Confused yet?
So I came across something today that pointed me to this survey of gay marriage polling tracks over time. I then noticed that this page also includes historical trends for interracial marriage as well.
My obvious question to myself was, "how much of that trend can you explain simply by assuming things are improving because old people die?"
That led to a search for detailed cohort data, because I need to know when people are dying. Google google google, and I find this page, which claims the image is from the Census people, although they've since rearranged their webpage so the link doesn't work anymore. I'm accepting that, because who lies about cohort data? Crazy statisticians, I guess.
 |
| Here's the cohort image, copied in case that link disappears too. |
I digitized that image, and then write a perl script to calculate the fraction of the population on a given date that were born before a certain year. This number subtracted from 100 gives the fraction born after that date. What does that look like?
 |
| Due to the five-year blocking of the cohorts, there's a stair-step effect. Interpolate a best fit line through that data, basically through the midpoints of each step. |
Huh. That worked far better than I expected. Basically, ignoring all other factors, you can pretty much explain the entire trend if you just assume that everyone born after 1960 (who would have been kids during the Civil Rights Movement) is cool with interracial marriage, and that everyone born after 1980 (I'm going to use the Simpsons as my cultural shift here, because I don't want to go with "people my age") is cool with gay marriage.
