,
8 tweets,
5 min read
Read on Twitter
Unlike @NickWolfinger, I don't have a dog in the fight over whether married people are happier than unmarried people. But building off of what he's pointed out, Paul Dolan's argument using ATUS data that many outlets have picked up is even more misguided than I first thought. 
Dolan expands on this argument in his book, Happy Ever After. The book also provides a few figures like this one. But notice: those six categories line up perfectly with values of the "marital status" variable, having nothing to do with who is present for the ATUS interview.
The marital status information isn't even collected as part of the ATUS phone interview. It's from the CPS interview that happened 2-5 months prior. A narrower question of the presence of a spouse or unmarried partner (in the *household*) is asked at the ATUS interview.
These are the values that the marital status variable takes. When I calculate mean "happiness" values over these, they roughly line up with the book figure. So it isn't measuring a spouse's presence for the interview, or even for any activities--just presence *in the household*.
As @NickWolfinger points out, the book doesn't note whether any of these differences are statistically significant. And since most people are pretty happy on this measure, it doesn't look particularly meaningful (but I disagree on using the GSS instead). ifstudies.org/blog/are-marri…
This claim, repeated breathlessly by many media outlets, appears to be based on a flawed analysis that actually compared slight differences in reported activity-level happiness for married people whose spouses live in the same household from those whose spouses live elsewhere.
I'd be glad to discuss the intricacies of these ATUS variables further with @profpauldolan. In the meantime, we can all take advantage of the extensive documentation @ipums staff have compiled (with links to BLS documentation) at atusdata.org.
