Unreasonable Doubt
http://www.wsj.com/articles/SB10001424052970204002304576628660692030584
Last the Guardian reported that an appeals judge in London threw out a murder conviction since the judge objected to an expert witness’s use of Bayes Theorem. The expert reached the conclusion that it was “somewhat unlikely that the observed correspondence would have been obtained as a result of a mere coincidence.” In the shoe print case, the expert testified that the degree of wear, size and sole pattern of the defendant’s shoes matched the prints at the scene. One of the main problems that judge had with the witness’s testimony was the execution behind it. During the cross-examination, the expert failed to clearly explain his reasoning that led to his ultimate conclusion. The judge also believed that statistics should not be permitted in the evaluation certain kinds of evidence. Statistics provide direction and not a precise answer, which is not very helpful in a courtroom when people are suppose to decide if a person is guilty without any reasonable doubt.
This article relates to our class discussion since we have spent lectures going over Bayes Theorem. Bayes Theorem looks at the probability of an event occurring given that a certain event or condition has already been met. Bayes Theorem looks from effect to cause which is why it would be useful in forensics. In the shoeprint case, the expert witness found statistics styles, sizes, and types of sneakers that are sold in the United Kingdom. He concluded that there was a very small probability that a different pair of sneakers could have produced the prints given the features of the shoe prints. In this example, given the effect, which is the features of the shoe print, leads to the small possibility that another shoe could have produced the print. The judge rules that even though the probability was a convincing figure, it was also misleading. The judge ruled that there weren’t actual hard numbers on sneaker distribution like there are hard numbers for DNA.