We don't really.
Organizers of Egypt's protests called for a million demonstrators to come to a rally at Cairo's Tahrir Square on Feb. 1, but actual estimates of crowd size have varied widely. Al Jazeera reported a crowd of "more than a million." The AP was more conservative, using "more than a quarter-million." The New York Times was vague, describing "hundreds of thousands" filing into Tahrir Square. How do they come up with these numbers?
They're probably guessing. The most widely used method of counting crowds was developed by the U.S. National Park Service in the early 1970s and relies on three factors: the size of the space where the event occurs, the amount of the space occupied by the crowd, and the density of the crowd. Crowd density can vary from "mosh pit" close -- one person for every 2 feet -- to the more common one person for every 5 feet -- your standard Tea Party rally. It's nearly impossible to get an accurate count from the ground because of an optical illusion called foreshortened perspective that makes crowds closer to the observer appear denser than they actually are, so aerial photos are generally needed.
U.S. newspapers still use a variant of this method, though the Park Service has been prohibited by Congress from releasing its own numbers since 1995, when Million Man March organizer Louis Farrakhan threatened to sue the agency for estimating the crowd size at around 400,000.
For preplanned, organized events like presidential inaugurations or the recent rallies on the Washington Mall held by Glenn Beck and Jon Stewart, news agencies contract with aerial analysis firms like the Virginia-based Digital Design and Imaging Service to get the most accurate numbers possible. Digital Design makes a computer model of the space divided into a grid. Then a balloon mounted with nine cameras is tethered over the crowd and takes photos at different altitudes and vantage points. Organizers generally try to schedule these surveys for the high point of the event, such as when Sarah Palin spoke during Beck's "Restoring Honor" rally. Digital Design then comes up with multiple numbers based on mathematical modeling, hand-counting heads within a given quadrant, and guesstimates from academics who study crowd behavior. The numbers are then averaged out.
Estimating gets exponentially harder in a more spontaneous and chaotic event like the Tahrir Square rally. The event was called only a day before, so obviously computer models and weather balloons were out of the question. The most accurate aerial photos of the event come from satellites, which means that images can be taken only when the satellite is directly overhead and when the square is not obscured by clouds. This makes it extremely difficult to get multiple images and to judge when the crowd is at its height.
Using a simplified version of the Park Service methods, analysts at the private intelligence firm Stratfor found that even the AP's 250,000 number is pushing it. University of Illinois sociology professor Clark McPhail estimated that Tahrir Square itself could probably hold a fairly dense crowd of around 100,000.
Of course, Tahrir is essentially a giant traffic circle with multiple feeder streets, also choked with people -- and news reports Feb. 1 stated that protesters were fanning out throughout the city. Because the demonstrators were not necessarily marching in one particular direction and were pushing through vehicles, police, and other impediments, a reliable estimate for the total number of people in downtown Cairo becomes mostly a matter of guesswork.
Even using the vaguest criteria, however, Al Jazeera's estimate seems extremely generous, though that was presumably of little comfort to President Hosni Mubarak as he announced his impending departure from power.
Thanks to Curt Westergard, president of Digital Design and Imaging Service, and Clark McPhail, professor emeritus of sociology at the University of Illinois.
Chris Hondros/Getty Images