Crowd counts: When The Post did it right
Sunday’s ombudsman column urged The Post to provide estimates of crowd sizes for rallies on the Mall rather than allowing event organizers to offer their own counts that go unchallenged. Actually, The Post did precisely what I suggested for President Obama’s 2009 inauguration, and it still serves as a good example of how readers can be well served by authoritative estimates.
Relying on satellite images, The Post was able to provide what remains a widely accepted estimate of about 1 million people between the Capitol and the Lincoln Memorial prior to the noon swearing in. Looking back on that estimate, what I liked most was how The Post explained it to readers in a terrific aerial graphic of the Mall.
It not only provided the satellite image online, but The Post also gave readers a step-by-step explanation of how the photo was analyzed in order to make an estimate. It noted those who were not included (people along the parade route or in side streets, for instance). Then it explained what experts calculate to be the area occupied by a single person in a crowd that is jam-packed (2.5 square feet), reasonably packed (up to 4.5 square feet) or loosely packed (10 square feet). The giant crowd was divided into zones, and each was assigned a density. Next, The Post noted what percentage of each zone was occupied by people. Finally, all the zones were totaled to produce the estimate.
Post database editor Dan Keating, who was largely responsible for the estimate, recalled this week that before making his crowd count he conferred with a retired U.S. Park Service official who made his own estimate using the same satellite image supplied by GeoEye. (About 14 years ago, Congress ordered the Park Service to stop counting crowds after organizers of the 1995 Million Man March threatened to sue because they felt the agency’s estimate was too low.) The retired official, working independently of the agency and The Post, came up with an inauguration crowd count that was only slightly higher.
Unfortunately, The Post’s news story at the time highlighted an estimate of 1.8 million released by the District government -- wildly higher than the 1 million estimated by The Post itself. The Post story even used the higher 1.8 million figure in the headline, and it got picked up by other media, thus diminishing The Post’s own more credible 1 million estimate. It still rankles Keating.
Sunday’s column prompted numerous e-mails and calls from readers who offered suggestions on how The Post could improve its crowd counting without relying on expensive satellite imagery. Many focused on the recent “Rally to Restore Sanity and/or Fear,” headlined by Jon Stewart and Stephen Colbert. The Post opted not to estimate that crowd, but its news coverage did note heightened ridership on Metro.
“They have exit polls at elections over a wide area. Why not have exit polls at airports and bus stations for major Mall events?” asked Lou Spevack of Brookeville, Md. “It seems to me that there are unexploited means to develop a methodology for measuring attendance at Mall events. What matters is that the methodology be applied consistently.”
Other readers noted the inherent difficulty in measuring crowds that grow or shrink within a relatively short period.
“I don’t believe anyone has noted that attendance should be keyed to the purpose and length of the event,” wrote Ken Masugi, who lives in Montgomery Village, Md. “If three students attend a 50-minute class for five minutes, they should not be counted as present for the class. But if people check in and out of a Mall event such as the recent rallies, they should count as having attended; maybe one event or personality was the focus of their attention. The purposes of the Mall event and of the lecture differ, and so should the method of counting attendance.”
Masugi said that he had attended the late August rally featuring conservative commentator Glenn Beck, “just to see what the crowd was like.” Masugi said he stayed for less than an hour, but “I noticed hundreds of people leaving the rally as I walked down 23rd St. Some complained they could not see or hear anything. Yet all showed up out of curiosity or to show support, and so should be counted as having attended, however briefly.”
“What crowd estimators should cease doing,” he said, “is to think that snapshots can capture the overall attendance, especially at lengthy events on hot, humid days.”
I also heard from Julian Bond, the Civil Rights Movement leader and former NAACP chairman, who wondered why The Post has not mentioned last month’s “One Nation Working Together” in comparing crowd sizes. That rally on the Mall included large numbers of self-described liberals and progressives, and its organizers said it drew 175,000 people. However, The Post did not offer independent estimates in its news coverage.
Most experts agree that images taken from high above a crowd are the best way to evaluate its size. And they caution that photos taken from lower altitudes -- even those looking down from the Washington Monument -- can present a misleading picture.
“The Obama inauguration is a great example of that,” said Steve Doig, a crowd-counting expert at Arizona State University’s Cronkite School of Journalism. “The TV and still images taken from the podium area (at the Capitol) looked like an unbroken sea of people back past the Washington Monument. But the GeoEye satellite images revealed how people were clustered near the JumboTrons but were much sparser in other areas.” Doig’s estimate of 800,000 was lower than the 1 million offered by The Post.
“Crowd estimation is difficult in D.C. because you get large free-entry events, don't have many high vantage points from which to take photos, and air space is very restricted,” Doig said in an e-mail. “I appreciate the problems the Post faces in covering those kinds of things, but I really feel it behooves the hometown paper to make a stronger effort on crowd estimation. As a news consumer, I was disappointed in how the Post...and most other news media punted on the Beck estimate and let the ridiculously high estimates go pretty much unchallenged.”
Beck claimed his “Restoring Honor” rally attracted “between 300,000 and 500,000” participants. But an estimate for CBS News, prepared by Digital Design & Imaging Service, Inc. of Falls Church, put the crowd size at 87,000. The CBS/Digital Design estimate for the Stewart/Colbert rally: 215,000.
For the next large rally on the Mall, Doig suggests that major media groups, including The Post, create a consortium to share the cost of providing satellite imagery and analysis to create credible crowd estimates. I think it’s a good idea.
“Barring that for whatever reasons,” Doig added, “I can think of a ground-level approach that would give you a defendable estimate. In advance of the rally, select a variety of locations -- at least a couple dozen -- spread evenly across the likely crowd area. Get a roster of volunteers -- staff, interns, j-school students, interested readers, whatever -- and assign each a location. Have each of them bring a cell phone camera. Agree in advance on a moment at the height of the gathering, like the start of the main speech or whatever. At that moment, each of your crew should hold up their camera and take pictures or video of the crowd around them. Then they should all head back to the Post to download their images (you can't count on cell phone service or Internet at these mass gatherings because the networks get overloaded). With the collection of images as a good sample through the crowd, Dan Keating or one of your other number crunchers can put together a good sample of densities and have a sense of the extent of the crowd, enough to make a decent fact-based estimate.”
Posted by: exNYer1 | November 12, 2010 1:08 PM | Report abuse
Posted by: B2O2 | November 15, 2010 5:49 PM | Report abuse
Posted by: B2O2 | November 15, 2010 9:37 PM | Report abuse
The comments to this entry are closed.