This seems like a good post for this question, but how much can you factor in errors related to demographic change since the last actual census?
I know that the census has yearly estimates but do particularly dynamic or fast-growing states such as Texas or Florida have extra error assumed in their polls?
Hi Alex: That’s a good reason, and one of the reason my MRP models seem to underestimate Biden in Texas and Arizona is for exactly that reason. I use population statistics as of 2016, but the electorate has gotten about 2 points less non-college white since then, which I think leads me to underrate Democrats. But this is a hard problem to really solve empirically and takes a lot of time to code!
While the map should speak for itself, I still wonder about the implicit conclusions about TX, GA, IA and OH. Biden has a better chance in these three states than he has in OH, or at least measurable enough that the shading is different on the map? The first three have moved left in the last 4 years while with Ohio, not so much?
Hey Craig! I think that’s right. This year might be one where we see lots of states continue their trends left/right relative to the national vote over the last couple of years. But until Election Day all we’re really doing is guessing.
This seems like a good post for this question, but how much can you factor in errors related to demographic change since the last actual census?
I know that the census has yearly estimates but do particularly dynamic or fast-growing states such as Texas or Florida have extra error assumed in their polls?
Hi Alex: That’s a good reason, and one of the reason my MRP models seem to underestimate Biden in Texas and Arizona is for exactly that reason. I use population statistics as of 2016, but the electorate has gotten about 2 points less non-college white since then, which I think leads me to underrate Democrats. But this is a hard problem to really solve empirically and takes a lot of time to code!
While the map should speak for itself, I still wonder about the implicit conclusions about TX, GA, IA and OH. Biden has a better chance in these three states than he has in OH, or at least measurable enough that the shading is different on the map? The first three have moved left in the last 4 years while with Ohio, not so much?
Hey Craig! I think that’s right. This year might be one where we see lots of states continue their trends left/right relative to the national vote over the last couple of years. But until Election Day all we’re really doing is guessing.
Phew! I am so glad you say so. Thanks so for your insightful findings with translations for the laymen. Aloha!