I bet Massachusetts changed behavior for Thanksgiving a lot more than Kansas or Ohio
In the sense that Kansas and Ohio generally were not socially distancing etc. before Thanksgiving, so their behavior on Thanksgiving was no different, whereas Massachusetts socially distanced before Thanksgiving, and then for one day said screw it, we're not socially distancing for this one day?
yes
Wish he would have done California but I think we were starting to spike sooner
I can see that in general, but I would be really interested in seeing what his model says the Thanksgiving bump for Kansas is. If you look at the line for the seven-day average for Kansas, it's hard to see how there could be any kind of Thanksgiving bump. There is a dip immediately following Thanksgiving almost certainly due to people not getting tested, counties not reporting and the like. There is a bump that follows it to make up for the cases that didn't get reported and then the downward trend continues.
https://www.worldometers.info/coronavirus/usa/KansasMy two thoughts on this are:
1. If your model is finding a positive effect from Thanksgiving for Kansas, your model is looking for a result and designed to do so.
2. If your model does not find a positive result from Kansas, is it really plausible that there is the same amount of interaction on Thanksgiving as prior to it. There's almost no way that people from around the state are all getting together on one day. Then add in the people that came in from out of state and the amount of social distancing is almost certainly less in Kansas than it was prior to Thanksgiving. In normal years, there is less social distancing on Thanksgiving than on the days leading up to it.