Given that the media is – mostly – in a full court press to ensure that voters do not make the apparent mistake of voting for Donald Trump, is it any surprise that polls are viewed skeptically? Especially given the wide divergence between different polls – or pollsters’ methodology to be more accurate – this late in the race. But are those skeptics – who tend to be Trump supporters – right?

There’s two main reasons why Trump could pull off a very unexpected upset on November 8th. The methodology of some pollsters is wrong. Or, people are lying.

Last Friday, 3 polls showed Trump leading slightly. The pollsters involved are:

LA Times: who get an A- rating from fivethirtyeight and have accurately predicted results 86% of the time.

Rasmussen: who get a C+ and have a 79% accuracy rating.

IBD/TIPP: who get an A- and have a 76% accuracy rating.

As you can see, one’s accuracy and one’s grade are not necessarily correlated. And it would be quite a slog for most of us to go through all the granular detail behind Nate Silver and fivethirtyeight’s methodology for ranking the methodology of other pollsters. So the best one can do is say that these 3 polls are from respectable pollsters and while they may very well be outliers, they might instead reflect real voter preferences. We’ll find out.

Are people lying to pollsters? Why would they? From shame? Is there a hushed army of bashful Trump fans out there? Or are they lying to pollsters due to a deep suspicion of pollsters and mainstream media in general? Even if linking these two sections of the communications industry is not always quite accurate. Both these reasons might be prompting more than the usual amount of misleading responses on the part of respondents. Again, we’ll find out.

While the Brexit results were propelled by similar political concerns on the part of British voters, the polls in the UK were significantly closer than the Real Clear Politics average currently shows for the presidential race. So a Brexit surprise is a little tougher ask on this side of the Atlantic. To pull it off, Trump has to reel in independents as well as bring back those GOP voters who have been turned away by the recent scandals.

But if he somehow does win, or makes it much closer than most are expecting, then absent a dramatic shift in the polls over the next 12 or so days, it will mean that polling is in trouble. Dead? The end of the polling era as an Observer article proclaimed? Not likely. But a Trump near-victory would send pollsters scrambling to update their methodologies (and that means everything from how often they try to contact you to more wonky statistical adjustments). And if Trump wins somehow? Pollsters will be even less trusted than the media. And even worse: they will have a tougher time getting people to pay them for their work. Unlike the generally profitable mainstream media.