Presidential campaigns don’t rely on public polls. Well, except for fund-raising. But not the campaign. Campaigns have their own pollsters who who have better ways of predicting voter behavior, means that are almost always much more expensive than those you see publicized. There are a lot of reasons why campaigns don’t rely on public polling, but the primary reason why is that public polls don’t (necessarily) use voter lists for their calls (or whatever method they use to contact people they are polling) and they don’t know how to weigh responses to accurately reflect the electorate.
You can see why when you look at the methodology from the new CNN poll: “The entire sample was weighted to reflect national Census figures for gender.” [p 29] That sounds nice, right? Except that this doesn’t reflect the electorate. Not this year, not last year, not in 2022, not in 2020, nor even in 2016. And, in fact, this means this poll is weighted heavily in favor of Trump.
Give me a break.
That CNN poll shows Donald Trump leading Joe Biden in the presidential race nationally, 49% to 43%. And there’s an important, critical reason why. It’s methodology sucks. “Men continue to be more likely than women to associate with the Republican Party,” and the CNN poll is weighted towards men.
”The female percentage in the USA is 50.41% while the male percentage is 49.59%.” Which, as we shall see, is the closest that men will come to women in the context of the voting electorate.
”In every presidential election since 1980, the proportion of eligible female adults who voted has exceeded the proportion of eligible male adults who voted.” Every single one. And there’s strong reasons to believe that 2024 won’t be any different, in fact, to think the “Gender Gap” in voting will be exacerbated.
Instead of a 0.8 percentage difference between men and women respondents in the CNN poll, i fully expect there to be, at minimum, a 5.9 percentage difference between the genders in the 2024 election. But that’s a conservative estimate. It is more likely that voter turnout in 2024 will exceed the male to female ratio of the Ohio amendment vote last year (53% female/47% male). The Ohio is a good base *but* the male population in Ohio is higher in that state than it is in the rest of the country. Any poll that doesn’t have a male to female breakdown of 47% male/53% female should be immediately suspect, and this one is...