Exit Poll Versus Reported Vote Count

By Theodore de Macedo Soares

As in 11 of the 17 state primaries elections prior to March 17, the discrepancies between exit polls projections and the results of the unobservable computer vote counts in the 2020 Missouri Democratic Party presidential primary is large and beyond the margin of error associated with the exit poll.  Like the 2016 Democratic Party primaries, with candidates Clinton and Sanders, all but one of these large discrepancies favors candidate Biden and disfavors candidate Sanders.

In Missouri, the combined discrepancies between the exit poll and the vote count for candidates Sanders and Biden totaled 9.6%, more than double the 4.6% margin of error for the exit poll difference between the two. All margin of errors calculated at 95% confidence interval (CI).  Values greater than the margin of error are considered statistically significant. The discrepancy in favor of Biden substantially exceed the margin of error even at 99% (CI).

In practical terms, given the total number of votes (664,400) in this election, Sanders computerized vote count underperformed his exit poll projection by 11.4% or 29,500 votes less than projected. Biden’s computer counts overperformed his exit poll by 9.3% or 34,000 more votes than projected by the exit poll.

Articles detailing the primary elections of additional states are underway. The overall picture:

This picture is very similar to the 2016 Democratic Party primaries between Sanders and Clinton where the discrepancies systematically favored Clinton.

In the 2016 Republican Party primaries, however, the computerized vote counts matched the exit polls within the margin of error for all but two states. The reader should also notice that the discrepancies were just as likely to favor Trump as not—just as expected with unbiased exit polls and vote counts.

[1] The exit poll (EP) was downloaded from CNN’s website on election night, March 10, 2020.   The exit poll is marked as having been updated at 8:09 PM ET—about ten minutes after the polls closed in the state. It is likely that this publication included the total early/absentee votes as they are usually counted during the day.  The exit poll just prior to the incorporation of these totals may have shown a much larger discrepancy with the computer counts. Thanks to Richard Hayes Phillips for the use of his screengrab of the MO exit poll used here.

Candidates’ exit poll percentage/proportion derived from the gender category. Number of EP respondents: 1,636. Exit poll proportions rounded to nearest integer as appropriate for data derived from whole integers. As this first published exit poll was subsequently further adjusted towards conformity with the final computerized vote count, the currently published exit poll differs from the exit poll used here and available through the link below.

[2] Candidates’ percentage/proportion of the total computer-generated vote counts derived from reported counts (100% reporting). Published by The New York Times. Total number of voters: 664,378

[3] The difference between the exit poll proportion and reported vote proportion for each candidate (subtracting values in column two from the values in column three). A positive value indicates the candidate did better and received a greater share of the total reported count than projected by the exit poll. For example, candidate Sanders, reported percentage / proportion of the total vote decreased by 4.4% compared to his exit poll share.

[4] This column shows the percentage increase or decrease from the candidate’s exit poll projection (difference in column four divided by exit poll proportion in column two).  The practical use of this statistic shows that Sanders received 29,500 votes less than projected by the exit poll—a 11.4% reduction of his exit poll share.  Biden received 34,000 more votes than projected by the exit poll number or 9.3% less.  Shown only for candidates with 4% or more share in the exit poll. 

[5] This column presents a distinct Margin of Error (MOE) of the exit poll (EP) for the differences between candidate Biden and each of the other candidate’s EP results. The exit poll MOE, for example, between Biden and Sanders is 4.6%.  For simplicity MOE not shown for candidates with less than 4% share in the EP.  MOE calculated at 95% CI according to multinomial formula in:  Franklin, C. The ‘Margin of Error’ for Differences in Polls. University of Wisconsin, Madison, Wisconsin. October 2002, revised February 2007. Available at:  https://abcnews.go.com/images/PollingUnit/MOEFranklin.pdf 

[6] The discrepancy between the exit poll prediction and the reported computer-generated vote counts comparing Biden with each of the other candidates (subtracting each candidate’s difference between exit poll and computer count from Biden’s difference of 5.1%). Discrepancy of 9.6% between Sanders and Biden is more than double their MOE. These discrepancies are significant as they cannot be attributed to the margin of error.

[i] Fittingly, according to a recent 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.

[ii] In 2009 the Federal Constitutional Court of Germany ruled that every important aspect of an election must be observable by the public and thus “meet the constitutional requirements of the principle of the public nature of elections.” No amount of machine testing, security requirements, and licensing procedures can compensate for the constitutional requirement that “the essential steps of the electoral procedure being examined by the citizens.” And “trust in the regularity of the election [can] only [be realized] by the citizens themselves being able to reliably retrace the voting.”

The court also noted that while vote fraud with hand-counted ballots would be easy to detect, “programming errors in the software or deliberate electoral fraud committed by manipulating the software of electronic voting machines can be recognized only with difficulty.”

[iii] During the 2007 presidential election, eighty-three municipalities (France has 36,569 municipalities) were allowed to use voting machines. Due to security concerns and the inability of voters to determine if their votes are counted correctly a moratorium, that remains today, prevents additional municipalities from introducing voting machines. In the 2012 elections only 64 municipalities continued their use. The French government desires a total ban on their use.

[iv] In Canada, the results of federal elections are determined exclusively by hand-counted paper ballots.  Some provinces have adopted voting machines for local elections. See here, here and here.

[v] The United States’ long ballots–containing federal, state, and local races–are commonly cited as being unwieldy for hand-counting. The use of Sweden’s method of providing different colored paper ballots for federal, state, and local races that are then sorted prior to hand-counting addresses this objection and allows for at least the hand-counting of federal elections with only three races per ballot.

***

Download Missouri 2020 Democratic Party Primary Exit Poll. Published by CNN at poll’s closing on election night.

Comments made on this or related pages that may be helpful to the reader (comments made on another page will open a new tab):

Why the exit polls accessible today differ from the exit poll used here.

Explanation of the method used to calculate the margin of error appropriate for an election with multiple candidates.

On errors in the conduct of an exit poll as the source of the disparities between the exit polls and the unobservable computer counts