/ 26 September 2003

How much use are stats?

National crime statistics, except for murder, have limited value in tracking trends. Station-level statistics are better indicators, Ted Leggett, senior researcher at the Institute for Security Studies (ISS), told the Mail & Guardian. But they are treated like ”the great national secret”.

”I don’t know how they justify that in a democracy,” said Leggett.

Such information is crucial to many decisions, from buying a house to establishing a business. Residents would also be able to lobby for better policing resources, while taking their own protective measures.

But police have long said local statistics are only for operational planning. This week police National Commissioner Jackie Selebi again ruled out access to such statistics. Upon request a station commissioner could only identify ”hotspots”, but not crime trends, Selebi said at the release of the 2002/03 crime report.

As a result, his claim that the anti-crime campaign Operation Tswikila has reduced the murder rate in Khayelitsha township, Cape Town, from 15 a weekend to two remains untestable.

According to the report, there has been a 1,3% drop in the national murder rate — from 48 to 47,4 per 100 000 citizens. But this changes dramatically at provincial level: the Western Cape murder rate stands at 84,8 per 100 000, in contrast to 12,1 in Limpopo.

The Democratic Alliance accused Minister of Safety and Security Charles Nqakula of ”misleading the public” by presenting the crime statistics as ratios. Instead it used the total countrywide number of murders (21 553) to show an increase from last year’s total of 21 405 killings.

Leggett said, however, that ratios were a better evaluative tool because they took population figures into account.