> i.e. random error. Did they say that their confidence interval was confidence against random error? I don't think that's a requirement of a confidence inter

> i.e. random error. Did they say that their confidence interval was confidence against random error? I don't think that's a requirement of a confidence interval, I think it's certainly valid to encode e.g. the chance of "shy Trump" respondents in a confidence interval.