2 ms·
> This is a sizable decline in just four years, albeit the two results cannot be directly compared. Hersh and Ghitza’s finding was based on a database of voter
by rsynnott 3y ago
> This is a sizable decline in just four years, albeit the two results cannot be directly compared. Hersh and Ghitza’s finding was based on a database of voter registration records. They focused on voters in 30 states and did not have direct data to define married couples. In contrast, the American Family Survey was based on a national sample of 3,000 American adults, and I looked at married adults and compared respondents’ party identification with their spouses’.
(From their main source article on the claim that marriages across party lines are falling.)
One of the more irritating habits of the media is to present two studies whose methods make any comparison, especially with a rather small change in numbers, fraught, explicitly acknowledge this problem, _and then proceed to compare them anyway_.