Americans overwhelmingly embrace the principle that all votes should count equally. For a candidate to win the most votes but lose an election violates this principle. Yet in America’s last five presidential elections, the electoral college twice delivered victory to a candidate who lost the popular vote. Forecasts from the Economist and FiveThirtyEight analyst Nate Silver suggest such a scenario is again possible. The current probability of President Trump losing the popular vote but winning the electoral college is about 10 percent.

If the electoral college winner loses the popular vote again this year, how would it affect Americans’ faith in the election’s outcome? Our new study shows that the damage would be substantial. Many Americans would see the outcome as less legitimate, even if their own party’s candidate benefits.

But partisanship does matter — just in a different way. Regardless of which party wins, Democrats see presidents who win the electoral college and lose the popular vote as less legitimate than those who win the popular vote. Republicans make no such distinction.

How we did our research

We conducted experiments in two surveys. The first was fielded online from March 23 to 30 by YouGov to a non-probability sample of 3,395 Americans that was weighted to mirror the U.S. population. The second was fielded online from May 12 to 22 to 7,749 Democratic and Republican identifiers recruited from the Lucid survey marketplace. The survey closely resembles the U.S. population in terms of age, gender, race, Hispanic ethnicity and region.

In each experiment, participants were asked to imagine that one candidate wins the electoral college and thus the presidency. The experiments randomly varied whether this candidate was in the respondent’s party or the opposite party. They also varied whether the candidate won the popular vote or not, and the size of the popular vote margin, ranging from a three-point victory for the electoral college winner to a five-point loss. (For comparison, Trump lost the popular vote by about two points in 2016.) We then asked participants whether they regarded the outcome as legitimate, as fair, and as having selected the rightful winner.

First, we find that Democrats and Republicans view the outcome as more legitimate when their party wins. This finding is consistent with an established pattern in which supporters of winning parties report greater trust in the results and satisfaction with democracy than do supporters of losing parties.

But Democrats also view the outcome as less legitimate if the electoral college winner does not win the popular vote. In this case, the proportion of Democrats who say the winning candidate’s presidency is “entirely” or “somewhat” legitimate decreases by 28 percentage points in the March experiment (from 84 percent to 56 percent) and by 18 percentage points in the May experiment (from 81 percent to 63 percent). Similarly, the proportion of Democrats who say the winning candidate was “definitely” or “probably” the rightful winner drops by 23 and 35 percentage points across the two studies.

Crucially, Democrats were less likely to think the election was legitimate even when told it was the Democratic candidate who won the electoral college and lost the popular vote. And surprisingly, the margin by which the winner loses the popular vote did not matter. Democrats appear to believe that any discrepancy in who wins the electoral college and the national popular vote is problematic.

By contrast, Republicans are seemingly not bothered by this discrepancy. Like Democrats, Republicans judged the election to be less legitimate when the other party won. But unlike Democrats, they did not impose any “legitimacy penalty” when a candidate won the electoral college but not the popular vote — regardless of whether the winner was a Democrat or Republican. For instance, the proportion of Republicans in the March study who said the electoral college winner is “entirely” or “somewhat” legitimate was 85 percent if they won the popular vote by 1 percent, and 89 percent if they lost the popular vote by 1 percent. In the May study, the corresponding percentages were 90 percent and 88 percent.

Prospects for 2020 and beyond

This difference between Republicans and Democrats probably stems from their different recent experiences with the electoral college. Of the Democrats’ last four candidates, two won the popular vote and lost the presidency, a pattern that is likely to continue benefiting Republicans, according to current research. By now, Democrats may be primed to view this scenario as illegitimate. A third election in which the party loses the electoral college but wins the popular vote would surely widen the party divide.

If the situation were reversed and a Democratic candidate won the electoral college but not the popular vote, Democrats might change their mind and see this outcome as legitimate after all. Republicans, of course, might flip-flop in the opposite direction.

And this suggests the challenge: Until both parties confront the prospect of the electoral college reversing their popular vote victories, bipartisan support of reforms binding election outcomes to the popular vote will remain elusive.

John M. Carey is the Wentworth Professor in the Social Sciences at Dartmouth College.

Gretchen Helmke is professor of political science at the University of Rochester.

Brendan Nyhan is professor of government at Dartmouth College.

Mitch Sanders is a principal at Meliora Research.

Susan C. Stokes is the Tiffany and Margaret Blake Professor at the University of Chicago.

Shun Yamaya is a graduate student in political science at Stanford University.