I shared this article with a friend (seriously) and she said this: “Actually, the author is comparing the margin of error to the marginal difference between them in the poll vs the actual voting. The margin of error has to do with the actual polling number. If Bernie polls at 65% with a 5% margin of error, then it could really be 60% or 70%. The author is looking at the difference in the margins and then applying the margin of error to that. Ummmm, that’s not how statistics works.”
Is she right, or is she misunderstanding your argument?