Still no fix for crowd-count conundrum

You would think that technology would have provided a way by now to get crowd counts and avoid the political fights over participation in the city's big annual rallies, like today's protest march.

But, unable to settle on a satisfactory method to tally the throngs passing through Hong Kong's streets, local police, academics and protest organisers continue to rely on a variety of relatively simple, low-tech methods. The results are likely to be at odds and subject to dispute.

'It would be a big day for Hong Kong if all parties can reach an agreement on how to count the number and if all these estimates are within reasonable range,' said University of Hong Kong social work professor Paul Yip Siu-fai, who does his own count each year.

In 2004, for instance, the newspaper Ming Pao took high-resolution aerial photographs to measure attendance at the July 1 rally. Grid patterns laid over the photographs broke the crowd area into smaller zones in which participants could be counted with better precision.

In the end, the newspaper estimated that 260,000 took part. That exceeded figures produced by three other methods, including the one used by police, but fell far short of the 530,000 claimed by march organiser the Civil Human Rights Front.

Robert Chung Ting-yiu, the director of HKU's public opinion programme, said Ming Pao's photographs were not clear enough, resulting in some marchers appearing lumped together.

It is also hard to pin down crowd sizes using such pictures because groups take up different amounts of space depending on the situation and environment, sometimes huddling together, sometimes spreading across the road.

Once, during the June 4 candlelight vigil three years ago, Chung's group borrowed an infrared camera to count heat sources. The results were disappointing, as the final estimates were far lower than those one obtained by their usual method.

Overhead photographs are good for counting static crowds, such as last month's June 4 vigil in Victoria Park, but researchers still prefer to be down on the ground to count a crowd on the move.

So, for today's rallies, some crowd counters will use a 'flow rate' to determine the numbers that pass a given point - a traffic light, for example - during the march.

Chung's group, for instance, will deploy 20 researchers to tally protestors as they pass under a footbridge in Wan Chai.

A police spokesman said the force based its estimates on how crowded the protest area appeared. The Civil Human Rights Front, meanwhile, positions 10 volunteers at four points along the protest route and averages the results.

'The rest of the groups do estimations, but what we do is real counting - not an estimation,' said Eric Lai Yan-ho, the group's convenor.

Academics say everyone would benefit if the parties involved could agree to one reliable method.

'Science and democracy are ways for society to advance,' Chung said. 'We may not get democracy, but we can do our best to do the science well.'