< The Art and Artifice of Counting Crowds


Friday, September 03, 2010

BROOKE GLADSTONE: Sometimes it’s hard to know who to trust, especially when it comes to numbers.


GLENN BECK: I have just gotten word from the media that there is over 1,000 people here today.



BROOKE GLADSTONE: On August 28th, from the steps of the Lincoln Memorial, Glenn Beck tweaked the press as he surveyed the crowd attending his “Restoring Honor” rally because he knew the truth. As he said later -

GLENN BECK: We'll have aerial photography here shortly on the numbers, but I can tell you that it was in the hundreds of thousands. Let's be on the low end, 300,000, and maybe as high as 650,000.

BROOKE GLADSTONE: But Minnesota Republican Michele Bachmann at her own post-Beck rally disagreed.


MICHELE BACHMANN: We're not going to let anybody get away with telling us there were fewer than a million people 'cause we were here. We are witnesses!


BROOKE GLADSTONE: Crowd estimates flew overhead like the geese over the rally. ABC News went with 100,000 plus, NBC News with some 300,000, and NPR with tens of thousands. Meanwhile, CBS News actually paid for some research and arrived at a figure between 80 and 90,000. Stephen Doig helped calculate it by using aerial photos.

STEVE DOIG: People way overestimate how many people can be packed into a given area using like two square feet per person. That is, you know, scary, mosh pit, dangerous kind of density. People couldn't fall over if they fainted.

BROOKE GLADSTONE: Doig says Washington is a tough town for crowd counters to get an angle on.

STEVE DOIG: There aren't very many tall buildings. I think there’s a municipal ordinance that doesn't allow anything over just a very few stories. The highest thing in D.C., to my knowledge, is the Washington Monument. But it was useless for crowd counting because the angle was too oblique. It wasn't overhead enough. So CBS went to a company called AirPhotosLive, which uses tethered balloons with cameras on them to get really beautiful detailed overhead images. Then I used their images to make my honest estimates of the density of the crowds in different parts of that whole sprawling area that it was in. It’s actually fairly low-tech science.

BROOKE GLADSTONE: You estimated about 80,000 and the other expert that CBS hired estimated about 87,000. They went with 87,000. You said, you know, same difference.

STEVE DOIG: Absolutely. I would never swear that exactly 80,000 people were there. Actually, I had one commenter on my blog about this who sent me a note saying, well, there had to be at least 100 people in Porta Potties at the time of the image. And I said, okay, my estimate now is 80,100.

BROOKE GLADSTONE: [LAUGHS] As for that 300,000 NBC number, that was provided by an NBC news staffer, Domenico Montanaro, who Tweeted last Saturday that that’s when an unnamed Park Service official said – NBC’s Brian Williams.

BRIAN WILLIAMS: At Glenn Beck’s rally on the steps of the Lincoln Memorial yesterday, he was able to attract a crowd – I've seen estimates 500,000. NBC News estimated the crowd at 300,000.

BROOKE GLADSTONE: And what did The New York Times do?

STEVE DOIG: They quoted NBC, which is really the chicken way out. NBC is not a source. And usually there is a way to at least produce a number that can be sort of grounded, at least, in reality, and that’s really what the science of statistics is for.

BROOKE GLADSTONE: Science, schmience. Doig was roundly condemned by the right for lowballing Beck’s rally.

STEVE DOIG: Well, obviously I must be a liberal dupe who was paid to underestimate it and, of course, others on the other side of the ideological stream said, oh, yes, you know, obviously you’re doing very scientific methods and clearly that’s very good. That really amuses me because 18 months or 20 months ago I did the Obama estimate. I came up with a number of around 800,000 being there on the mall for the inauguration. That number was lower than the pre-event predictions, some of which were laughably high, like five million. So my prediction then became embraced by those who didn't really want to see Obama draw a big crowd and was sort of ignored by supporters. So, exactly the same methods predicted exactly the opposite reaction from the different ends of the political stream.

BROOKE GLADSTONE: Carl Holmberg, who retired from the U.S. Park Police in 2000 after 35 years on the job, assisted The Washington Post’s Dan Keating in his calculation of attendance at Obama’s inauguration. Using aerial photography, they came up with a range of 1 to 1.2 million. But after an earlier Post story tossed out 1.8 million, that’s the number that stuck. Carl Holmberg.

CARL HOLMBERG: And I'm not exactly sure where that 1.8 number came from. Somebody came out with a figure. There was no protocols applied. There was no procedure explained. And I would just hope that our history would be written a little more accurately than just based on an unnamed source.

DAN KEATING: I like data, I like facts rather than just numbers that people throw around based on their opinion.

BROOKE GLADSTONE: Washington Post database editor Dan Keating, who learned his methods from Stephen Doig when they were at The Miami Herald, saw his own science-based estimate of 1 million overshadowed in his own paper by that more memorable figure of 1.8. So how did The Post report attendance at Beck’s rally last week?

DAN KEATING: I believe we've made reference to Glenn Beck’s number. I believe we've made reference to Michele Bachmann’s million number, and I believe we may have made reference to the number that CBS derived.

BROOKE GLADSTONE: But if they just gave each of those three numbers equal weight, that’s a problem. Wouldn't the normal reader see one as the low estimate, one as the high estimate and maybe wind up with Glenn Beck’s number as the right one? And isn't that, given that there was no science behind Glenn Beck’s number, foisting a misrepresentation upon the public in order to appear fair?

DAN KEATING: Well, I think it’s hard to completely ignore what people say. We put it in the best possible context we can. We'll go to a lot of effort to put some science behind a number, and then it kind of goes up into the media atmosphere and gets sucked up in with every other number that people just whip off the top of their head. And I wish there was an easy way of saying, you know, hey, my number’s better than all your numbers. But the number that we worked really hard to put a lot of facts behind sometimes loses its weight as compared to all the other numbers bandied about.

BROOKE GLADSTONE: So the solution would be an official number. And, in fact, the Park Police still uses aerial photographs and density calculations to estimate crowds for planning purposes. But it can't release those numbers, not since the mother of all crowd estimate disputes erupted into the threat of a lawsuit in 1995.


LOUIS FARRAKHAN: A million sober, disciplined, committed, dedicated, inspired black men. Long live the spirit of the Million Man March!

CROWD: Long live the spirit of the Million Man March!


CARL HOLMBERG: Well, the name was attached to it six months prior to the event.

BROOKE GLADSTONE: Retired Park Service estimator Carl Holmberg:

CARL HOLMBERG: And when you call it the Million Man March, I guess you expect to have a million people there, but our count didn't go up to a million people.

BROOKE GLADSTONE: What was your count?

CARL HOLMBERG: I believe it was like 250,000.

BROOKE GLADSTONE: So when you come up with a quarter of the advertised amount, what happens?

CARL HOLMBERG: People get upset. [LAUGHS] The sponsors do, anyway. And the Million Man March certainly is not the only one. We had others that wanted to take us to court. Certainly many antiwar rallies thought that we had shortchanged them. And so Congress in I believe it was our 1997 Appropriations Bill directed the Park Police to stop estimating crowds.

BROOKE GLADSTONE: Do you wish that Congress would once again reauthorize the Park Police to reissue those numbers?

CARL HOLMBERG: Well, if they want accurate numbers, or as accurate a number as they can get, then they should. If they're happy with just having everybody name their own number, then I guess that’s up to Congress.

BROOKE GLADSTONE: So, there you have it folks, just another of many inconvenient truths.

[MUSIC UP AND UNDER] The difference here is that this one has been neatly legislated out of existence.