Thread by @Trinhnomics

I'm reading about polls (we actually have votes now as it is T-4) so polls increasingly less important.

What do you know about polls? As in how they are done? Let me summarize this article I read on why u should be skeptical

politico.com/news/magazine/…

https://twitter.com/Trinhnomics/status/1321756340088315905

In 2016, Arie Kapteyn (Dutch Economist for USC) & Robert Cahaly (Republican pollster for Trafalgar Group) saw Trump coming while others didn't as they predicted Trump getting Penn, Michigan, FLorida & NC & he won all.

Let's talk about how they think current polls underestimate.

3 factors they discuss that u should know:
1) Existing polls that use PHONE is outdate b/c who has time to answer 23 mins of phone conversation & tell someone what they really think if they're a Trump supported & so people answer are EXTREME or BORED. Short & anonymous is better.

So they discussed the issue w/ the methodology itself as phone/in person only gets motivated people to answer esp if the survey is pretty long.

2) Beyond the way polls are conducted, which they think should be short & anonymous, polls should address the issue of social stigma.

They already discussed the difference b/n popular vote vs electoral where general polls get right for popular but not electoral but frankly who sits in the White House is electoral so whatevs.

The issue of social stigma goes back to human psychology & that we want acceptance.

As in, we are aware of what the others think & in 2020 the way the mainstream media defines a Trump support is WORSE THAN 2016. In 2016, they were just "deplorables", yes Hillary said that. But in 2020, they get punished for expressing outside mainstream opinions & even beat up!

Say you live in a neighborhood that has 1 person that is very ANTI TRUMP. And you are generally polite & u want that person to like you, then you NEVER WOULD SAY U ARE PRO TRUMP, unless u get a kick out of antagonizing your neighbor b/c anti Trump voters are very vocal & visceral

People know that they're judged badly for being a Trump supporter, esp if they are a nice person & don't fit into the mainstream narrative of a Trump supporter then they'd never say it. Reverse happens for Biden supporter in a super Trump area but the stigma for Biden isn't there

Word that they use is "social desirability bias" in human psychology & how that plays out in polls, especially if polls are conducted over the phone/in person.

Concept is similar to Adam Smith of the Theory of Moral Sentiment where u want to be liked b/c u're a social creature.

Because of this, they do the following to ADDRESS the issues of a "social desirability bias" & also outdated ways in which polls are conducted that puts off busy people & tends to get the extreme:

They do anonymous polls & tailor their questions to gauge sentiment.

Let's look.

Beyond simply asking voters whether they support Biden or Trump, USC asks a “social-circle” question—“Who do u think ur friends and neighbors will vote for?”—makes it easier for people to share their true opinions without fear of being judged for their views.

Surveys short too.

Finally, polls, beyond methodology issues discussed, DON'T take into account LOW-PROPENSITY voters for Trump (people that usually don't vote) & REPUBLICATION VOTER REGISTRATION.

There u go. Take that information however u may but at least u learn something about polling & bias.

Fine, some people asked me to do a poll, let's do it. Imma do their style & ask this:

WHO DO YOU THINK YOUR "FRIEND" & "NEIGHBOR" WILL VOTE FOR? 😉