Excellence in Polling

[Husky_Khan]

So I had these graphics professionally created in MS Paint this morning.

Susan Collins v. Sara Gideon... Gideon projected to win in every poll, in some by upwards of twelve points!

Spoiler: 2020 Maine Senate Elections

Gary Peters v. John James in Michigan... several elections projected the incumbent Gary Peters winning by 12... 13... even 15 points!

Spoiler: 2020 Michigan Senate Elections

2020 United States Senate election in Michigan - Wikipedia

en.wikipedia.org

Thom Tillis v. Cal Cunningham... Of the past twenty plus or so polls... one has a draw and two project a Tillis victory. But most project a comfortable Cunningham win just above the margin of error!

Spoiler: 2020 North Carolina Senate Elections

2020 United States Senate election in North Carolina - Wikipedia

en.wikipedia.org

Lindsey Graham v. Swamp Creature Jaime Harrison... STATISTICAL DEAD HEAT!

Spoiler: 2020 South Carolina Senate Elections

2020 United States Senate election in South Carolina - Wikipedia

en.wikipedia.org

2020 Presidential Elections between Donald Trump and Sleepy Joe Biden. Biden wins every poll except one and often by double digits... aggregates project a 7.5% to 8.5% point victory.

Spoiler: 2020 Presidential Election

Nationwide opinion polling for the 2020 United States presidential election - Wikipedia

en.wikipedia.org

Well done all around I think. They definitely improved from 2016 in their projections.

 

[MelancholicMechanicus]

It's quite simple what has happened here: Outdated methodology, outdated basis of data sourcing, dishonest reporting and wish fulfillment came together to destroy the science of political polling.

 

[ShieldWife]

Polls are primarily a method of disseminating information, not gathering it. Polls give Biden the edge to create a bandwagon effect, to energize his supporters while discouraging Trump supporters.

 

[Husky_Khan]

Four years ago we were all universal in our praise of political polling. From what we can tell today... it's only gotten better now that the race is in a statistical "dead heat" across multiple swing states for multiple days in a row...

Which of course... is also statistically near impossible...



A one in 1.95 trillion chance? Truly... we live in the most unfortunate of Antifascist Timelines.

As Nate Silver explains...

All told, we'd expect 55 percent of the polls to show a result within 2.5 points in a tied race. Instead, almost 80 percent of them did. How unlikely is that?

Based on a binomial distribution — which assumes that all polls are independent of one another, which theoretically they should be — it's realllllllllllllly unlikely. Specifically, the odds are 1 in 9.5 trillion against at least this many polls showing such a close margin.

How could this happen?

Because pollsters are lying and cheating.



Shockingly he says some of the higher quality polls actually come from periodicals such as The New York Times which apparently wants to show they have some shreds of journalistic credibility after a likely Trump victory.

There’s more herding in swing state polls than at a sheep farm in the Scottish Highlands

Some pollsters aren’t telling you what their data really says.

www.natesilver.net