Lies, damned lies and statistics
Apr 05, 2006
OXFAM has caused a storm of protest by claiming that deaths through violence in northern Uganda are three times higher than in Iraq.
OXFAM has caused a storm of protest by claiming that deaths through violence in northern Uganda are three times higher than in Iraq. The so-called Oxfam report lists 50 NGOs as joint authors although some now want to disassociate themselves from it.
The assertion was both ridiculous and offensive. UN statistics show that 168 people died from violence in northern Uganda between July and December, less than the number who die from violent deaths in one week in Iraq.
Essentially Oxfam was massaging the figures. They compared the mortality rate in Iraq in 2003 immediately after the American invasion with the average mortality rate for northern Uganda over the last 20 years. This ignored the fact that Iraq enjoyed a period of relative calm after the invasion and that the security situation in Uganda has substantially improved in the last year.
The Oxfam report sought to make some legitimate points (that government should negotiate with the LRA, etc.) but by overstating its case in this manner, it wrecked its credibility.
Similarly, the Besigye appeal team brought in a statistician from Makerere University who demonstrated that Museveni apparently only got 48 per cent of the votes instead of the 59 per cent declared by the Electoral Commission.
The statistician based his analysis on less than half of the tally sheets supplied by the Electoral Commission to the FDC and then applied a weighting factor based on assumed population growth.
How reliable was this analysis? The Supreme Court judgement will reveal whether the judges found these figures convincing.
The famous adage states that there are ‘lies, damned lies and statistics’ meaning that statistics can be manipulated to prove anything. Sometimes statistics are true and useful, but they can also be very wrong and misleading. Statistics - whether they are for death rates, opinion polls, or radio rankings - have to be treated with great caution.
Ends
The assertion was both ridiculous and offensive. UN statistics show that 168 people died from violence in northern Uganda between July and December, less than the number who die from violent deaths in one week in Iraq.
Essentially Oxfam was massaging the figures. They compared the mortality rate in Iraq in 2003 immediately after the American invasion with the average mortality rate for northern Uganda over the last 20 years. This ignored the fact that Iraq enjoyed a period of relative calm after the invasion and that the security situation in Uganda has substantially improved in the last year.
The Oxfam report sought to make some legitimate points (that government should negotiate with the LRA, etc.) but by overstating its case in this manner, it wrecked its credibility.
Similarly, the Besigye appeal team brought in a statistician from Makerere University who demonstrated that Museveni apparently only got 48 per cent of the votes instead of the 59 per cent declared by the Electoral Commission.
The statistician based his analysis on less than half of the tally sheets supplied by the Electoral Commission to the FDC and then applied a weighting factor based on assumed population growth.
How reliable was this analysis? The Supreme Court judgement will reveal whether the judges found these figures convincing.
The famous adage states that there are ‘lies, damned lies and statistics’ meaning that statistics can be manipulated to prove anything. Sometimes statistics are true and useful, but they can also be very wrong and misleading. Statistics - whether they are for death rates, opinion polls, or radio rankings - have to be treated with great caution.
Ends