During the last election cycle, even up through the first half of election night, every major media outlet predicted a comfortable victory for Hillary Clinton. Here you can see one much-referenced election calculator, timed out like a broken stopwatch, displaying the best prediction the pollsters could give us: 71% chance of a Clinton victory, last updated only hours before the election was called for Trump. Of course, in a way, these predictions weren’t all wrong. Hillary Clinton may have fallen short in the electoral college, but she won the popular vote by a substantial margin.

It’s interesting to look at where Hillary won. Of course she won the major blue states like New York and California. I want to zero in on one fact: Hillary Clinton won big in the areas where most media outlets are based. According to a new report in Politico (based in Washington, D.C.), “Nearly 90 percent of all internet publishing employees work in a county where Clinton won, and 75 percent of them work in a county that she won by more than 30 percentage points.”

Fully three quarters of “internet publishing employees,” which is to say, staff at cutting edge web-based journalism outfits like Slate, Salon, HuffPo, or The Daily Beast, not counting more traditional newspaper and television media, live in heavily blue counties. Not only are these scions of tomorrow’s media industry probably Democrats—only 7% of reporters identify as Republicans according to a Politico survey, down from 18% in 2002—they tend to live in trendy, gentrified, liberal areas like Williamsburg in Brooklyn, the Pearl District in Portland, or Silver Lake in Los Angeles.

Let’s bring this into the real world. When a reporter at HuffPo goes to Whole Foods, she will rub shoulders with Democrats and more Democrats. If she stood on the street corner below her loft apartment and took a poll, she would have been quite confident that Hillary Clinton was destined to give Donald Trump a drubbing at the polls.

This is technically called an “echo chamber.” When you believe something, and everyone else around you believes it, you believe it even more strongly. So when someone at HuffPo holds forth on how well Hillary Clinton will do among moderates, someone at Slate will be be even more convinced that his new piece on how well Hillary Clinton will do among Hispanics is accurate. And once these pieces have been published, “thought leaders” will pontificate on Twitter, the ultimate echo chamber, about how they had foreseen these trends months ago. And everyone will go home to their loft apartments, listening to an NPR podcast or Vampire Weekend. And everyone will be more convinced than ever that despite Trump’s improbable nomination, despite the groundswell of support for him in rural areas, despite Brexit and the winds of economic nationalism, Hillary Clinton has the election tied up with a little bow.

But of course, that’s not what happened. Not even close. But because the pollsters and the pundits don’t live in Mahoning County, Ohio, or Luzerne County, Pennsylvania (both of which shifted massively rightward in 2016), they didn’t see the way the wind was blowing. Before the Internet, news had to pass from place to place by physical transportation, or by phone or telegraph. In that more localized age, local newspapers carried stories, conducted polls, and created jobs in their local communities. Nowadays, the big new name in political polls is FiveThirtyEight, a NYC-based data journalism site founded by pollster Nate Silver, and owned by ESPN (based in Connecticut). All while regional newspapers are dying off at a frightening rate, as the Politico report illustrates.

The point here is that there is a geographical and cultural element to the truth. The prognosticators didn’t forecast Trump’s win because they didn’t live where it happened. Trump’s America, the parts of the country that were always for him and the parts that surprised everyone by flipping his way, are geographically and culturally different than the Left’s America. If you live in one, you have to go out of your way to understand the other. And journalists and pollsters in the “new media” aren’t going out of their way. That’s why they were blindsided by Trump’s improbable rise, and his almost unbelievable win in November.

If I’m right about this media blindside, it will have implications far beyond politics. The political geography of this country and its religious geography are tied hand-in-hand. How can Christians from Middle America hope to be understood, sympathized with, reported on fairly, by a media establishment with such a massive blind spot? How can media apparatchiks claim to be objective about, say, the issues of abortion or adoption by homosexual couples, when something as simple and obvious as geography kept them (in part) from accurately calling the election? Christians should move beyond the point where we “trust, but verify” media coverage and regard talking points as dubious until proven otherwise.

It is not surprising that most pundits, pollsters, and other avatars of the “new media” live in homogeneous, almost entirely secular enclaves that wealthy, white liberals have carved out on America’s coasts and in its college towns. What is surprising, and worrisome, is that this geographic blindness can keep the media from accurately and fairly covering issues of politics and religion. Even in the Internet age, where you are matters. And many journalists, it seems, are in the wrong places.