Posts
The US election is tomorrow: the fate of civilization itself almost seems to hang in the balance.
The polls show Biden will win.
On the other hand, a couple of polls do show a Trump win. Either way, then, some polls are going to be wrong. So can we trust “the polls”?
There is, it seems, a late surge for Trump. All the polls pick this up. Even the polls that predict a Biden win show the race as close in key swing states, although he is winning the national popular vote. Even if the Bidenite polls are dead on, then, this late surge might unpredictably flip the result.
One might, of course, want to argue that a late surge matters little this time, because so many people have already voted. Yes; but those who have already voted will be those who had already firmly made up their minds. Those who were open to changing their minds will be disproportionately those who have not yet cast a ballot.
Then there is the possibility of “shy Tory” voters, people who intend to vote for Trump, but lie to pollsters. This sometimes seems to have been a factor in the past: in Brexit, in the last British general election, in Trump’s victory in 2016, or in the last Florida governor’s race. Pollsters claim they have allowed for this—but how can they, really? The inclination will vary election to election, and seems to me unpredictable by any definite metric. Given the hostility of the anti-Trumpers, it seems possible that this is a larger factor this time than it has ever been before.
Some are arguing that the way to circumvent this effect is to ask people how they think “their friends” will vote. This technique, they say, has been more reliable than the standard sort of polling in four recent elections in which it was tried. And this technique reputedly predicts a Trump win.
Scott Adams also raises the possibility that some Trump supporters may have deliberately lied to pollsters just for the satisfaction of watching a leftist meltdown on election day. Compilations of these, of women screaming into their smartphone cameras, or talking heads looking grim, have been popular ever since 2016 on YouTube. Perhaps a significant proportion of poll respondents really want more of that.
I have been disappointed before, and wrong before, but my gut says a Trump win.
The polls show Biden will win.
On the other hand, a couple of polls do show a Trump win. Either way, then, some polls are going to be wrong. So can we trust “the polls”?
There is, it seems, a late surge for Trump. All the polls pick this up. Even the polls that predict a Biden win show the race as close in key swing states, although he is winning the national popular vote. Even if the Bidenite polls are dead on, then, this late surge might unpredictably flip the result.
One might, of course, want to argue that a late surge matters little this time, because so many people have already voted. Yes; but those who have already voted will be those who had already firmly made up their minds. Those who were open to changing their minds will be disproportionately those who have not yet cast a ballot.
Then there is the possibility of “shy Tory” voters, people who intend to vote for Trump, but lie to pollsters. This sometimes seems to have been a factor in the past: in Brexit, in the last British general election, in Trump’s victory in 2016, or in the last Florida governor’s race. Pollsters claim they have allowed for this—but how can they, really? The inclination will vary election to election, and seems to me unpredictable by any definite metric. Given the hostility of the anti-Trumpers, it seems possible that this is a larger factor this time than it has ever been before.
Some are arguing that the way to circumvent this effect is to ask people how they think “their friends” will vote. This technique, they say, has been more reliable than the standard sort of polling in four recent elections in which it was tried. And this technique reputedly predicts a Trump win.
Scott Adams also raises the possibility that some Trump supporters may have deliberately lied to pollsters just for the satisfaction of watching a leftist meltdown on election day. Compilations of these, of women screaming into their smartphone cameras, or talking heads looking grim, have been popular ever since 2016 on YouTube. Perhaps a significant proportion of poll respondents really want more of that.
I have been disappointed before, and wrong before, but my gut says a Trump win.
No comments:
Post a Comment