500,000? 750,000? 1 Million?
Why We Crave A Real Number
By Philip Kennicott
Sunday, May 2, 2004; Page B01
When Xerxes set out on his ill-fated attempt to conquer Greece in the 5th century B.C., he counted his army by gathering 10,000 men tightly into a circle, drawing a line around them, and building a corral around the space. Then he marched the rest of his men into the corral, group by group, until he arrived at the figure of 1.7 million, according to Herodotus (not always the most reliable historian). Counting crowds, in ancient Persia at least, was easy.
Not so in Washington, where the legions of protesters who descend on the National Mall each spring not only paralyze city traffic, but have defeated the most basic efforts at human accounting. The Mall is a stage for political dissent, and the political force of that dissent, for lack of better metrics, is measured by the size of the crowd. This number, it seems, is absolutely crucial to placing protests in the larger context of American political life. And yet what Xerxes did so simply eludes us today.
Take last Sunday's March for Women's Lives. Marches are a dime a dozen this time of year, but this one was different. It was clearly huge, but how huge? No one could quite say, which is as frustrating as sweating through the hottest day of the year without the satisfaction of knowing exactly what the temperature was. Nonetheless it invited comparison to the biggest marches of the past, and enticed pundits to consider the gathering not just as another item on the spring calendar of political spectacle, but as yet one more little tea leaf at the bottom of the election-year teacup. Washingtonians process protests mainly for their nuisance value, in terms of street closings, Metro crowds and other disruptions. But this one, even though no one could put a precise number on it, had transcended nuisance and hit a magic number of some sort. It was an event.
Without the essential fact at their disposal, the media went into contortions to give an accounting. In its front page story on Monday, this newspaper described those in attendance as "hundreds of thousands," then gave the organizers' estimate of 1.15 million and an informal police estimate that called the crowd "at least the biggest since the 1995 Million Man March." (Independent researchers put that one at about 870,000.) The Post also cited Police Chief Charles H. Ramsey, who said the march met or exceeded organizers' expectations of 750,000. The New York Times led with "hundreds of thousands," and then cast doubt on the organizers' claims of more than a million by adding that the higher number was offered as "a clear demonstration of political clout." Grasping for a hard number, the Times also mentioned a Metro ridership figure of 320,138 -- up from 133,448 the Sunday before -- and then acknowledged that of course not all marchers took the Metro.
If you simply averaged all the numbers definite enough to plug into a calculator (let's say 1.15 million, 870,000, 750,000 and 186,690 -- the difference between the Metro ridership numbers), you would end up with a lovely, meaningless number (in this case 739,1721/2 people). Useless, yes, but it's worth noting that this kind of mental averaging may well reflect how many people read, and make sense of, conflicting numbers when the media can't come up with a conclusive one.
A more nuanced process would put each number into context and consider the emotions, habits and political goals of those who offered it. That analysis would look something like this: The 750,000 figure reflects the organizers' hope, and the 1.15 million reflects their exultation. The Metro ridership figure of 320,138 unmasks the journalists' pseudo-scientific preference for a "hard" number over a perhaps more accurate, but subjectively derived, one. Although the police no longer officially provide estimates (controversy over the U.S. Park Police estimates of crowd size at the Million Man March helped put an end to that), they still offer "informal" estimates, each of which has to be seen within the context of "crowd control." Yes, the cops are connoisseurs of "crowds," but they may also be exercising a good deal of "control" in their estimates, reflecting, perhaps, the tension, the overtime and a certain sniffy impatience that people in authority generally show toward those who pay their salaries. There's no greater form of crowd control than control over the number that gets in the history books.
All of this makes these numbers seem more like vectors (def: a mathematical expression reflecting "magnitude and direction"). There is a number, and there is a secondary factor, a direction (political, emotional, epistemological) in which that number is heading. Protesters head toward an impressive number (1 million has emerged as a sacred goal of big protests); cops head to a deflating number (their final act of control).
And journalists? They head to a safe number. The most curious thing about not having a hard number is how thoroughly it messes with the journalist's basic hard-wiring. Without a number in the first paragraph to start the article, the story sputters. And there's a habit of underestimation built into the basic semantics of journalistic caution: Go for a low figure and then say something like "more than" or "at least." If you think there were 900 people in a room, "more than 500" is definitely true; "almost a thousand," which is closer, may also be riskier (in part because protesters who complain about low numbers are already considered complainers -- they're protesting, after all -- while people who complain about inflated numbers are seen as more objective). The Washington Times, for instance, led its coverage of Sunday's march with "thousands." Of course, you can't dispute that. (Nor can anyone dispute the fact that "hundreds" or "dozens" gathered.) But at some point, journalists become guilty of an inverse hyperbole -- and suspected of bias.
What about a real number?
Although scientists are pursuing ever more sophisticated systems of crowd estimation, using multiple sensors and complex algorithms, one of the most widely available, relatively accurate methods is not much different from Xerxes'. Using aerial photographs and grids, and estimates of varying crowd densities within the squares on the grid, researchers can come up with more reliable numbers.
A Boston University professor, who looked into the Million Man March controversy to find a number that might satisfy all the various constituents, used grids and estimates to arrive at a number around 870,000 (used by The Washington Post when it cited "independent researchers"). It was a far cry from the Park Police's estimate of 400,000, a number so low that the Nation of Islam threatened a lawsuit.
The BU number had a margin of error of about 20 percent, which would be intolerable if this were a public opinion survey. But of course protests aren't polls (though we have a habit of forgetting this because anything that can be counted must be a poll, no?). They don't measure the breadth and density of political sentiments among the population. Rather, they measure something more akin to an interest group's ability to get out the vote. It tests logistics and organization as well as commitment.
Let's indulge a scientific-political fantasy, for a moment. If researchers came up with a counting method much more accurate than a 20-percent error margin, and if protest organizers were willing to subscribe to an "accounting" service that would provide a third-party reliable estimate (rather like newspapers and magazines submit to independent circulation audits), would we be any closer to a "satisfying" number? It's likely that good hard numbers wouldn't even satisfy marchers who scored a numbers success like the one last week. Good numbers would only corrupt the essential purpose of the rally (forcing organizers to direct their efforts ever more at getting bodies on the ground, rather than crafting a message). It would do to marches what fundraising has done to elections.
Despite our fascination with crowd sizes and numbers, numbers don't get at much of what makes marches on the Mall distinctive. The Mall isn't just a space, it's a canvas on which people gather like pixels on a computer screen. There's a natural sense of "quorum" (perhaps anything above 100,000) that animates the Mall to transform it from grassy avenues into something seething with color and motion.
When filled with people, the Mall's long, straight expanse is reminiscent of those large cardboard thermometers, colored red as the money flows in, that churches and charities use during fundraising drives. Actual numbers matter less than degrees of completion toward a full Mall. And a photograph showing a densely filled Mall may be worth more than any hard number.
Fill the Mall with enough people to march out and encircle the White House, and you meet another goal: the symbolic possession of the executive branch. Never mind that the president is probably away at Camp David, or clearing brush in Texas; a march that embraces his home sends a message (to the media and to the organizers and their opponents).
How that message is received or processed, protesters can't know. The wise journalist will forgo crowd estimation, and the wisest journalist will forgo attempting to make political prognostications based on protest sizes.
But despite the journalist's reticence (and this president's habit of outwardly breezy dismissal of all protest against his policies), marches have value beyond the numbers. They are informal, outdoor conventions, places for networking and sharing information. And they have a high "fun" factor, as daylong carnivals that tap into something primitive in the democratic spirit.
The Mall is also sized in such a way that invites us to gather in numbers that have historic resonance in this country. That half a million or 1 million people can fill this space and surround the most sensitive and symbolic government buildings in the country without sparking a fear of insurrection, is a remarkable fact in itself. But these magic numbers, whether they're blocks of 100,000 or 1 million, also recall the first scale of America. The original census put this country's population at just shy of 4 million. It isn't an accident, perhaps, that even when a quarter of that first American population figure gathers on the Mall, we feel a small thrill of collective memory, even these many years later.
Author's e-mail:kennicottp@washpost.com
Philip Kennicott is The Post's culture critic.
© 2004 The Washington Post Company |
