Exit polls in the battle ground states and pre-election polls for the 2022 elections projected the Republican Party to win control of the U.S. Senate. The computerized vote counts reversed these projections with the Democratic Party maintaining control of the Senate. Most countries, unlike the United States, do not allow computers to count their ballots—all ballots must be hand-counted and the process observable. (See highlighted section below).
As in past elections, as noted in this website, the discrepancies between the exit polls and the computerized vote counts overwhelmingly favored one party. If the discrepancies were due to random errors, the discrepancies would favor both parties about equally. As the discrepancies favor mainly one party, a systemic problem, with either the polls or the computerized vote counts, is the cause.
In the past two decades, the discrepancies between the polls and the vote counts overwhelmingly favored the Republican Party. That is, the computerized vote counts totaled more votes for the Republican Party candidates than projected by the polls. In this 2022 midterm election the opposite occurred. The computerized vote counts totaled more votes for the Democratic Party candidates than projected by the polls.
Exit polls and pre-election polls, imperfect as they are, provide, unfortunately, the only means to check the veracity of computer vote counts: The computerized vote counting found in all U.S. States is A) inherently not observable B) logic and accuracy tests (L & A tests) of the machines before and after elections are easily defeated by software that operates only during the election [1] C) audits of the machine vote counts varies in every state D) precincts to be audited may not have been randomly chosen; Arizona, for example, preselects batches of votes to be audited [2] E) only the District of Columbia conducts highly transparent audits where observers can verify the actual marks on ballots [3]
In the 2022 election, exit polls were conducted by Edison Research in nine battleground states. In seven of these states the Democratic Party candidate exceeded their exit polls totals and pre-election poll projections. In three states this discrepancy was greater than the margin of error for these polls.
In two states the Republican Party candidate computerized vote count totals exceeded their exit poll and pre-election poll projections. These two states, Florida, and Ohio, have a long history with claims of fraudulent elections. In Florida, Clint Curtis, a computer programmer, testified under oath at a U.S. Congressional Hearing that he wrote, in 2000, vote-flipping software for Tom Sweeny, a Florida Congress person to rig Florida’s elections. In Ohio, in 2004, Presidential candidate Kerry lost to Bush with claims of massive, and documented vote fraud. If Kerry had won Ohio, he would have won the Presidency.
[1] Walker, J., Bajaj, N., Crimmins, B.L., Halderman, J.A. (2022). Logic and Accuracy Testing: A Fifty-State Review. In: Krimmer, R., Volkamer, M., Duenas-Cid, D., Rønne, P., Germann, M. (eds) Electronic Voting. E-Vote-ID 2022. Lecture Notes in Computer Science, vol 13553. Springer, Cham. https://doi.org/10.1007/978-3-031-15911-4_10 (“Research has long recognized that L & A testing cannot reliably defeat an adversary who manages to execute malware on voting machines, because the malware could detect when it was under test and only begin cheating during the election itself.”)
[2] Arizona’s audits, according to the former Secretary of State, Ken Bennet, preselects a small number of batches of votes to be audited and party representatives choose about half of these preselected batches to be audited. In the 2020 Presidential election between Trump and Biden, in Maricopa County, there were over 10,300 batches of ballots and only 52 of these batches were preselected to be hand-count audited. Party representatives then chose about half of the batches that were set aside for a recount. See https://smartelections.us Counting & Auditing the Election (beginning at -9.20 minutes prior to end of video).
[3] See https://verifiedvoting.org, Audit Law Database. (Transparency chosen in drop-down menu, scroll to bottom to see result)
Notes:
[1] Exit polls (EP) conducted by Edison Research and published by CNN shortly after the closing of polls for the state and downloaded by TdMS. Edison Research conducted EPs for Senate races in nine states. Exit poll results are derived from the gender category–the proportion of men and women voting for each candidate. As these first published exit polls are altered/adjusted to conform with the unverified computer vote counts, the discrepancies shown above are adjusted to near zero in the final EPs.
[2] New York Times reported vote counts. https://tinyurl.com/323kca43
[3] The margin columns subtract the Republican Party candidates’ vote percentages from the Democratic Party candidates. A Democratic Party candidate’s win is shown by a positive sign and the Republican Party win by a negative sign.
[4] Note that the Margin of Error (MOE) is for the differences between the two candidates (at 95% CI). This MOE is about double the usual MOE for each candidate. MOE calculated with multinomial formula discussed in sections 2 and 4 in: Franklin, C. The ‘Margin of Error’ for Differences in Polls. University of Wisconsin, Madison, Wisconsin. October 2002, revised February 2007. Available at: https://web.archive.org/web/20221110150249/https://abcnews.go.com/images/PollingUnit/MOEFranklin.pdf
[5] Pre-election polls as compiled and averaged by the FiveThirtyEight. https://projects.fivethirtyeight.com/polls/senate/2022/
The United States remains one of the few major democracies in the world that continue to allow computerized vote counting—not observable by the public—to determine the results of its elections [1]. Countries such as Germany [2] Norway [3], Netherlands [4], France [5, 6], Canada [7] , Denmark [8, 9], Italy [10], United Kingdom [11], Ireland [11], Spain [11], Portugal [11], Sweden [11], Finland [11], and most other countries [11], protect the integrity and trust of their elections with publicly observable hand-counting of paper ballots.
[1] According to a 2020 Gallup World Poll, only 40% of Americans say they are confident in the honesty of U.S. elections. Finland and Norway with 89% of their citizens expressing confidence in the honesty of their elections along with the citizens of 25 other countries have greater confidence in their elections than do Americans.
[3] “Norwegian votes to be counted manually in fear of election hacking“
[4] “Fearful of Hacking, Dutch Will Count Ballots by Hand“
[5] “Voting in France: Paper ballots, in person, hand-counted“
[6] French Senate: Making the moratorium on use of voting machines permanent. CONTINUE, AS IS, THE MORATORIUM
[7] Canada Elections Act (S.C. 2000, c. 9). Part 12. Counting Votes. Election officer to count votes in the presence of others
[8] Denmark’s election law does “not permit electronic voting technologies to be used during voting”
[9] Folketing (Parliamentary) Elections Act, Translation (Consolidated Act No. 1260 of 27 August 2020) CHAPTER I: GENERAL ELECTIONS AND REFERENDUMS. Counting of the Votes Cast at the Polling Station, Part 9 “[T]he polling supervisors and the appointed electors, … shall count the votes cast at the polling station. The counting is public.”
[10] Election of House of Representatives and the Senate of the Republic September 25, 2022. Chapter 23. Ballots counted by officials.
[11] The ACE Electoral Knowledge Network. Vote Counting. Country List. See “Count votes by hand”
Exit polls downloaded from CNN’s website at poll’s closing: Arizona–Florida–Georgia–Nevada–Ohio–Pennsylvania–Wisconsin–New Hampshire–North Carolina.
The exit poll vote proportions were derived from the gender category (all the other categories would have approximately the same result). Democrat’s proportion of the male vote was multiplied with the total male proportion and added to Democrat’s proportion of the female vote multiplied with the total female proportion to arrive at Democrat’s exit poll vote share in the state. The same procedure was applied to arrive at Republican’s proportions of the exit poll vote. As these first published exit polls are altered/adjusted to conform with the unverified computer vote counts, the discrepancies shown are adjusted to near zero in the final published exit polls.
Great work. How do we know the exit poll data (gender domegraphics) was not pre-adjusted before sending to CNN? Is it possible they adjust based on past exit poll/tally deviations?
Hi Chris. Thanks.
I doubt that they specifically adjust their polls for their past deviations from the actual vote count. But such adjustments may be reflected in how they weigh non-responders to their polls.
Edison mainly adjusts for non-response by the voters. To exit poll election day and early voters, they’ll have “interviewers”–they do not interview anyone, just hand out anonymous questionnaires to the voters–and note the age, race, and gender of those refusing to take the questionnaires and adjust their exit poll results accordingly. Older people tend to vote Republican, blacks, Democrat for example. See historic voting trends since 1972 in page 6-18 in:
https://www.aei.org/wp-content/uploads/2022/01/The-Exit-Polls.pdf?x91208.
They also conduct telephone interviews to capture the ever-increasing share of vote-by-mail and early voters and adjust for non-response.
Edison Research “typically have people at 700 to 1,000 precinct locations. In 2020, we also had people at 200 early voting centers around the country. We interviewed about 30,000 people by telephone before Election Day to identify by mail and early voters. Every survey you saw in 2020 was some mix of these three modes.” See link above.
That said, pre-election polls, according to Nate Cohn, the NY Times chief political analyst, are specifically weighed to adjust for their past “misses” by adding an average of two points to their polls of Republicans in this election. An average of two points would not significantly alter the above table. See:
https://web.archive.org/web/20221128231327/https://www.nytimes.com/2022/11/02/upshot/polls-2022-midterms-fix.html
The main problem, of course, is that the hidden computerized vote tabulation is trusted as accurate while pre-election and exit polls are not. In actuality, no one can know if the computer counts are accurate, with much evidence showing they are not.
We need to follow Germany’s example, and most of the world, and count votes transparently by hand.
Diverse film genres find a home in Conti’s thoughtfully curated portfolio.