Anyone remember 2016 and the polls that had Clinton up 8 - 12 points? Yeah?
I can believe Biden is winning but come the last two weeks of the election, a lot of "undecided" voters are going to still vote for Trump in the voting booth. It happens every election. These margins just aren't happening, as much as I would like them to.
This is not directed only at you, because you're far from the only person who repeats it, but: while I agree that this margin looks too high, the bolded talking point is getting really tiresome.
First:
we all know that Clinton led by good margins in early polls and subsequently lost. Anyone interested enough in politics to hang out here is quite aware of that, and doesn't need to be reminded of it. Of course things can change a lot between now and Election Day, just as they did in 2016.
That's the lesson to be taken from early polls: things can change -- not that they'll necessarily change
in a specific direction.
Second: outliers happen from time to time even with good pollsters. This poll certainly looks like one, as the majority of polls show Biden up by mid-single digits. The Fox poll showing a tied race is an outlier in the other direction. This is why poll averages are much more useful than looking at any single poll.
Third: you and others are assuming that Trump will win the lion's share of undecided voters. While this is certainly possible, it's far from a sure thing; and in fact I suspect it's more likely they'll break for Biden, for the following reasons:
a) Assuming undecided voters will vote for Trump in 2020 because they did so in 2016 is extrapolating from a sample size of ONE election.
b) Undecided voters more typically break against the status quo and for the outsider. In 2016 Trump was the outsider, but now he's the status quo.
c) Crosstabs from most polls have shown that undecided voters tend to disapprove of Trump by significant margins.
