Looks like adjusting has moved MI up a few spots in the state rankings (4, from #43 to #39, to be precise): South Dakota and northern Wisconsin have particularly odd distributions which, to me, suggest there is still some weird statistical stuff going on in this data for rural areas, which have very low denominators. It’s still interesting how sparse the gays seem to be in the Midwest compared to the Coasts & the Sunbelt.
First, we have updated national maps of the distribution by county, which based on a cross-check of Michigan counties with the previous county map, appears to reflect slightly more intuitive distributions. So, did the process change the results? A little bit. The Census only releases adjusted state-level data, so for the county-level estimates the attempt at adjustment was performed by Williams Institute demographer Gary Gates & his team. The Census Bureau preferred estimates adjust original Census tabulations… to account for the likelihood that a small portion of different-sex married couples miscode the sex of a spouse and are incorrectly counted as a same-sex couple. This one is based on “adjusted data.” What, you may ask, does it mean to adjust the data? The Institute explains:
The Williams Institute has taken a second crack at counting our gays.
