Science convicting the innocent

This paper focuses on the dangers of scientific evidence in criminal cases. The general reliability of scientific evidence is reviewed. Second, cases are evaluated thus tracing the development of the rules set out for the admissibility of scientific expert evidence. Third, cases are analyzed and the existing rules are criticized; this analysis grounds several proposed changes set forth as a protection for an accused person from possible dangers accompanied with scientific evidence. And finally, this paper proposes that the trial judge should exercise more active and comprehensive discretion to exclude potentially dangerous scientific evidence. Instead of assessing threshold reliability questions, the trial judge should determine the actual reliability of scientific evidence under the ad-hoc circumstances of the particular case before him. The trial judge should then determine if the scientific evidence meets the new tests, as proposed in this paper, for scientific admissibility, depending on whether the defense or the prosecution relies on science and how close science is to the ultimate issue of guilt or innocence. In judge-rule trails, judges must rule on the scientifically more accurate theory, but not on the more convincing expert.