This chart, documenting the change of support for gay marriage in the US on a state-by-state basis, has been doing the rounds of the blogosphere recently (click to enlarge):
The chart is based on a hierarchical model of polling data by Jeff Lax and Justin Philips. They used the glmer function in R to aggregate poll data from various sources and create an “opinion estimate” for each state in three periods: 1994-1996, 2003-4 and 2009-9. It shows clearly how support for gay marriage has been steadily increasing everywhere (with the notable recent exception of Utah), but is generally increasing fastest in those states where some same-sex rights have already been conferred. Andrew Gelman provides further detailed commentary of this chart in an op-ed for the political website NewMajority.com. It’s nice to see such in-depth statistical analysis and graphics from R having an impact on political discourse.
