Friday, May 27, 2005

Schapelle - a statistical analysis

Now I have heard this statement a few times with respect to Indonesia's legal system. "Only facts relevant to the case should be considered" implying that things such as the relative percentages of people carrying drugs from Indonesia to Australia and visa versa, the prevalence of drugs in luggage in domestic Australian flights, unrelated cases of drugs mysteriously appearing in peoples luggage are completely irrelevant. This reminds me of a tragic case of a young male getting a screening for HIV. His initial screening came back positive, and his doctor told him that the test had a 4% false positive rate. Because this young individual only took into account facts relevant to the test, he calculated that he had a 96% chance of having AIDS. He committed suicide, not realising that the actual probability was only 1% that he had HIV. Anyone who knows statistics knows that in situations of imperfect information, a strictly statistical approach including (even highly approximate) data about the population in general will give a much better idea of probability of guilt. By the percentage of drug cases that the judge has found innocent, it seems that he is just rubberstamping the prosecution's cases. The Indonesians don't seem to agree with the adage that it is better for 99 guilty to be set free than for 1 innocent to be found guilty.

No comments: