The Numbers Game in Tunisia: Nawaat Estimates as a Case Study

The numbers game was an important part of the 3 July coup in Egypt, implicating the army, political and youth groups involved in the 30 June demonstration, local media, some international media as well as certain western politicians and commentators. A similar game can also be perceived in Tunisia. As protests and counter-protests continue in Tunis, wildly differing numbers continue to circulate with rival protests using inflated and downplayed numbers ranging from 10,000 to 500,000 for the very same gathering.

The Tunisian multilingual blog Nawaat published estimates of the 3 August 2013 pro-government gathering in the Kasbah Square and the 6 August 2013 anti-government gathering in Bardo Square.

For the Kasbah gathering, Nawaat’s conclusions are the following:

1. “The surface area of the Kasbah square AND the adjacent 20 Mars Street is 23,650 m²“, which according to the site is a generous estimate, since it does not exclude the stage, sound equipment, martyrs’ monument, trees, benches, etc. Nawaat refers to Google Planimeter for their surface area conclusions.
2. “Crowd behaviour experts adopt the average of 1.5-3 people per square metre” – According to Nawaat reporters, protesters’ density was high near the stage, but “average to low as we get further away”, justifying the adoption of an average of 2 people per square metre.
3. Nawaat concludes that the number of protesters on 3 August was between 35,475 and 70,950 protesters, and that “the most correct estimate is thus 50,000 based on the average of 2 people per square metre”.

For the Bardo 6 August gathering, however, Nawaat’s techniques are strangely dissimilar:

• The surface area of the Bardo square is taken as 11,000 m² (no mention here of fountains, trees, stage, etc)
• The diagram used for the number estimate includes additional areas including adjacent streets which together add up to even more than the central square: 17,475 m².
• Curiously, when calculating the average density of protesters in Bardo, Nawaat declines to use the same estimate based on “crowd behaviour experts” used in the case of the Kasbah square. Instead of using the estimate of 1.5-3 people per square metre, they curiously stretch the density to 4.5 protesters per square metre throughout the whole of Bardo square. The density then reduces to 3 people per square metre outside the square, then to 2/m² further away then finally to 1/m² more than a kilometre away.
• The Nawaat final estimate for the number of protesters in Bardo on 6 August is 95,000 protesters.

Questions

A number of questions pose themselves:

• What is the basis for a very conservative estimate of the Kasbah Square area, and on the other hand a very generous estimate of Bardo Square?
• Why is a density of 2 protesters per square metre adopted in the case of Kasbah (initially 1.5-3, but the average of 2 was concluded by Nawaat to be “the most correct”), and an unprecedented 4.5 protesters per square metre adopted inside the Bardo Square? Are the Bardo protesters assumed to be half the size of the Kasbah protesters?
• Moreover, why are Kasbah protesters restricted to the square (while hundreds of images exist of protesters filling the surrounding areas all around the square); and on the other hand why are there significant areas (greater than the Bardo square area and containing more than the protesters estimated within the square itself) included in the Bardo protest area estimate? In fact, no aerial image supports the suggestion that half the protesters were outside the square or that protesters did spread out to adjacent streets over such a long distance. There is no proof that protesters covered a significantly extended area beyond the square, and there is certainly no proof that significant areas beyond the square contained the high density of 3 protesters per square metre adopted by Nawaat for an area as large as 7,325 m² (as well as an area of 6,350 m² containing an alleged density of 2m²).

Alternative Calculations

To address these intriguing questions, we use the same tool mentioned by Nawaat – Google Planimeter – to calculate the areas of the two squares. Google Planimeter gives us the results below:

• Bardo Square = 3,000 m²
• Kasbah Square = 25,600 m²

Using the very same tool, the area estimate we obtain for Bardo Square is very different from Nawaat’s, reaching around a quarter of the area estimate reported by Nawaat for Bardo square. We use two other programmes to calculate the Bardo Square area, the results are below:

Google Planimeter

Daft Logic Area Calculator

Free Map Tools

The same tools give the following area estimates for the Kasbah squares:

To avoid polemics about whether any side streets were full and how large those outside were and how full they were, and to avoid polemics about the density of protesters in either gathering, we make the following assumptions for both gatherings:

1. That the bulk of protesters on both occasions were within the squares.
2. That the same density estimate of 4 protesters per square metre can be used for both squares.

These produce the following numbers:
Kasbah: 25,000 x 4 = 100,000 protesters
Bardo: 3,000 x 4 = 12,000

The above are very conservative estimates. There is a great discrepancy in the Bardo estimate. If we double the area estimate to give the benefit to the claim that there was a large overflow into surrounding areas, and use a generous area estimate of 6,000 m² (which agrees with a quick naked eye non-expert estimation that the Bardo square is at the very most less than 4 times the area of the Kasbah square), this still gives us a MAXIMUM number estimate for the Bardo gathering of 24,000 protesters.

Our questions to Nawaat, leaving aside its inclusion of such large overflow areas around Bardo, remain the following:

1. How do you explain that your estimate of the area of the Bardo Square is half of the Kasbah Square, which is unsupported by even a cursory look at the two squares?
2. How do you explain your choice of a reduced protester density of 2 p/m² in the case of Kasbah and an unprecedented high density of 4.5 p/m² in the case of Bardo?
3. How do you explain your conclusions of a number of protesters at Kasbah that is half the number of protesters in Bardo (50,000 vs. 95,000)?