Daniel Weitzel @danweitzel.net · Oct 23 We find that common features of commercial polls like not asking don’t know, inserting a partisan cue, and treating unconfident answers as knowledge inflate the partisan gaps by up to 50%. 3/7
14 likes 1 replies
?
Replies
Removing these inflationary features from survey questions produces knowledge estimates that are more strongly associated with established predictors of political knowledge, such as interest, participation and education. 4/7