Election polls in 2020 produced ‘error of unusual magnitude,’ expert panel finds, without pinpointing cause (gee...what could that cause POSSIBLY be...?)

The Conversation ^ | July 20, 2021 | W. Joseph Campbell

[DoodleBob]

More than eight months after the acute polling embarrassment in the 2020 U.S. elections – that produced the sharpest discrepancy between the polls and popular vote outcome since 1980 – survey experts examining what went wrong say they have no definitive answers about why polls erred as markedly as they did.

That inconclusive finding reported by a polling industry task force will do little to assuage popular skepticism about election polls which, in one way or another, have misfired in all U.S. presidential races but one since 1996.

And if the source of the 2020 polling error cannot be pinpointed, then addressing and correcting it obviously becomes daunting.

Moreover, as I discussed in my book “Lost in a Gallup,” polling failures in presidential elections since 1936 rarely have been repetitive. Just as no two elections are alike, no two polling failures are quite the same.

Over the years, pollsters have anticipated tight presidential elections when landslides have occurred. They have signaled the wrong winner in closer elections. The estimates of venerable pollsters have been singularly in error. Wayward exit polls have thrown Election Day into confusion by identifying the losing candidate as the likely winner. Off-target state polls have confounded expected national outcomes, which essentially was the story in 2016.

Trump support underestimated

In 2020, election polls pointed to Democrat Joe Biden’s winning the presidency. But collectively, the polls underestimated backing for then-President Donald Trump no matter how close to the election the survey was conducted and regardless of the methods pollsters chose. Surveys in races for U.S. senator and governor were beset by similar flaws.

Those were among the findings described in a report made available on July 19, 2021, that noted that voter-preference surveys in 2020 “featured polling error of an unusual magnitude” and that the discrepancy in the presidential race was the greatest in 40 years.

The experts, who comprised a task force of the American Association for Public Opinion Research, a survey industry organization, speculated that some Republicans may have been less willing than Democrats to be interviewed by pollsters – a hypothesis that could explain some of the polling error. But the task force report said “identifying conclusively” why polls erred “appears to be impossible with the available data.”

The task force, which included 19 members from the polling industry, the news media and academia, said it reviewed data from more than 2,800 polls and found that surveys in the 2020 presidential race overstated Biden’s popular vote advantage by 3.9 percentage points.

This marked the fourth presidential election in the past five in which the national polls, at least to some extent, exaggerated support for Democratic candidates.

Masking dramatic miscalls

Averaging the polling errors, as the task force did in conducting its months-long analysis, is broadly revealing about the extent of those errors. But it also has the effect of masking several dramatic miscalls in late-campaign polls conducted in 2020 by, or for, leading news organizations.

The final CNN poll had Biden ahead by 12 points. Surveys for The Wall Street Journal-NBC News and by the Economist-YouGov had Biden winning by 10 percentage points as the campaign wound down. A few polls, such as Emerson College’s survey, came close in estimating the outcome.

Biden won the popular vote by 4.5 percentage points.

The report said the task force rejected several prospective causes of polling error in 2020 – including those that likely distorted survey results in key states in 2016 when Trump unexpectedly won an Electoral College victory. Those factors included undecided voters swinging to Trump late in the campaign and a failure by some pollsters to adjust survey results to account for varying levels of education.

White voters without college degrees were understood to have voted heavily for Trump in 2016, but those voters were underrepresented in some polls in key states such as Pennsylvania and Wisconsin, where Trump won narrowly and surprisingly.

The task force also rejected as a factor in 2020 any errors pollsters made in projecting the likely makeup of the electorate in terms of age, race, ethnicity and other factors – an estimate common to preelection surveys.

The task force reported finding “no evidence that polling error was caused by the underrepresentation or overrepresentation of particular demographics” in the preelection surveys.

Additionally, it is unclear whether Trump’s sharp criticism of preelection polls in 2020 dissuaded his supporters from participating in surveys.

“So it’s possible that these may be short-term phenomena that will abate when Trump is not on the ballot,” Daniel Merkle, the then-president of the American Association for Public Opinion Research, said in a speech in May.

“On the other hand,” Merkle said, “it could be a broader issue of conservatives becoming less likely to respond to polls in general because of a decline in social trust, or for some other reasons. It will take further evaluation to understand this nonresponse issue and to adjust for it.

"This may not be an easy task.”

Overblown characterizations

In the immediate aftermath of the 2020 election, several media critics declared that polling could be “irrevocably broken” and faced “serious existential questions.”

Such disquieting assertions seem overblown; polls are not going to melt away. After all, election polling represents a slice of a multibillion-dollar industry that includes consumer and product surveys of all types.

And if election polling survived the debacle of 1948 – when President Harry S. Truman defied predictions of pollsters and pundits to win reelection – then it surely will live on after the embarrassment of uncertain origin of 2020.

[DoodleBob]

Now...lemme see..what could account for the colossal failure in polling in 2020...?

And, again, pointing to the popular vote winning margin is like looking at the NY Yankees' 55 runs in the 1960 World Series vs the Pirates' 27 runs, and declaring that the Yankees were the winner (the Pirates won that WS 4 games to 3).

For the record, Trump's win in 2016 by 77,744 votes (the sum of the winning margin in theTipping point states of PA, MI, and WI) as a percentage of the total popular vote was a narrow 0.0569%, but on par with Carter's winning percentage and higher than that for Truman and Bidet*.

By contrast, Bidet's winning* margin of 0.0277% was the 4th lowest since 1868, and on par with that of Kennedy's (also shady) win. By contrast, the biggest margin percentages were 1936-FDR, 1972-Nixon, 1964-Johnson, and 1984-Reagan. And for what it's worth, Obama and Wilson are the only Presidents (aside from FDR) who's second term winning margin was lower than that in the first term.

Winner	Total Popular Vote	Pop Vote Margin	Popular Vote Margin/Total Popular Vote	Tipping Point States	Sum of Pre-Tipping Point Popular Margins	Sum of Pre-Tipping Point Popular Margins to Total Popular Vote	Sum of Pre-Tipping Point Popular Margins to Total Tipping Point Popular Vote
2020 Biden*	158,383,935	7,052,120	4.453%	3	42,918	0.0271%	0.3673%
2016 Trump	136,669,237	(2,868,691)	-2.099%	3	77,744	0.0569%	0.5577%
2012 Obama	129,085,410	4,982,291	3.860%	4	527,737	0.4088%	2.5770%
2008 Obama	131,313,820	9,550,193	7.273%	7	994,143	0.7571%	3.6067%
2004 Dubya	122,294,846	3,012,166	2.463%	3	134,648	0.1101%	1.7063%
2000 Dubya	105,405,100	(543,895)	-0.516%	1	537	0.0005%	0.0090%
1996 Clinton	96,275,401	8,201,370	8.519%	10	1,388,367	1.4421%	5.8860%
1992 Clinton	104,423,923	5,805,256	5.559%	11	613,637	0.5876%	2.8642%
1988 GHW Bush	91,594,686	7,077,121	7.727%	12	1,231,838	1.3449%	4.0101%
1984 Reagan	92,653,233	16,878,120	18.216%	17	5,947,627	6.4192%	12.2299%
1980 Reagan	86,509,678	8,423,115	9.737%	18	1,605,352	1.8557%	4.2728%
1976 Carter	81,531,584	1,683,247	2.065%	2	46,361	0.0569%	0.7462%
1972 Nixon	77,744,027	17,995,488	23.147%	16	6,812,738	8.7630%	16.1019%
1968 Nixon	73,199,998	511,944	0.699%	3	172,177	0.2352%	1.9917%
1964 Johnson	70,639,284	15,951,287	22.581%	24	4,653,332	6.5875%	15.7769%
1960 Kennedy	68,832,482	112,827	0.164%	3	18,953	0.0275%	0.2756%
1956 Eisenhower	62,021,328	9,551,152	15.400%	15	2,625,170	4.2327%	10.8569%
1952 Eisenhower	61,751,942	6,700,439	10.851%	14	1,896,387	3.0710%	7.7873%
1948 Truman	48,793,535	2,188,055	4.484%	2	24,972	0.0512%	0.3589%
1944 FDR	47,977,063	3,594,987	7.493%	10	720,296	1.5013%	3.3634%
1940 FDR	49,902,113	4,966,201	9.952%	11	1,145,860	2.2962%	4.2824%
1936 FDR	45,647,699	11,070,786	24.253%	17	4,873,549	10.6764%	16.8014%
1932 FDR	39,751,898	7,060,023	17.760%	14	2,003,798	5.0408%	9.6534%
1928 Hoover	36,807,012	6,411,659	17.420%	12	1,052,714	2.8601%	8.0695%
1924 Coolidge	29,097,107	7,337,547	25.217%	16	685,717	2.3567%	9.6009%
1920 Harding	26,765,180	7,004,432	26.170%	18	1,335,434	4.9894%	16.0643%
1916 Wilson	18,536,585	578,140	3.119%	2	28,016	0.1511%	0.3517%
1912 Wilson	15,044,278	2,173,563	14.448%	20	417,028	2.7720%	6.7548%
1908 Taft	14,889,239	1,269,356	8.525%	8	149,980	1.0073%	4.2272%
1904 T. Roosevelt	13,525,095	2,546,677	18.829%	8	445,961	3.2973%	10.8031%
1900 McKinley	13,997,429	857,932	6.129%	7	237,695	1.6981%	6.2735%
1896 McKinley	13,938,674	601,331	4.314%	5	70,914	0.5088%	2.8443%
1892 Grover Cleveland	12,068,027	363,099	3.009%	6	45,168	0.3743%	1.9839%
1888 B. Harrison	11,383,320	(94,530)	-0.830%	2	16,721	0.1469%	0.9006%
1884 Grover Cleveland	10,060,145	57,579	0.572%	1	1,149	0.0114%	0.0984%
1880 Garfield	9,219,477	9,457	0.103%	3	28,339	0.3074%	1.7543%
1876 Hayes	8,418,659	(252,666)	-3.001%	1	889	0.0106%	0.4866%
1872 Grant	6,471,983	763,729	11.801%	10	140,259	2.1672%	5.9524%
1868 Grant	5,722,440	304,810	5.327%	6	58,060	1.0146%	3.7761%

[DennisR]

“...that produced the sharpest discrepancy between the polls and popular vote outcome...”

Polls mean absolutely (or nearly) nothing. The only thing that matters is the popular vote outcome; that is:

1. Were only legal votes counted; and
2. Were the legal votes counted accurately

[ProtectOurFreedom]

As Rush (may he RIP) said so many times, public opinion polling is NOT designed to measure public opinion, it IS designed solely to SHAPE and MOLD public opinion. Hence the polling that showed Biden with insurmountable, unbeatable leads up to 19% -- even though he didn't campaign, drew a handful of people to a couple of events, and stayed holed up in his rec room basement.

[j.havenfarm]

I have been very successful in my field. But I’m not a math guy. Got a math
Block. Statistics was the easiest math requirement I could take in college. It took me SIX times to pass it. Even so, I know the answer, geniuses: your sample sets are screwed up. They’re too heavily weighted for democrats. Consistently. Over and over. I’d call them f’in morons, but the fact is they do it on purpose to try to affect the election results, and this introspection is fake.
Ok, I’ll call them f’in morons anyway.

[JohnBrowdie]

stop it. the polls have been ****ed for a decade. this is just dumb.

[CatOwner]

The polls were necessary in 2020 to provide cover for the fraud that was about to be perpetuated. And most of the country bought it.

[monkeyshine]

survey experts examining what went wrong say they have no definitive answers

Well, let's Occam's Razor this. My opinion, it's because polling outfits are largely FOS. They rig polls to give the answers their paymasters want. Oh, forgive me... they don't "rig it", they just manipulate the questions and their contacts and censor the results to fit their paymaster's wishes. Oh forgive me... that is rigging it.

Anyway I don't care what anyone says. There is no way Joe Biden won that election legitimately. No way he got 20 million more legitimate votes than Hillary did in 2016. No way Trump lost AZ, PA, MI, WI or GA. Probably not NV either. Trump won by a landslide and the cheaters had to go over the top to make it appear legit. So in 2020, in the middle of a pandemic, more votes were cast then ever before, and the 2nd place finisher got more votes than any other candidate ever, except the winner? I call BS.

[Pikachu_Dad]

How did we gain 22,000,000 boyers in just 4 years ?!?

[null and void]

ROUS? I don’t think they exist.

[FlyingEagle]

Polls are now about spoofing citizens into not voting because propaganda about lopsided victory “polling” forecasts.

Polsters are all beholden to the totalitarians.

[unclebankster]

Trump was supposed to get blown out in 2016 and 2020, like Goldwater or McGovern in 1964 & 1972.

Guess who I’m supporting in 2024?

Hint: The guy who was our 45th President.

I like him a lot.

[lewislynn]

yea, yea, yea, none of that matters. Who's buying Hunter's art work? Did VP Biden follow Hillary's lead using private e-mail for official business? Is the Obama/Biden/Harris NSA spying on only one private citizen named Tucker Carlson?

The deafening silence of the likes of the Tucker/Ryan/Fox News cabal and the expected deaf, dumb and blind fake news media and Mike Pence performing his constitutional duty (GAG!) of rubber stamping the election as legit is all anyone needs to know.

ALL HAIL OBAMA/BIDEN/HARRIS!!!

what could account for the colossal failure in polling in 2020...?

Their failure to figure out how to pol 10's of thousands of non-existent voters for Biden/Harris in the largest voter fraud network ever...Don't take my word for it Biden said it himself...Even though the apologist media, including Fox will tell you that was just a typical Biden gaffe we should laugh off.

[Dick Bachert]

HOOVERVILLE & THE FBI
4 MINUTE VIDEO
https://www.brighteon.com/06cf4d20-5f17-438c-87d0-4f32e00f48a4

[kiryandil]

Statistics was the easiest math requirement I could take in college. It took me SIX times to pass it.

Sometimes, they just don't teach it correctly to some students.

I have a math minor. I know what I'm talking about when it comes to moron professors.

[5th MEB]

Polls are requested and paid for by the ENEMEDIA, the ENEMEDIA tells the pollsters what they “EXPECT” the results to be.

It’s really simple once you know how the system works.

[kiryandil]

We didn't.

They're "virtual" voters...

[kiryandil]

Hence the polling that showed Biden with insurmountable, unbeatable leads up to 19% -- even though he didn't campaign, drew a handful of people to a couple of events, and stayed holed up in his rec room basement.

Middle of May, 2020.

[Still Thinking]

Easy peasy. Next time just poll each election worker 5,000 times, make sure to ask the tabulating machines, and don’t bother with asking any voters. You’ll be right on the money.

[TBP]

No surprise.