,
4 tweets,
1 min read
Read on Twitter
I guess I’d start the discussion by congratulating that fellow for getting his methodologically strange experiment published in the NYT without having to explain why the confidence interval for his key result with independents is blown out to 12 percentage points.
The experimental design is to take a couple groups of voters, make one group read a “news snippet” composited by the investigator of Dem primary candidate positions he apparently considers most objectionable, and then ask who they’re going to vote for.
The result is that independents forced to focus on one party’s primary, as summarized by a hostile observer, in an environment where the other party apparently isn’t described either don’t love it or get wise to what they’re supposed to say to the tune of 0-12%.
One more thought about that study Chait highlighted: maybe that huge confidence interval indicates that—even in experimental conditions—independent voters are relatively scarce, and basing your electoral approach exclusively on them forecloses too many opportunities.
