Donald Trump at the Fredericksburg Expo Center on Aug. 20. (Molly Riley/Agence France-Presse via Getty Images)

Much has been made lately of Democratic presidential nominee Hillary Clinton’s substantial polling lead in Virginia, combined with her campaign pulling ad buys from the state and reports that Republican presidential nominee Donald Trump’s campaign is focusing its ads on four rather than five swing states, with Virginia so far off the list. Some observers suggest Virginia is blue this year.

This once-reliably Republican state at the presidential level has undergone a remarkable political transformation in a relatively short time. From 1952 to 2004, Virginia went to the GOP in every presidential election cycle except the Lyndon Johnson landslide of 1964. It was the only former Confederate state not to support Southerner Jimmy Carter in 1976.

Barack Obama’s 2008 win here broke the GOP’s perceived presidential hold on the Old Dominion. Demographic changes, particularly the rising population of the more liberal urban corridor, substantial growth statewide in minority communities that vote heavily Democratic and an impressive voter-turnout operation by Obama’s campaign contributed to this major shift.

After two consecutive presidential victories for the Democrats against well-regarded GOP opponents Sen. John McCain (Ariz.) and former Massachusetts governor Mitt Romney, and an acceleration of the demographic shifts that seemingly turned Virginia blue, why would anyone take seriously the idea that Trump has the formula to turn Virginia red? It doesn’t seem credible.

Unless you consider that in many states in the primaries, Trump substantially outperformed the results of public opinion polls. Sometimes the difference was dramatic — in or near double digits. In Maryland, Trump’s vote percentage beat the average of major pre-primary polls by 6 percentage points. A recent statewide poll in Virginia has Clinton up by 8 percentage points. Apparently that has made a lot of her supporters feel very comfortable.

I am not betting on Trump in Virginia. But some electoral background gives room for pause. Simply put, what we might call the reverse Wilder effect may be at play in this election, giving Trump a credible chance of winning Virginia despite what polls are telling us.

Democrat L. Douglas Wilder’s 1989 gubernatorial win in Virginia made him the nation’s first elected African American governor. His victory was touted as a historic shift in U.S. politics, especially because it happened in the one-time capital of the Confederacy. Wilder won what was at the time the closest statewide race ever, by less than 7,000 votes (out of 1.8 million), leading to a recount. Pre-election polls had him way up over his GOP opponent. Even more telling, exit polls that asked people how they voted showed that Wilder should have won the race in a landslide. How could the polls, and especially exit polls, have been so incredibly wrong?

Analysts suggested that many voters lied. When asked by a pollster, they concealed their preferences to give what they considered the socially acceptable answer. Thus, to admit to supporting the Republican was to openly profess not being on the side of making history by electing a black governor for the first time. This outcome reminded many of what previously had been called the “Bradley effect.” In 1982, Los Angeles Mayor Tom Bradley lost his bid to become the nation’s first elected African American governor, despite pre-election polls showing him clearly leading.

Could there be a similar effect going on in this election, with many voters secretly supporting Trump while telling friends, relatives and pollsters that they will not vote for him? This would represent perfectly rational behavior on the part of those Trump supporters who have good reason to believe that many will look down on them for admitting their true preferences. It is natural for people to want to be seen as smart, open-minded and tolerant.

There is no reliable way to test the existence or the extent of hidden Trump support. But given that his voting percentage in primaries showed that he often substantially outperformed pre-vote polls, the current margin in favor of Clinton doesn’t seem strong enough to count Trump out. Well-informed observers are taking Virginia off the electoral map. Don’t be fooled. Trump has a chance here.

The writer is dean of the Schar School of Policy and Government at George Mason University and a longtime analyst of Virginia politics.