Poll Analysis by a Facebook Friend

And so, as promised, here is a post as to why polls can be wrong in the U.S (they can be wrong elsewhere, but for other reasons). The interest of this analysis is that right now, polls are pointing in one direction while ALL secondary indicators point the opposite way: […]

And so, as promised, here is a post as to why polls can be wrong in the U.S (they can be wrong elsewhere, but for other reasons).

The interest of this analysis is that right now, polls are pointing in one direction while ALL secondary indicators point the opposite way:

– Rally attendance and social media following give a Trump advantage measured in orders of magnitude, not percentages;

– Arms sales (a figure correlated with Republican support) are at a historic high;

– The change in voter registration since 2016 has favoured Republicans in all contested states;

– Institutions which “go Democrat” lose public support. The audience for basketball has dropped by 45% since it became politicised.

Therefore, either the secondary indicators are misleading, or the polls are wrong.

Historically, the second explanation happened more often, so I’ll focus on it.

A first methodological weakness of U.S polls is their tradition of sampling respondents by party affiliation. It partly presumes the answer in the question and gives rise to very different polls for the same state, through arbitrary changes of the D/R/I divide. It would be statistically safer to sample by geography, age and professional status only, according to accessible statistics.

A second source of error is the atmosphere of intimidation – moral and physical – of political opponents that currently exists in America. If you’re hesitant about stating your political preference in public, you will sometimes be hesitant about stating it to pollsters.

And finally, there are many ways through which differences in motivation can be honestly missed by pollsters.

If your assumption about youth participation is based on past elections, but this time there is a pandemic and the young’uns are at home instead of being on campus where political groups organise them, they’ll vote less than usual and your polls will miss it.

If your assumption about rural area participation is based on past elections, but this time people are terrified of the sight of chaos in Democrat cities and determined to go out, they’ll vote more than usual and your polls will miss it.

I’m not even talking at this point about those polls that are deliberately misleading in order to discourage the opponent’s supporters. If you see anything called a “CNN poll”, it’s not worth reading further: why would everything else that these people do be insanely dishonest, except their polling?

I could be wrong. But my current assessment is that the difference in direction between average polls and secondary indicators is best explained by the abovementioned phenomena.

If anyone thinks that it’s the secondary indicators which mislead, I’ll be glad to read your analysis.

Next Post

Amazon Launched a Bunch of Hidden Prime Day Deals That Will Help You Beat the Rush

amazon.com The official start of Prime Day may still be a few days away, but that isn’t stopping the retailer from kicking off some epic sales.  Amazon dropped tons of early Prime Day deals days ahead of the annual shopping event, which begins on October 13. With discounts up to […]