the bias in the cheater technique. So why did they not use the more appropriate analyses throughout their research? Maybe because a more defensible way of testing the supposed benefit of getting married would not produce the desired results. In an American longitudinal study of marriage, the researchers conducted analyses that did and did not involve the cheater technique. They looked at happiness and other outcomes, too. They wanted to know if getting married resulted in benefits that remained after the first few years. When they used the appropriate non-cheater technique to look at the outcomes for people who had gotten married or partnered at least four years ago, they found that those who had gotten married were not happier, they were not any less depressed, they were not healthier, and they had no higher self-esteem.

There is one more claim made by the authors and repeated in the media that I want to critique – that the reason marriage makes people happier (a claim which I dispute) is because of the friendship between the spouses. The way the authors frame that issue is really telling. I’ll save that discussion for later.

Notes: (1) Thanks to Erin Albert, Kim Calvert, Carol Hynson, and Elizabeth Saenger for the heads-up about this latest bout of matrimania. (2) Image is from Google Images, labeled for reuse. (3) In their working paper, the authors asked that anyone quoting from it provide full credit, including the copyright. So I fully credit Grover and Helliwell, [copyright symbol] 2014 by Shawn Grover and John F. Helliwell, for NBER working paper 20794. (4) If you want to read other critiques of other claims about getting married and getting happy or healthy or living longer, try Marriage vs. Single Life: How Science and the Media Got It So Wrong or the shorter version, The Science of Marriage: What We Know That Just Isn’t So