official: https://t.co/6cw8WbXiv3 ungated: https://t.co/LSsPFbVWGJ
1,382 followers
1,382 followers
But this misses that candidate preference was pretty stable through 2016. A model predicting support in December 2015 does not decrease in performance through the primary season as it ought to if primary campaign strategies changed & this changed suppo
1,382 followers
@JakeMGrumbach @arpitrage Though doesn't seem to directly translate into vote choice changes. In 2016 assuming a one day national primary would not have clearly led to different outcomes in either nominating contest. Basic models predict pre-primary suppor
1,382 followers
@JohnHolbein1 2016 primary election outcomes don't seem to change candidate support, suggesting they're not unique producers of "momentum" https://t.co/jjehMW74zJ