community-based, non-corporate, participatory media
how many in DC on J18? An Estimate
by Paul Donahue Sunday, Jan. 19, 2003 at 9:04 PM
This article summarizes an attempt to get a handle on actual turout in DC for the January 18 protests. We may want to consider employing some kind of count for the parade and march this weekend in Pittsburgh.
I have been dismayed by the numerous vague estimates on the numbers of demonstrators at the two Washington, DC demonstrations I have participated on both October 28 and on January 18. The situation in San Francisco is similar. Now, I fully agree that numbers shouldn't be so important; our message should. Unfortunately much of the perception of the importance of our movements by the politicians is in our numbers, rather than the merit of our ideas, so an honest estimate IS important. At the same time, wild guesstimates, usually optimistic, don't help either.
The intuitive estimates I have read or heard have varied by a factor of 10 or more. Predictably, the pro-war media estimates are obviously too low, and the demonstration organizer (ANSWER in the DC case) provide estimates that seem to be high (although to their credit not as high as the media's are low). Now, intuition is a great thing in many situations, but when dealing with large-scale phenomenon such as large numbers, sizes, weights, etc. intuition fails us, because we just don’t have a "feel" for many situations. For example, we don’t know what a couple tons weight "feels" like, because if we find ourselves under such weight, we don't live to remember it. Likewise, once we are in a crowd of more than 10,000 or so, it just seems to be a sea of humanity, be it 10,000 or a million.
So, we need to do some kind of analysis. The actual estimate of the size of a crowd or march should not, in theory, be a big deal. First, it should be pretty simple to just place volunteers at the ingress and egress points to the rally and just count them! Considering how important the numbers have become, ANSWER or other organizers should already be doing this! Unfortunately, they are not. So we are left to other estimating devices. Since the standard demo consists of a rally then a march, two methods come to mind.
1. For a stationary rally, or a "snapshot" sampling of a march of a known length and width, it is just:
Rally (or march) area * average of spacing between persons squared
The critical thing is that we have to take the average of the square of the spacing, not the square of the average spacing. Just taking the square of an estimated average spacing is going to result in an underestimate, which becomes quite big if the crowd is widely varying in density. Now, the average rally usually DOES have widely varying spacing with tightly packed people in the inner parts and sparse spacing on the outskirts. This spacing also varies widely from demonstration to demonstration. At the DC events, the spacing at the October rally (when the weather was warm) was somewhat wider than the January 18 rally, when there was definitely a lot of huddling together for warmth.
Nonetheless, we should get a fairly good estimate if we break the gathering into 2 or 3 zones of assumed equal density and add them together.
2. For a moving march, we can just take a sampling point, preferably near the beginning of the march, and take the average persons per unit time (seconds or minutes) times the duration of the march. This method has the advantage in that any errors in your estimate don't get squared, like they do in method #1 above. Of course, not every rally attendee marches, and some marchers drop out early. So this method can also result in an underestimate.
So with this in mind, here's my estimated from January 18.
1. Rally size -
I ended up missing the bus and driving to DC instead. I was separated from my group and basically on my own, so I had the opportunity wandered about the rally area (with my peace flag on a 15 foot pole) and got an idea of the size of the rally perimeter. I also paused at a couple points to observe the rate that marchers passed by. Back home, I used a mapping utility combined with old air photos and known landmarks to calculate the areas occupied by the rally. I broke the 1,450,000 area the rally occupied into 3 zones; a dense inner the area of the mall between Fhird and Fourth St.); a less dense middle zone area extending a bit beyond the second set of loudspeakers, and along the streets surrounding the mall, and an outer zone extending to the far end of the Smithsonian Air and Space museum). The innermost, packed part was estimated to be 400,000 square feet and huddled tight - I assumed 2.0 feet between persons (center to center) The second zone was a 250,000 square foot transitional zone at 2.5 feet between persons, the outer zone was800,000 square feet extended beyond 4thstreet and was assumed to be 6 feet average between persons. The resulting total is 162,000 people.
2. March size -
Very rough observations indicated between 10 to 50 persons per second crossing a point on M street SW near 8th street. I assumed 25 per second. Various reports in Indymedia reported the parade continued at this point for about 1 hr 50 min. This yielded 165,000 marchers.
Most errors are likely to be on the conservative side, so I believe about 170,000 people were in DC on January 18, 2003 to protest US aggression against Iraq.
Many of you probably feel this number is low. I had felt that about 300,000 showed up until I started crunching numbers, but I was not able to get close to 300,000 , much less ANSWER's "500,000" using credible data.
But, I may be off in some of the parameters used for the numbers, especially the crowd-density figure. If anyone, especially ANSWER organizers or volunteers have better figures I'd like to hear from them.
But the main conclusion is that nothing is better than actually performing a count. If we feel that turnout is important, we need to get some volunteers out there to actually do a count.
by R. Workman Monday, Jan. 20, 2003 at 2:57 AM
Good mathematical proposal for estimating the crowd. I think it might be easier in Pittsburgh.
I may try a count
by pjd Monday, Jan. 20, 2003 at 8:52 AM
I may just volunteer to sit at a strategic point on Sat and Sunday and perform a count.
BTW, to expand on the remark about 300,000 not being credible, I found the maximum credible mumber to be about 200,000, so this 200K should be considered the "no more than" number.
...and apologies about belaboring this issue - I'm a pointy-headed engineer with OCD.
There was a count
by JM Monday, Jan. 20, 2003 at 8:11 PM
There was a count in progress along the march route in DC on Saturday. There were a few people (one woman sitting in a lawn chair concentrating on the crowd) on the left side of the road after a sign that read, "count in process--stay in road." Of course, people were overflowing and going behind the sign (and the counters) without having seen them.
I think I've read somewhere that the technique is something like this: visualizing clumps or rough rows of people as they pass (you develop an eye for it) by tens, twelves, whatever is easiest given the width of the street.
I was under the impression that it's these people who provide the estimated figures announced at the end of marches like this.
Anyway, there were a lot of us! And quite a diverse group!
I wish you a strong turnout on the 25th and 26th!
by Sam Thursday, Jan. 23, 2003 at 4:52 PM
Well, there definitely weren't 500,000 or (as some were claiming) 800,000.
I'm angered by ANSWER's deliberate exaggeration; it reflects very poorly on the activist community and damages our credibility.
by pjd Friday, Jan. 24, 2003 at 10:52 AM
Fiddling with the two different calcs., I realize that most likely, there were likely many more at the march than at the rally, due to late arrivals and feeders.
Final likely count - 170,000 at rally, about 210,000 at march.