Woman + (man * testosterone) = divorce?
Or, the search for a number in the testosterone-divorce connection.
Anneli Rufus at the Daily Beast offers an extremely helpful column called "15 Signs You'll Get Divorced." And by "extremely helpful," I mean as a source of material for me.
I was especially attracted to Divorce Sign #4:
If you're a man with high basal testosterone, you're 43 percent more likely to get divorced than men with low testosterone levels.
Many writers who try to bring research findings to a general audience go to great lengths to get to a specific number. As a researcher, I have tried to oblige journalists -- to find the sweet spot where they have a specific number to use while I can preserve a kernel of truth distilled from the months or years it takes to produce a single journal article.
Often, the quest is futile, as the resulting number is either misleading or -- as in this case -- just plain wrong.
The number
To her credit, Rufus includes a specific reference for this 43% number. The reference is to a book called Biosociology of Dominance and Deference by Allan Mazur. I haven't read the book, but the reference is fortunately found on a page (125) that Google Books showed me a preview of.
The study referenced in the book was published right down the hall from me in the journal Social Forces -- 17 years ago. Using what was then state-of-the-art evolutionary theory and hormone-testing technology, the authors analyzed testosterone levels in more than 4,000 military veterans, which they then compared with their previous marital history. The result showed that those who had been divorced were more likely to have higher testosterone levels (measured once, years later). Unfortunately, the authors, Alan Booth and James Dabbs, described the finding like this: "men producing more testosterone are less likely to marry and more likely to divorce" -- as if the testosterone were measured before the family history occurred.
So, in fairness to Rufus, the fact has a source. Unfortunately for Rufus, on the next page the book -- which Google also showed me -- Mazor explains that testosterone levels change over a man's life, and that, in particular, "testosterone is highly responsive to changes in marital status, falling with marriage and rising with divorce." (Short-term fluctuations in response to competition -- like those on election nights -- are well documented.)
And, in fairness to Booth, in 2006 he wrote more moderately (also in Social Forces, with Mazur as well as other co-authors) that "the evidence is weak and additional research is needed to reach a confirm conclusion about reciprocity" in the relationship between competition (including marital conflict/divorce) behavior and testosterone levels. Others are also on the trail of this relationship.
There remains no definitive proof that high testosterone levels increase the likelihood of divorce. There is no number. In the absence of a proven relationship between divorce and testosterone, I can't recommend shelling out for a T-test before a trip to Vegas.
On the other hand
However, if you want to stick with this idea, I am honor-bound to offer you this comparison of trends.
A. American men's testosterone levels have been falling -- not just as men age, but over time at the same ages:
B. Divorce rates have been falling since the 1980s:
Cross posted from the Family Inequality blog.