Hacker News Reader

Comment by bryanlarsen | Hacker News Reader

Somebody who got it wrong and actually admitted he got it wrong is often a credible source on why he got it wrong.

I disagree. It's FAR too early for him to be writing about "why we got it wrong" because if they knew that now, they would have known that last week.

Look, if it was 2 points in a couple states here or there, I get it! But... if the guy was claiming the possibility of a 400 vote Biden landslide? Nah man, that is either massive incompetence or it's intentionally misleading. Neither scenario do I trust this person.

> because if they knew that now, they would have known that last week.

Last week he didn't have access to the results or exit polling data like he does now.

> if the guy was claiming the possibility of a 400 vote Biden landslide?

A polling error of this magnitude in the other direction would have been a massive landslide.

Which one of Cohn's swing state predictions are within "polling error"? If you can't even consider the possibility that the polls were pushed to demoralize, that's not on me.

Here... Let me give you a source of pollster who explains just how crazy Cohn, Silver, Wasserman's polls were and how drastically they overshot.

This guy is right leaning, but no fan of Republicans, but the view here is that he's pissed about how poorly everyone did this and how it's incompetence or malice.

https://www.youtube.com/watch?list=PLHe3u4zZfY7K1N04z6Xu_JbY...

Watch from 30min to 40min. That's 10 minutes that I think will make the point with numbers that I can not make here.

> how crazy Cohn, Silver, Wasserman's polls were

Silver didn't do any polling, and is the guy that wrote [I'm here to remind you that Trump can still win](https://fivethirtyeight.com/features/im-here-to-remind-you-t...) before the election.

> incompetence or malice.

IIRC, Cohn said his number on Latinos in many districts was off by 50%. Call it incompetence if you want.

> Which one of Cohn's swing state predictions are within "polling error"? If you can't even consider the possibility that the polls were pushed to demoralize, that's not on me.

A normal polling error is about 5%, so all of them are within that except for Florida and Wisconsin.

First, I think you misunderstand poll margins of error. For example, the NYT poll is +-4 for each candidate. Meaning it could swing a total of 8 points.

That isn't anything new.

"In a new paper with Andrew Gelman and Houshmand Shirani-Mehr, we examined 4,221 late-campaign polls — every public poll we could find — for 608 state-level presidential, Senate and governor’s races between 1998 and 2014. Comparing those polls’ results with actual electoral results, we find the historical margin of error is plus or minus six to seven percentage points. (Yes, that’s an error range of 12 to 14 points, not the typically reported 6 or 7.)"

(written before 2016 election) https://www.nytimes.com/2016/10/06/upshot/when-you-hear-the-...

Reply on news.ycombinator.com