Health
Avoid Stroke by Marrying? A Case Study in Misrepresentation of Marriage Findings
I prefer science to matrimania and other misinformed hype
Posted March 2, 2010
A recent set of posts here at Psych Today provide a case study in how findings about marital status and life outcomes get misunderstood and misrepresented, and then set off a chain reaction of false proclamations about the transformative powers of getting married. The myths get perpetuated, and the purveyors of bad social science writing can usually count on not getting called on it.
Not this time.
Matrimania Unleashed
PT bloggers Charles and Elizabeth Schmitz were so psyched. With exclamation points galore, they proclaimed what they understood to be the results of the latest study. "Get married and avoid a stroke!" That was the headline of their post. Start reading and you will find this: "So here's the deal, stay single or have an unhappy marriage, and increase the risk of having a stroke by 64%! Wow, is this an incentive for a happy marriage or what!!"
(Answer: "or what.")
With this unfortunate start as a lead in, they then go on to propagate many of the other myths about getting married - that people who get married live longer, have better health, and all the rest. Sometimes their claims are qualified - you have to have a happy marriage to get all the benefits. That's telling, too.
Another blogger quickly jumped in to cheerlead for marriage. He touted the finding, which he described as "remarkable," that 90% of people who are currently married say they would marry their spouse again. (If it is not immediately clear what's wrong with that claim, my critique is here.)
Then the comments started. James Littleton (I don't know if that's his real name) said he followed the link to the report of the stroke research, and declared that the Schmitz duo "stated the current research exactly correct!!!" (He's excited, too.) Based on his Google search, he added that "there are many findings from studies that report all of the other benefits stated in this article." He then thanked the Schmitz's for their "good job of reporting for their readers."
I wish the Schmitz's, the other blogger, and James Littleton were the only perpetrators of these false tales about getting married. But claims like this get posted and published all the time, setting off even more misinformation, more exclamation points, and just plain bad science reporting. The problem is, though, that if the basic research paradigm does not support causal conclusions (e.g., "get married and avoid a stroke"), it does not matter how many such flawed studies there are or how many exclamation points you add at the end of your false statements - they are still false.
What that Stroke Study Really Did Show
So let's see what the research really did show. First of all, your alarm system should go off as soon as you hear that the results were reported at a conference and not in a peer-reviewed journal article. The only thing online is an abstract (here and here) and a press release (and numerous pieces drawing solely from those documents). Readers of this blog know that I always want to read the original research report before I make claims about what the study showed.
I emailed both the author of the study and the person who was asked to comment on it at the conference, and asked for the published journal article. Both responded promptly. There is no published version, and no copy of the talk as presented at the conference was available either (or they just decided not to send it to me). I asked some specific questions, and the study author answered some of them right away.
So now let me tell you what I was able to learn about what the study really did show.
1. First, the men described as single included men who were divorced, separated, and widowed, as well as those who had always been single. So, as is typical in so much of the research touting the ostensible advantages of getting married (recall the title of the Schmitz-pair's post, "Get married and avoid a stroke!"), many of the men in the "single" group DID get married. They just hated it, and got unmarried. The always-single men were not analyzed separately from the previously married men. That matters quite a lot. In many studies, the always-single group does better than the previously married group, and very similarly to the currently married group (even though the latter group is a select subgroup including only those who got married and stayed that way).
2. Second, this was not a longitudinal study. The men were not studied at various points in time over the course of their adult lives, as they got married or unmarried or stayed single. So the findings that have gotten so much attention were based primarily on three things: the marital status of the men in the study in 1963; for the married men, their ratings of how successful their marriages were in 1965; and for all the men, how many had died of a stroke by 1997. (The abstract also reported the overall rates of death in 1997 for the two groups, but I haven't seen that mentioned in any of the media reports or commentary.) Again, as with point #1, the implication is that this study did not, and could not, tell us anything about the implications of getting married for stroke or anything else.
3. Third, about that 64%: Recall the Schmitz's statement: "So here's the deal, stay single or have an unhappy marriage, and increase the risk of having a stroke by 64%!" That sure sounds like a lot of stroked-out single men and unhappily married men. (The study is actually about fatal strokes only, not all strokes.) It sounds like a huge difference. I asked the author about the death percentages corresponding to the unmarried and married men, and to the unsuccessfully married and successfully married men. He did not respond to that question, but I can give you some idea by referring to what was reported in the abstract. Of the men who were currently married at the start of the study, 7.1 % had died of a stroke within the subsequent 34 years. Of the unmarried group (divorced, widowed, separated, and always-single), 8.4 % had died of a stroke within the same time frame. Probably not the sort of difference you imagined when you heard that 64% figure.
4. Fourth, what about rates of death from all causes? For the married men, 64.9% of them had died by 1997. For the group in which the divorced, widowed, separated, and always-single men were all glommed together, 69.7% had died by that time. Again, this is a difference, but probably not the one you expected when you heard that men could cut their risk of a fatal stroke by 64% if only they did not "stay single or have an unhappy marriage." And again, remember, that difference between 64.9 and 69.7 is NOT the result of getting married, because many of the men in the unmarried group did get married - and hated it and got divorced.
5. Fifth, notice the practice, so commonplace in studies of marital status, of looking closely within the married group to find the men with the best outcomes, while not doing anything comparable for the singles. So, the author compares the men who were most positive about their marriages to the men who were least positive. He finds that the ones with the more successful marriages were less likely to have a fatal stroke. He never compares the most happily single men with the least happily single men, or those who wanted to be single with those who did not. It is as if the goal were to find something great about marriage, rather than to apply the same even-handed scientific method to all groups.
The bottom line is this. Contrary to the title of the Schmitz's blog post, the research did not show that you can "get married and avoid a stroke." The very design of the study made that conclusion impossible to demonstrate.
Still More False Claims
Said the Schmitz's, "You see, the benefits of marriage are numerous." Actually, they are not. I've explained this so many times that I won't go over it all again here. I'll just refer you to other places where I did spell it out, such as in Singled Out, Single with Attitude, and posts such as this one on this Living Single blog.
Finally, thanks to Kath for the heads-up about the stroke study.
Added later: I looked up the Wendy Wood research mentioned by the Schmitz's. It is a review paper published in Psychological Bulletin in 1989. In it, the authors compared currently married people to all unmarried taken together (divorced, widowed, and always-single). So, just like all the rest, this work cannot be used to support the claim that getting married has some transformative effect.