Forecasts and polls got the 2016 election results dead wrong

Donald Trump won the 2016 presidential election, beating out Hillary Clinton. Although many forecasts had Clinton as the favorite.

--------------------------------------------------

--------------------------------------------------

Business Insider is the fastest growing business news site in the US. Our mission: to tell you all you need to know about the big world around you. The BI Video team focuses on technology, strategy and science with an emphasis on unique storytelling and data that appeals to the next generation of leaders – the digital generation.

[con_m_computerscience]

Whose ready for “How they got the 2020 election polls dead wrong Part 2” 💀

[Agooo]

3 more highly impactful states not mentioned in the video but made a huge difference:
Ohio
Michigan
Arizona

[coleconverse]

Here's the problem: State polling in the Midwest: Many White Republican Americans in this election and will likely increase over time will be less likely to talk to pollsters about their true voting intentions. Hence, polling in the Midwest/Rust Belt will likely be the biggest challenge for pollsters going forward.

Polling Errors in Midwest/Rust Belt (nearest .5%):(From RCP Average); (-->)= Actual
North Dakota- Trump 14-->Trump 36.5 OFF 22.5
South Dakota- Trump 10.5-->Trump 30 OFF 19.5
Nebraska- Trump 27-->Trump 26.5 OFF -.5
Kansas- Trump 16--> Trump 21 OFF 5
Minnesota- Clinton 9--> Clinton 1.5 OFF 7.5
Iowa- Trump 3--> Trump 9.5 OFF 6.5
Missouri- Trump 11--> Trump 18.5 OFF 7.5
Wisconsin- Clinton 6.5--> Trump 1 OFF 7.5
Michigan- Clinton 3.5--> Trump .5 OFF 4
Illinois- Clintion 11.5--> Clinton 16 OFF -4.5
Indiana- Trump 10.5--> Trump 19 OFF 8.5
Ohio- Trump 3.5--> Trump 8 OFF 4.5
Pennsylvania- Clinton 2--> Trump .5 OFF 2.5
West Virginia- Trump 18--> Trump 42 OFF 24

Polls on average in these 14 states were off on average 8.2% and if you take out Illinois with a substantially higher minority population than any other state on the list (37% minority, all others < 25% minorities), your polling error is 9.2% on average. And the polling error for states from RCP either lean Dem, tossup, or lean GOP from these states is 4.6%. Most polls have a 3% margin of error, hence actual results are well outside their polling margins of error even in the projected closer states.

[declinerrexlete]

If even he had a 0.1% chance of winning, its
still not impossible lmao

[parthin]

What were they polling? Landline telephones?

[cqdgqlj]

Love how they assume everyone who didn’t vote would automatically have voted Hillary. This is what happens to people who live in a bubble and never talk to anyone from flyover country.

[skx]

How is this wrong exactly? It said there was a 30% chance of winning, that's probably pretty accurate as Trump's win was an considered an upset, even by happy Trump supporters

[marcofeature]

28% winning means Trump will win once every 3~4 times, and he did win at this election, narrowly. It's not a terribly bad prediction I would say.

[skatefang]

I mean to be fair it didn’t predict Clinton’s stupidness by not visiting Wisconsin and also neglecting Pennsylvania and Michigan. She got wayyyy too cocky

[the_boss]

What about Michigan? Michigan was polled for Clinton but went to Trump instead.

[TheFranchiseCA]

To put these in context, imagine rolling a pair of dice. Most of us are familiar with that from playing Monopoly, Clue, Risk, etc. Winning Florida or North Carolina was like rolling a 2, 3, 4, 9, 10, 11, or 12. Barely worse than a coin flip. Pennsylvania was like rolling a 9, 10, or 11. Wisconsin was like rolling doubles.

But unlike rolling dice, getting a good result in one place (say, Florida) means the odds actually go up other places.

[shubhampawaskar]

55% chance in NC and Florida is a tossup. So it could have gone either way. Maybe they were off in Wisconsin and Pennsylvania.

[kenfarlez]

The media couldn’t beat Trump with their lies and fallacies about him. They need to try harder lmao.

[bobsaggat]

You say :mistakes" I say they deliberately lied to push their own agendas

[mike]

Greatest Election Ever!!
You couldn't write it any better.
A modern day revolution occured.
Amazing!

[Joshua-Samarita]

Yeah, poll asked over thousand of people and then you expect the remaining of people to vote the same? Seems legit.

[janelle]

Wisconsin going to Trump was so tuff to the Democrats and their supporters

[cianmcdonagh]

Nevada polls had trump ahead? Same with Florida and North Carolina on the day of the election. The major shock was Pennsylvania.

[javiersblog]

Maine shouldent be colored red with blue stripes. It should be colored blue with a red stripe. Trump only got one electoral vote out of there and Clinton won the popular vote their pretty handily, even more so then New Hampshire

[rsriddel]

well, as it's a 'poll', it is by definition a statistical approximation. And its approximation was much more accurate than chance. So...no.