Home > News > Do Republicans really believe Trump won the 2020 election? Our research suggests that they do.
116 views 7 min 0 Comment

Do Republicans really believe Trump won the 2020 election? Our research suggests that they do.

- January 6, 2022

All credible evidence tells us that the 2020 election was very secure. Experts on both sides of the political aisle, and even President Donald Trump’s own Justice Department, have confirmed that 2020 was a free and fair election. Even a Republican-sponsored audit of Arizona’s results found no evidence of fraud or malfeasance.

Nevertheless, the vast majority of Republican voters say they agree with Trump’s unsubstantiated claims that the election was stolen. In our most recent University of Massachusetts at Amherst poll, fielded online Dec. 14-20 by YouGov among a nationally representative sample of the U.S. voting-age population, only 21 percent of Republicans say Joe Biden’s victory was legitimate. This is nearly identical to what we found in our April poll, in which just 19 percent of Republicans said Biden was legitimately elected. Other universities, media outlets and polling firms have found nearly identical results.

How could the “big lie” campaign convince so many Republicans that Trump won an election he so clearly lost? Some observers wonder whether these beliefs are genuine or just an example of expressive responding, a term social scientists use to mean respondents are using a survey item to register a feeling rather than express a real belief. In this case, it would mean that these Republicans, upset about Biden winning, say his victory was not legitimate even though they know deep down that it was.

What does the evidence tell us?

While it is difficult to firmly establish what respondents truly believe, clues suggest this is a genuine belief. One piece of evidence is that the result is nearly identical in phone surveys and online surveys. When people respond to phone surveys, those responses are often biased by what social scientists call “social desirability,” in which respondents say what they think makes them look good — even if that is not what they actually believe. Web-based surveys are known to reduce social desirability bias.

Other survey responses appear consistent with a true belief that the election was stolen. In our December UMass Poll, we asked those who said Biden’s presidential victory was illegitimate to select all the reasons they believed so from a list of conspiracy theories floated by those pushing the “big lie.” As you can see below, fully 83 percent say that “fraudulent ballots supporting Joe Biden were counted by election officials”; 81 percent that officials counted “absentee ballots from deceased people”; 76 percent tell us that “non-citizens and other ineligible voters were allowed to vote for Joe Biden”; 69 percent that the victory was illegitimate because “some states changed election rules in ways they should not have”; and 65 percent that election officials destroyed ballots supporting Trump.

In other words, not only do they say that Biden’s victory was not legitimate, but they endorse several (though not all) specific theories about how fraud was perpetrated.

Further, Republicans in our UMass Poll say they would be more likely to vote for 2022 GOP congressional candidates who questioned Biden’s victory and less likely to vote for those who concede that Biden won. Using a conjoint experiment, an approach likely to reduce social desirability bias and expressive response, political scientists Vin Arceneaux and Rory Truex found that Republicans do indeed reward and punish candidates in this way.

How we did further research

To further explore how genuinely Republicans believe the “big lie,” we embedded a list experiment in our December UMass Poll. A list experiment is a technique scholars use to reduce social desirability bias and expressive responding, and it lets us see what percentage of respondents select an item without having to tell us so directly. This way they neither have to tell a pollster something embarrassing or feel the reward of telling a pollster something that expresses a feeling rather than a belief.

In our study, we divided respondents into two groups. We asked all of them: “How many of the following do you believe? Don’t tell us which ones. Just how many.” However, the control group was asked about four statements, while the other group was asked those four, with the addition of this: “Joe Biden’s victory in the 2020 presidential election was legitimate.” You can see all five statements in the survey image below.

Since the lists are otherwise identical, we can see what percentage selected the additional sentence by comparing the control group’s average number with the second group’s average number.

The survey experiment question in the December UMass Poll.
The survey experiment question in the December UMass Poll.

Only 28 percent of Republicans believe that Biden was legitimately elected. That’s what we find when we subtract the average number of statements the Republicans in the control group claim to believe (1.80) from the average number of items the Republicans in the experimental group claim to believe (2.08), giving us 0.28. That’s 28 percent.

That’s almost identical to the percentage of Republicans in our survey (27 percent) who, when asked directly, either told us that Biden’s victory was legitimate (21 percent) or selected “I’m not sure” (6 percent).

Apparently, Republicans are reporting a genuine belief that Biden’s election was illegitimate. If anything, a few Republicans may, for social desirability reasons, be using the “I’m not sure” option to hide their true belief that the election was stolen.

Republican respondents consistently tell pollsters that they doubt the legitimacy of Biden’s election. Apparently, that’s a genuinely held belief.

Alexander Theodoridis (@AGTheodoridis) is an associate professor of political science at the University of Massachusetts at Amherst and associate director of the UMass Poll.

Lane Cuthbert (@LaneGCuthbert) is a PhD student in political science at the University of Massachusetts at Amherst and a research fellow for the UMass Poll.