08 May, 2011

Statistics and Judicial trials

An issue hotly debated in academic circles, but treated as little more than disdain by most judges, has been the extent to which mathematics can inform the trial process. An extreme level, the courts to accept mathematical evidence pointing to the design inference. Fingerprints and DNA evidence, for example, a powerful only because those who give it also tell the court, in mathematical language, that the odds of anyone else having the same principal DNA material are almost impetus similarly small. The question which seems to arises is whether it is rational not to act on such evidence.

Judicial skepticism of attempts to quantify the probabilities in a given case has often been justified. In some cases, the statistics are obviously been phoning. The prosecution in the case of Capt Dreyfus in 1899 claimed a remarkable incidence of similarities between Capt Dreyfus' handwriting and the letters to the Germans which the prosecution alleged and the defense denied were written by the same hand. In fact, the number of similarities were statistically insignificant. There are often good reasons to be dubious statistical evidence. One can rarely be assured of its correctness, let alone its applicability to the case in hand. And it is almost impossible to imagine a case in which all the odds can be precisely calculate it. How, for example, the character led the odds of a witness lied? But even excepting the limited place for mathematics in assisting quantification of the odds for or against proposition with any precision, the fact is that we do occasionally decide issues, or test the appropriateness of a tentative decision, by reference to vague notions of the chances for or against an innocent connection between the party and a proven fact.


चंद्रमौलेश्वर प्रसाद said...

Finally, shall we conclude that statistics is lies and damn lies as one of the definition of Statistics says :)

चंद्रमौलेश्वर प्रसाद said...

But the courts want definite proof and sometimes these proofs are also taken with a pinch of salt. That is the reason why most murderers go scotfree. Theory of probables means inference not definite proof !