A crowd scientist says Trump’s inauguration attendance was pretty average

This Saturday very likely marked the largest day of demonstrations in American history. But Friday’s inauguration had, well, pretty average turnout for a presidential inauguration. And the new administration had a problem with that, as we saw with an alarming series of false statements from President Donald Trump and his press secretary Sean Spicer on Saturday.

To try to settle the question of how many people attended the inauguration ceremony on the National Mall in Washington, DC, we reached out to Keith Still, a professor of crowd science at Manchester Metropolitan University in the UK. (He analyzed aerial shots of the crowds of both President Obama’s 2009 inaugural address and President Trump’s for the New York Times.) His conclusion is that the crowd on the Mall on Friday was roughly one-third the size of President Obama’s.

If Still is right, and Trump’s inauguration attracted a third of the 2009 crowd, then there were anywhere from 300,000 to 600,000 people on the Mall on Friday.

President Obama’s historic 2009 inaugural address drew 1.8 million people, which officials consider the largest gathering on the Mall ever. But as Still told Vox, not all 1.8 million were on the Mall — crowds extended behind the Washington Monument and outside the view of the aerial photos used in analysis.

As you can see in the chart above, presidential inauguration attendance isn’t very consistent. In recent years, the first inauguration has been better attended than the second for a two-term president. But this wasn’t true for President George W. Bush — his second inauguration purportedly drew a larger crowd than the first.

The inaugural attendance numbers above largely come from various news outlets that arrived at these estimates through interviews and on-the-ground reporting with inauguration planners to assess crowd size. The National Park Service and the Office of the DC Mayor told me they won’t be releasing estimates of Friday’s event. (NPS was banned from releasing official crowd estimates after an especially controversial count involving attendance at the Million Man March in 1995. In 2009 it broke tradition and provided the 1.8 million estimate for Obama’s inauguration attendance.)

As you can see in this tweet from Matt Viser at the Boston Globe, President Trump might have perceived a large crowd from the podium, but as you move farther down the Mall, the crowd is less tightly packed.

“When you are barely above the head level of a crowd that extends back several hundred yards, it does look an unbroken sea of humanity,” said Steve Doig, a professor at Arizona State University with experience in estimating large crowds. “But people’s ability to estimate numbers bigger than what they can count on their fingers is not something we do well.”

This is a satellite image taken from space that helped analysts measure crowd attendance at Obama’s inauguration in 2009.

Charles Seife, a mathematician and professor of journalism at New York University, has written about the perils of counting crowds. He told me that using satellite imagery is one of the best ways to measure the size of a crowd if it’s not a strictly ticketed event, but warned it’s not a perfect tool.

But unfortunately, given Friday’s inclement weather we’re unlikely to get comparable images with such clear resolution for the recent gatherings on the Mall.

Doig said that for future events, organizers and journalists should work to get real-time estimates that don’t rely on aerial photos, especially for large events in DC, as DC falls under a no-fly zone, making it difficult to get aerial shots to begin with.

Doig said one thing you could do to get reliable real-time estimates is to have a team of 20 or so people scattered throughout a large event, and have them measure by square feet how closely people are standing next to each other.

“I’d assign random locations for people to stand, and the key thing they’d have to report is how dense the crowd is, measuring it by square feet,” he said. “Then we’d use those estimates on a grid [of the Mall’s area] to average across and total up what we might think is an overall reasonable count.”

Doig thinks a methodology like this could be implemented relatively easily and only take 30 minutes or so to provide reasonably accurate estimates.

As Vox’s Ezra Klein writes, what is truly frightening in how the Trump administration has responded to media stories covering the size of the crowd present at the inauguration is its apparent utter disregard for facts.

Among Spicer’s slew of falsehoods on Saturday was this: “This was the largest audience ever to witness an inauguration, both in person and around the globe.”

This simply isn’t true. Metro ridership did not surpass 2009 levels, and television ratings indicate not as many people tuned in as they did in 2009 either.

And as Doig told me, “[Spicer] offered no reason for us to believe they have attempted to do a reality-based estimate of the crowd, other than quoting President Trump saying it looked like a million and a half [people].”

As my colleague Alvin Chang illustrates, the consequences of the Trump administration repudiating numbers it simply doesn’t agree with could have dangerous repercussions when it comes to government data sets we rely on for accuracy and the health of democracy.