Election Polls' Confidence Levels are Total B.S.

Election polls are 95 percent confident but only 60 percent accurate, UC Berkeley Haas study finds

By Laura Counts, UC Berkeley Haas
Wednesday, October 28, 2020

https://www.universityofcalifornia.edu/news/election-polls-are-95-percen...

How confident should you be in election polls? Not nearly as confident as the pollsters claim, according to a new UC Berkeley Haas study.

Most election polls report a 95 percent confidence level. Yet an analysis of 1,400 polls from 11 election cycles found that the outcome lands within the poll’s result just 60 percent of the time. And that’s for polls just one week before an election — accuracy drops even more further out.

“If you’re confident, based on polling, about how the 2020 election will come out, think again,” said UC Berkeley Haas professor Don Moore, who conducted the analysis with former student Aditya Kotak, B.A. '20. “There are a lot of reasons why the actual outcome could be different from the poll, and the way pollsters compute confidence intervals does not take those issues into account.”

Many people were surprised when President Donald Trump beat Hillary Clinton in 2016 after trailing her in the polls, and speculated that polls are getting less accurate or that the election was so unusual it threw them off. But Moore and Kotak found no evidence of declining accuracy in their sample of polls back to 2008 — rather, they found consistently overconfident claims on the part of pollsters.

More at the link

Sorry, I won't be here when this hits

be well and have a good one

Share
up
26 users have voted.

Comments

QMS's picture

Also, WTF is consumer confidence they measure? Does someone call you and ask what your buying mind has on a list? Of things to consume, with confidence that either you need it or it is actually available to buy? For what then, commingled with other datum. Advertising dollars no doubt.

Polls mean jack. Commissioned to make a point. Not an actual questioning of the masses as to if blue or red is the *in* thing this season. Or anything else, for that matter. It is a sign of our (r)age. One neighbor puts up a black lives matter yard sign. Next door a sign saying blue lives matter. One a Biden, the next a Trump. Divisiveness displayed.

We will have to work together to get thru this mess.

Namaste

up
17 users have voted.
thanatokephaloides's picture

@QMS

Polls mean jack. Commissioned to make a point.

That's why I hang up on them! Smile

up
9 users have voted.

"US govt/military = bad. Russian govt/military = bad. Any politician wanting power = bad. Anyone wielding power = bad." --Shahryar

"All power corrupts absolutely!" -- thanatokephaloides

The Liberal Moonbat's picture

@QMS If no one else is bright enough to do it, where are all the Poli-Sci scholars?

If you want to accomplish anything constructive in life, especially on the large scale, first order of business has to be figuring out what is already true, otherwise your plans will inevitably fall apart. Even after all these years of seeing Orwellian Evil in action, I still fail to understand how these people expect to succeed if their plans are not grounded in truth; all they accomplish, ultimately, is a planetload of gratuitous suffering and loss.

"We can ignore reality, but we cannot ignore the consequences of ignoring reality."
- Ayn Rand

up
10 users have voted.

In the Land of the Blind, the one-eyed man is declared insane when he speaks of colors.

To paraphrase Jodie Foster: Human is not something to aspire to, it's something to get away from.

and isn't synonymous with 'confidence.' The 95% confidence level means that it would be rare for the true number to fall outside the number +/- the MOE. The smaller the sample size, the larger the MOE and vice versa. Although larger sample size doesn't always mean more accurate until the pollster uses some means to obtain a representative and random sample of the whole population. Easier said then done.

Using an example -- the last 2016 Quinnipiac FL poll.

Clinton 46%
Trump 46%
MOE 3.3%

Actual results -
Clinton 47.8%
Trump 49.0%

The actual results were within the 3.3% MOE. So, the statistical confidence level was accurate.

up
19 users have voted.
Roy Blakeley's picture

@Marie the vote may not fall within the so-called confidence limits (e.g. if the sampling is bad). Then there is the issue of accurate tabulation of votes.....

up
10 users have voted.
snoopydawg's picture

and yet they started projecting Biden the winner even before he entered the race. But boy did they keep people distracted for those 2 years.

What is the purpose of this information? Does anyone believe that Biden would let either Warren or Bernie into his cabinet?

up
17 users have voted.

It is not until the tide goes out that you discover who has been swimming naked.

@snoopydawg that ultimately didn't work out that well.

up
8 users have voted.
Roy Blakeley's picture

@Marie A dispassionate observer might conclude that he was working for someone other than the Democratic Party.

up
19 users have voted.

to tell them to go f--- themselves, as I do anybody seeking political donations.

up
10 users have voted.
Bollox Ref's picture

I've willingly partaken of was for the Minnesota Orchestra, years ago. I suggested that they play a symphony by Franz Schmidt for the upcoming season, and lo and behold, they included his 1st Symphony, so we had to go.

up
16 users have voted.

Gëzuar!!
from a reasonably stable genius.

@Bollox Ref Glad you attended, hope they rocked it!

up
5 users have voted.
Bollox Ref's picture

@on the cusp

They did a good job. Huge round of applause.

up
7 users have voted.

Gëzuar!!
from a reasonably stable genius.

@Bollox Ref but will listen to it. Within 12 seconds, I will know if I have heard it.
I have John Hyatt paying now, can interrupt his greatness.

up
2 users have voted.

@Bollox Ref never heard of it. Saw on wikipedia his 2nd was a hit.
Another piece I had ready access to was the intermezzo from Notre Dame, which I assume was one of his operas.
Man! I almost fell out of my chair!
BEAUTIFUL! Such a nice way to spend 4 minutes of your life!

up
5 users have voted.
Bollox Ref's picture

@on the cusp

was my first introduction to him, years ago. His first two symphonies inhabit the 19th Century. The last two, the 20th.

up
1 user has voted.

Gëzuar!!
from a reasonably stable genius.

RantingRooster's picture

I don't need no stinking polls.
Crazy

What is polling, but the commodification and exploitation of the public's opinion for profit, not a public good. At least imho.

Considering the sobering results of the Martin Gilens and Benjamin I. Page study, "Testing Theories of American Politics:Elites, Interest Groups, and Average Citizens", we see that when public opinion comes to influencing actual public policy of our government...

"Multivariate analysis indicates that economic elites and organized groups representing business interests have substantial independent impacts on U.S. government policy, while average citizens and mass-based interest groups have little or no independent influence. The results provide substantial support for theories of Economic-Elite Domination and for theories of Biased Pluralism, but not for theories of Majoritarian Electoral Democracy or Majoritarian Pluralism."

snip

When a majority of citizens disagrees with economic elites or with organized interests, they generally lose. Moreover,because of the strong status quo bias built into the U.S. political system, even when fairly large majorities of Americans favor policy change, they generally do not get it.

A perfect example issue is MedicareForAll.
According to Newsweek.com: (4-24-20)

A newly released poll shows that 69 percent of registered voters support Medicare for All, a plan which would create a national health insurance plan available for all Americans.

The poll also showed 46 percent of Republican voters supporting Medicare for All alongside 88 percent of Democrats and 68 percent of Independents.

A clear majority of Americans support the idea of a Medicare For All type system. Neither Trump nor Biden support what clearly a majority of Americans want in their healthcare system.

I'm like, why should they (pollsters) get paid for my opinion that, in the grand scheme of things, really doesn't matter anyway? Why should I give them free, raw materials (my opinion) for their faulty questions (based on linguistic gymnastics), and ultimately worthless products, polls & commentary when public opinion has no influence on public policy?

It generally goes like this...
Me - We live in a capitalist society correct?
Pollster - Well, yes
Me - You get paid to solicit my opinion, right?
Pollster - Well, yes
Me - You want my opinion, pay me. That's how capitalism works, is it not?
Pollster - WTF? (This guy's nuts!) click.

Crazy

up
19 users have voted.

C99, my refuge from an insane world. #ForceTheVote

Probably the only truth that Bill Clinton ever told.

But who do they blame for being out of work? Trump? Or Dem governors?

up
4 users have voted.

I've seen lots of changes. What doesn't change is people. Same old hairless apes.

CalvinV's picture

to include the 2016 polls and assume that the 95% confidence reflects the final results.
There were almost 10% of undecided voters in 2016 because they disliked both major party candidates (I include voters who thought they would vote for third-party candidates as undecided). At the last minutes. many of them will vote for one of the two party candidates anyway, usually for the one they know less and therefore hate less.
The state polls are even worse with some 12% undecide.
Therefore, the polls with high percentages of undecided have built-in error as predictors for the future since we're almost guaranteed that the future will change. The confidence level and margin of errors are supposed to apply to the polls at the moment they were taken to reflect the population at the time. If hundreds of polls were taken at the same time, then 95% of them will fall within the margin of error of the average.
If the study removes all polls which have a more reasonable percentage of undecided, like under 5% then I bet it will fare better than 60%. But, fundamentally, they are wrong to assume that there are no major events between the time the polls are taken and the future it's supposed to predict that might change what people think.

up
4 users have voted.