You are reading a single comment by @andyp and its replies.
Click here to read the full conversation.
andyp-
Thanks, these were interesting articles and, inevitably, the more you get into it, the more complex it gets. But I see lots of discussion of herding, sample weighting and segment turnout assumption tweaking, but nothing that looks to me like it is trying to correct for systematic lying on the part of a chunk of respondents, eg shy Trump voters.
frank9755
Nate Silver comes with a lot of caveats, but this might be useful