Saw a great point made about Rasmussen and the pollsters that were "right" in 2016

[Dr. Jack]

I was reading comments on a political betting market this even and of course all of the Republicans are sooooper excited about that Rasmussen poll that shows the orange turd up by 1 point. Of course that ignores all of the other polls out today showing Biden by up to close to double digits. But I'm preaching to the choir on those points.

But here is the point one user made that is absolutely perfect when it comes to those right wing polls that "got 2016 right" that I never thought of before.

"The entire "Republican Pollster" industry is just a bunch of grifters who got lucky in 2016. These kinds of people were pushing these kind of polls before maybe knowing that, eventually, there would be probably eventually be an election where Republicans outperformed polls and now they're hanging onto that."


For me that was a "oh shit, that makes perfect sense" moment for me. Looking back, these pollsters were saying the same shit in 2012, 2008, 2004, 2000, etc. Just showing very good results for Republicans that are always way off. They got lucky in 2016. It was a coincidence, not some brilliant secret knowledge about the electorate. That definitely destroys the "Rasmussen got 2016 right and that is why they should be trusted in 2020".

[Stallion]

1. Rasmussen's Final Poll in 2016 Had Clinton +2 but

they also had numerous polls with a more favorable Republican lean in the weeks preceding the election. They always move toward the norm as election apporoaches

[Awsi Dooger]

2. I made the same type of comment in a Reddit post earlier tonight

I'll just paste it here. It was posted in Political Discussion subreddit within the Polling Megathread, in regard to some posters who were hyping the Trafalgar polls based on 2016 and Florida 2018:

"I am not impressed with outliers, in polling or anything else. It is sucker mentality to believe that one model has some secret sauce that will consistently defeat wisdom of the crowd. I've seen that flawed thinking so many times, in following polling and wagering on politics since 1992. I remember when Mason-Dixon was the wise guy pollster. Then James Zogby. And so forth. Inevitably the way it plays out is that they stumbled onto a combination that happened to nail one cycle, and received hype for it, so everyone scrambles in that direction for the next cycle. One cycle too late.

It is similar to a squawker at a race track who gives out one longshot winner, and everyone suddenly views him as a seer, in eager anticipation of his pick in the next race. The squawker feels compelled to give out another long shot. This one is beaten 45 lengths.

Independents have moved sharply from 2016 to 2020, favoring Trump by 4-5 points in every respected poll of how 2016 played out, to 10+ points in Biden's favor this time. Independents shifted away from Trump early in 2017 and have not returned. That variable in itself lends weight to the notion that Trafalgar is forcing something that is not there. USC Dornslife tracking poll favored Trump in 2016 but has reported Biden significantly ahead this time.

In the 28 years I've been following I've seen dozens of trends come and go. The only one that is basically undefeated is Alaska polling consistently inept, always overstating the Democrat to absurd degree. I won plenty of wagers on that trend and started emphasizing the incompetent Alaska polling on political websites in 2002. Nate Silver finally caught on about a decade later and ran a lengthy article detailing that Alaska polling errs heavily toward the Democrat every time. But he did not go back nearly far enough in his examples. Maybe he didn't have those examples in his database."