“Dirty Secret”: why the polls were so wrong

Thursday, November 05, 2020

By Benjamin Konietzny

Surveys are rarely accurate, that’s normal. However, the differences in elections in the United States are unusually high. There are several reasons why this could be the case.

It has not yet been determined who won or lost in the US presidential election two days after the start of the election. There is also much to suggest that it will be some time before the winner is finally clear. But many Americans have now identified a loser: pollsters. In fact, some of the predictions differed significantly from the actual results.

In the state of Ohio, for example, the latest figures from Real Clear Politics predicted a narrow victory for Donald Trump. In the poll, the incumbent was one percentage point ahead of challenger Joe Biden. Now it’s clear: Trump is more than eight percentage points ahead of Biden in Ohio. In Florida, “FiveThirtyEight,” also one of the best-known poll sources in the United States, predicted Biden’s victory by 2.5 percentage points. It turned out differently: Trump won in the Sunshine State, by a 3.4-point margin. According to a Real Clear Politics poll, Biden should win by 2 percentage points in Iowa. What happened? Trump took the state by more than 8 percentage points.

It should be noted that neither “Real Clear Politics” nor “FiveThirtyEight” conduct polls by themselves. Both portals aggregate polls from different institutes and therefore claim to be able to make particularly accurate predictions. But they weren’t the only ones wrong. Multiple polls across the country recorded a stable lead for Biden for weeks before the election. On average, the Democrat was nearly 10 percentage points ahead of Trump. In reality, however, this advantage is less than 2 percentage points. Where does this important discrepancy come from?

There are several possible reasons why surveys fail:

1. This year’s turnout was unusually high, the highest in more than 100 years. In the past, high turnout has led primarily to good Democratic results. This is due to the fact that in previous elections the “blue” candidates, that is, the Democrats, were able to attract a disproportionately high number of minority members to the polls in these cases. In this year’s election, however, Trump in particular was able to mobilize his clientele. The proportion of white voters without a college degree was unexpectedly high.
2. US political scientist Salvatore Babones describes another reason in a guest article for the “Sydney Morning Herald” as the “dirty secret of public opinion polls.” Consequently, response rates to telephone surveys in the US have completely collapsed. According to the Pew Survey Institute, 20 years ago 36 percent of those called participated in surveys by telephone. This figure is now 6 percent. Babones also writes that there are rumors in the industry that only about 3 percent will participate. His guess as to what could be behind this is: in an age when almost everyone is equipped with a smartphone, hardly anyone answers calls with a hidden number. However, what speaks against this thesis is that many online surveys were not more accurate either.
3. An effect that also exists in other countries, and is also known in Germany, could also have played a role. This is the concept of social desirability developed in 1972 by sociologists Derek Philipps and Kevin Clancy. According to this, some people knowingly refuse to take surveys or give wrong answers because they fear that providing truthful information violates social norms. As a result, predictions about opinions outside the mainstream are significantly less reliable.
   In the elections in Germany, this effect could be observed in the surprising success of the Republicans in the state elections of Baden-Württemberg in 1992 and in the state elections of Saxony-Anhalt in 1998, where the far-right DVU obtained 12.9 per percent of the votes from zero, according to the latest polls. the party in only 6 percent. And the electoral results of the AfD are traditionally characterized by this effect, in some cases considerably. In the last state elections in Baden-Württemberg, Rhineland-Palatinate and Berlin, the polls on election day (exit polls) showed a deviation of 25 percent. In the 2013 federal elections it was almost 50 percent.
   In the United States, this effect is also explained with “timid Trump voters”, with “timid” Trump voters. It cannot yet be said how big the effect really is. The Republican camp has also tried in the past to develop a political narrative from it: that of the “silent majority,” the silent majority who supposedly back Trump. This is supported by the fact that a poll in Florida, which was conducted by computer voting, was much closer to the actual result. Because in front of a computer nobody has to be ashamed of their preferences. Against this, as mentioned above, the fact that even anonymous online surveys were sometimes unable to provide accurate results.
4. Trump may have been able to convince many Americans in the last meters of the election campaign. Unlike his competitor, the president held massive events in major changing states despite the corona risk. The incumbent may have been able to mobilize many voters in the last days before the elections, and it was too late for this development to be reflected in the polls.

In this context, it should be remembered that surveys have also been wrong in the past. In 2016, Trump was also underestimated by polls, in some of the undecided states by 4 percentage points. And in 2012, too, the forecasts for Barack Obama were off by roughly the same amount. However, the scope of this year’s deviations is new.