re: #96 thedopefishlives

Actually, to piggyback on this topic, I just realized something. There probably is a selection bias in picking people who watched the speech, and as I mentioned before, it would be unavoidable. However, it could be counteracted by selecting the sample with this selection bias in mind. That is to say, selecting the sample such that the political balance of the sample matched that of political America at large. thus, you minimize the selection bias of being forced to include only people who watched the speech.

I think that’s very likely the case.