Amid the speculation on whether the electoral college will refuse to make Donald Trump president, many Trump opponents are pinning their hopes on one glaring fact: Hillary Clinton's sizable win in the popular vote.
Clinton's lead now exceeds 2.8 million votes (more than 2.1 percent of the total vote) and continues to grow. Many Democrats hope this fact alone might persuade Republican electors to reject Trump in favor of some alternative.
But this hope faces a serious challenge: Half of all Republicans actually think Trump won the popular vote.
In a nationally representative online survey of 1,011 Americans conducted by Qualtrics between Dec. 6 and 12, we asked respondents, "In last month's election, Donald Trump won the majority of votes in the electoral college. Who do you think won the most popular votes?"
Twenty-nine percent said Donald Trump won the popular vote. This is a slightly larger proportion than in a recent Pew survey in which 19 percent said Trump won the popular vote.
Respondents' correct understanding of the popular vote depended a great deal on partisanship. A large fraction of Republicans - 52 percent - said Trump won the popular vote, compared with only 7 percent of Democrats and 24 percent of independents. Among Republicans without any college education, the share was even larger: 60 percent, compared with 37 percent of Republicans with a college degree.
This same pattern of partisan bias didn't emerge on other factual questions in our survey. For example, we asked respondents to estimate the size of the country's African-American and Latino populations. As is typical, people tended to overestimate the size of these groups: On average, respondents think 27 percent of Americans are black and 28 percent are Latino. (The correct answers as of 2015 are 13.3 percent and 17.6 percent.)
But these numbers do not vary by party affiliation. Democrats and Republicans make similar guesses, on average.
These results align with something social scientists have long recognized: We choose facts to be consistent with our prior beliefs. In this case, Republicans are more likely to endorse erroneous claims about Trump's victory because it aligns with their partisanship. They do not, however, have any partisan motivation when estimating the size of minority groups.
If the Republican electors are anything like the party rank and file, they may think voting for Trump is in line with the choices of the American people.
John Sides is an associate professor of Political Science at George Washington University. He specializes in public opinion, voting, and American elections.