In the 1950?s and early 1960?s if you ever saw photographs of a politician?s headquarters on election night, did you ever wonder who those people on the telephones were talking to? They were listening to reports from their contracted pollsters, giving them information on what voters were saying as they left the precincts. Occasionally there would be huddles as groups compared notes.
That data was what America?s electronic media introduced to us as ?exit polling.? This was the new ?gimmick? each of the networks would use in the late 1960s to hook viewers into watching their coverage. Now the pundits were going to have the information on what the voters really thought; we wouldn?t have to stay up all night for results. Magically only minutes after the voting ceased, they would be able to tell us who had won. It was going to be wonderful. The shadowy dark world of politicians and their pollsters was coming into the light. At long last, drama was returning to election night broadcasting.
A Brief History Of Exit Polling
The statistical concept is simple to understand. Determine the demographic factors most in alignment with the candidate’s views, then interview those leaving the polling place who matched those characteristics about how they voted. PRESTO! A totaling would reveal the mathematical likelihood the politician would win. It required a great massing of data, analysis, and armies of volunteers with clipboards and questionnaires; historical data needed to be amassed and analyzed. Precincts matching the desired results needed to be identified. Computers had a useful and necessary purpose.
At long last science was going to trump guesswork. Punch cards took the place of pundits. And all night diners and take-outs were going to lose tons of money since they would no longer be needed to keep those TV anchors aware and alert and well-fed.
The first major exit poll was for a network during the 1967 Kentucky governor’s race. By the 1970s, exit polling had become an industry practice. In 1980, NBC, relying on its painstakingly gathered data, reported Ronald Reagan’s 1980 victory over Jimmy Carter almost immediately after the voting ceased on the East Coast, and nearly three hours before polls closed on the West Coast. There were Congressional hearings on whether it discouraged voter turnout.
To avoid federal strangulation of the practice, networks promised to not project a state’s winners until polls in that location closed. Local jurisdictions made attempts to restrict it, but since exit polling is protected by the First Amendment there have been few successes. Ironically, the U.S. government is a big supporter of exit polling abroad: the practice is widely used by pollsters hired by monitors to verify if elections are being conducted legitimately. The U.S. government has even financed exit polls in former Soviet republics and satellites to ensure votes are counted accurately.
For a brief time in the early 1990s, the network election consortium offered the promise of an end to the network competition to make calls first. On Election Night 1992, Voter Reporting Service, the precursor of Voter News Service (VNS), made calls that each network reported. It seemed that the networks had stepped onto a new playing field. No longer would they compete to call races first; instead, the competition would be to try to offer the best analysis, or produce the most interesting programming.
That new playing field was plowed under during the very next election, in 1994, when ABC hired its own consultants to give the network an edge. It worked. With its strengthened Decision Desk, ABC beat VNS in calling Senator Charles Robb as winning in Virginia and Governor Mario Cuomo as losing in New York. By 1996, each network had made the decision to hire analysts and operate its own Decision Desk.
Then, in 2000, relying on data from VNS, the networks declared the race for Al Gore around 8 pm, only to switch to George W. Bush at 2 am and declared the race locked at ?too close to call? two hours later. Blame varied from ?complete data? to ?premature data? to ?data input error.?
Computer glitches in 2002 and then in 2004 exit poll data was leaked online around midday on Election Day, and bloggers declared John Kerry the presumptive winner. In 2006, the pollsters began quarantining representatives of the service to prevent such leaks from occurring.
An Even More Recent History
The magic began to crumble. Americans in 2012 were told before the presidential elections Mr. Obama was struggling and would likely not be re-elected. Mitt Romney was poised to sweep into the White House. But the props were knocked out from under the media as a wave of unanticipated (or misjudged) support emerged and the president won re-election. Surprise, surprise.
The dominoes didn?t stop falling then. In the fall of 2014, the British held a general election and polls showed a crumbling of support for the Conservatives, so much so the BBC and ITV were discussing the possibilities of a coalition government in the first hours after the polls closed. A few weeks later American Democrats were shocked when an unsuspected flood of support for Republican candidates turned into ?a wave.? In 2015, commentators in Israel were proclaiming the end of the Benjamin Netanyahu era, only to have him finish with a resounding triumph over his opposition. The British once more were having elections in 2015 and this time, the pundits said, the right wing was doomed.
David Cameron, Conservative, appointed in 2010, was elected by an overwhelming majority that also brought many more of his party into the House of Commons and the House of Lords.
In each of these five elections, those pontificating talking heads were backtracking as the nights wore on. Early projections were retracted. Having learned from the 2000 disaster, most media outlets, once they saw their first calls were going to be wrong, didn?t venture new opinions, preferring instead to declare the matter ?too close to call.?
So, Do They Work Or Not?
Exit polls, like other surveys, are subject to statistical distribution errors. Therefore, before news organizations report any results or make projections, the results are compared to pre-election polls and precinct voting history. As the actual vote count comes in, the exit poll results are updated to reflect that information.
Exit polls, properly conducted, can remove most sources of polling error. Exit polls are not confounded by speculation about who will actually show up to vote, or by voters who decides to change their mind in the final moments. They condense identify the entire voting populations into representative precincts and can obtain very large samples in a cost-effective manner, thus providing even greater degrees of reliability. So long as your methodology is good and you read your measures correctly, your results will be highly accurate.
Steve Freeman and Election Integrity.org on the website goes to great lengths to conclude the 2004 presidential election was manipulated in some way since the exit polling data (that was leaked earlier in the day) ?clearly? showed John Kerry had defeated George W. Bush. Mr. Freeman presents arguments but little hard evidence that a fraud was perpetrated.
Eight years later the reverse scenario played out. In 2012 many cases of confirmed fraud were reported, but the disposition of these has long since faded away without any final resolution. Because of the imposed embargo on the release of exit polling information in place in 2012, commentators were a bit hamstrung in supporting their positions.
In truth, exit polling doesn?t do a good job of predicting outcomes; it exists to explain them. They help develop statistical models of voting behavior, and ought not to be used to predict final results. That is exactly what happened in 2004, when media outlets used non-predictive data in predictive ways, and while the data sets were still being compiled.
In the end, though, the completed exit polling only gives us a short-lived sneak peek at potential results. Within a couple of hours in most states, we?ll get the actual results anyway. The best course is to be patient, but in any case, don?t put any value in exit polling until those data sets are complete.
What do you think – does exit polling work, is it an accurate measure of results? Do you believe coverage of election results needs to be more strongly regulated, either by industry watchdogs or by the government? Do you think voter fraud is on the increase in the United States? Why do you think exit polls have been such poor reflectors of recent election results?