Shame on the Associated Press for leading with this: “Love can sometimes break a heart but marriage seems to do it a lot of good.” Even more shame on Time magazine for cueing up this misleader, soaked in singlism: “Lovelorn singles, that ache in your heart will subside once you get married.” (At least they added “sort of” as the next two words.) Really, Time, all singles are lovelorn, and you believe everything you read unthinkingly and uncritically?
Double, triple, quadruple shame on the authors, who ended their PowerPoint presentation with this gem, presented as a cutesy cartoon: “Here’s the deal: Stay married or put yourself at risk.”
The heading of the article you are reading, “Get Married, Get Heart Disease: Study of 3.5 Million Adults,” is one way to represent the study that has been making the media rounds, starting, I think, with the Associated Press report and then picked up promiscuously by all sorts of media outlets since then. I don’t think my headline is totally accurate, for reasons I will describe. But it is at least as accurate as many of the other claims you may have seen – all of which insisted that Married People Win.
Here’s what my claim is based on. In slide 5 of the slide show (appears as a Power Point presentation if you paste this into your browser: timedotcom.files.wordpress.com/2014/03/marital-status-abstract-ppt-presentation-web-briefing-edited1.pptx), you will see the results of a greatly flawed study. One of the cardiovascular diseases they assessed was coronary heart disease. In the slide, people who have always been single are the comparison group. The groups compared to them are the currently married, the divorced, and the widowed. All three groups have higher rates of coronary heart disease than the always-single. One way to characterize these results is to say that if you have ever gotten married, you are more likely to have coronary heart disease than if you stayed single. Again, that’s not how I would describe the results of this wholly assailable study – or at least not without qualification. I’m using these headlines to make a point.
Other results seem less promising for single people. For example, the first part of the graph summarizes the results from all vascular diseases. It shows that the currently married have 5% lower odds of having any kind of vascular disease than the always-single. So does that mean, as the AP so irresponsibly put it, that “marriage seems to do it [your heart] a lot of good”? Is Time right in stupidly saying that your heart problems will subside “once you get married”? Look at the people who divorced. They got married, and they have a 5.1% greater chance of having any kind of vascular disease than the people who stayed single all their lives.
But none of these claims – not even the ones I like – should be taken seriously. Let me tell you just a few things about this study and I bet the long-time readers of this blog will instantly know what is wrong with each – even if they have no training whatsoever in research. So what excuse do paid writers for AP and Time have? What excuse do the authors of the original study have for the claims in their presentation?
What follows are just a few critiques. If you look at the authors’ presentation and the reports in the media, you can probably find even more problems.
Fact 1: The study compares currently-married, always-single, divorced, and widowed at one point in time.
The Problem: This is one of the most fundamental flaws of bad research. If you ever took a research methods class, even as a beginning undergraduate, you should have learned this on the first day. In fact, if you ever took a decent class in social psychology or even introductory psychology or any other class in the social sciences, you should have learned this: The study is correlational. Correlation is not causality. (How can someone be a social science or health writer for the AP or Time magazine and not know something this basic?) Any way in which the currently married people differ from the single people – other than the fact that they are married – is a possible alternative explanation of why their health differs.
Fact 2: The study compares the currently-married to the always-single, divorced, and widowed people.
The Problem: This is the classic cheater technique. I critiqued it in detail in Singled Out, and I referred to it often in debunking lots of other flawed studies making scientifically irresponsible claims about the implications of getting married for health, happiness, longevity, and more. The authors are skimming the currently married off the top of all of the people who ever got married, then they or others (such as Time magazine) are using those findings to suggest what would happen if you, as a single person, got married, or if you, as a married person, stayed married.
But the divorced people got married, too! If you want to talk about the implications of marrying, you need to include all of the people who ever got married. Otherwise, you are just cheating. As I’ve often said before, looking only at the currently married is like a drug company who wants you to evaluate their drug based only on the results of the people for whom the drug worked – when close to half of the people who took the drug got sicker and refused to continue.
About the authors’ cutesy cartoon telling married people to stay married or put themselves at risk: Seriously, authors? You think your study shows that if those people who were in horrible, conflict-ridden marriages – maybe even abusive marriages – had stayed married instead of getting divorced, they would be healthier?
If you want to see the implications of marrying or divorcing, you need to follow the same people over time to see how their health or happiness (or anything else that interests you) changes as they go from being single to married or married to unmarried. Even then, you need to look at all the people who ever married, not just those who got married and stayed married. And it still would not be as good as a true experiment that randomly assigns people to different conditions, but we can’t assign people to get married or divorced or stay single.
Fact 3: The people who participated in the study were those who sought screening from the company, Life Line Screening.
The Problem: This is not a random or representative sample of Americans. It is not a random or representative sample of single people or married people or divorced people or anyone else. They are people who self-select to get screened by a particular company. (The critique of this company or the kinds of screenings they recommend is beyond my scope here.) Who knows how they differ from everyone else, or what those differences might mean for what we make of the results of the study.
Fact 4: The study is based on more than 3.5 million Americans.
The Problem: The numbers are a red herring. In general, bigger studies are better studies. But if your study is deeply flawed, adding more people does not help. I would not be any more impressed with the results of this study if it were based on 4 billion people or on all of the adults alive in the universe. There was an old saying where I grew up (maybe it is popular everywhere): “You can’t shine sh*t.”
To read more debunking of all sorts of claims about marriage, and learn some general methodological principles for critiquing marital status studies, take a look at Singled Out. You can also read lots of blog posts critiquing claims about:
Getting married and (not) getting healthy
Getting married and (not) getting happier
Getting married and (not) living longer
Getting married and lots of other outcomes
[Note: Thanks to Jeanine, Valerie Park, and everyone else who sent me a heads-up about this study. If you are one of them and it is okay for me to thank you by name, please let me know and I’ll add it here.]