from the lies,-damn-lies,-and-the-bsa dept
Glyn Moody points us to an article looking at the report's coverage of South Africa, and notes not only did IDC/BSA not survey anyone in South Africa, they're using these totally made up numbers to push for new copyright laws. As for how ridiculous the numbers are, well, here's the quick explanation:
How was the 35 percent rate arrived at? It's a guess, or rather, a combination of guesses combined with some market data and presented as a final authoritative percentage.Yup. It's not statistics when that makes IDC/BSA look bad. But when you ask them how they made up their numbers, suddenly it's a statistical analysis.
South Africa wasn't surveyed at all for the current report.
The BSA says surveys were conducted in 28 countries representing "a mix of geographies, levels of IT sophistication, and geographic and cultural diversity". These included, among others, China, India, Pakistan, Thailand, Taiwan, Vietnam, Hungary, Poland, Russia, Germany and Italy.
IDC extrapolates numbers from the 28 countries to form conclusions about the 111 that appear in the final report. But even for those countries that are surveyed, the sample size is woefully small given the number of PCs there are in the world: just 6 000 consumers and 4 300 business respondents are the world's global proxy for the final results. This works out to 150 businesses per four countries.
What's the statistical error and standard deviation of this sample? The BSA in the UK told Brainstorm that the study "is not a statistical estimation or survey that lends itself to probability analysis, so there is no standard deviation."
How then is the number of applications per South African PC calculated? Using, er, statistical analysis.
But where it gets worse is that the BSA in South Africa is apparently using these totally made up "findings" to push for ever more draconian copyright law in South Africa:
Andrew Rens, an intellectual property lawyer with the Shuttleworth Foundation, is unimpressed with the proposed changes.The Brainstorm article includes some laughably generic "woe is the software industry" quotes from the local BSA representative. It also does a nice job debunking the ever-popular "ripple effects" that the BSA loves to tout, but only counts in one direction (ripple effects work both ways), and ignores the fact that it's often double, triple or quadruple counting the same dollars. It's nice to see more publications challenges the bogus BSA numbers. It really does make you wonder why so many press reports still quote them as if they were actually representative of something real.
"Over the last five months lawyers for the BSA, as well as the BSA chairperson, have been quite vocal about changes that the BSA wants made to South African law," he says. "Those changes would shift the onus of proof so that when the BSA brings a case against someone, that person would have the onus of proof on him. They claim that these changes are necessary to combat the allegedly high level of software infringement in South Africa. But now it emerges that the claims themselves are suspect...."