Easy reporting of hard DNA: computer comfort in the courtroom

Perlin, M.W. Easy reporting of hard DNA: computer comfort in the courtroom. Forensic Magazine, 9(4):32-37, 2012.

Downloads

Article

Abstract

Forensic analysts currently testify about mixtures and other ambiguous DNA with justifiable trepidation. While their laboratory data are extremely reliable, the human interpretation of this data may lack rigor. CPI and other threshold-based methods modify data, ignore variables and discard information in order to enable approximate human solution. The alternative may be reporting nothing at all (e.g., "inconclusive"). But these simplifications are not based on a tested statistical model or on solid mathematics.

Much important DNA evidence, often crucial to a case or public safety, has been discarded by overly simplistic interpretation. Analysts often agonize over DNA mixtures, spending days wondering whether there is even a reportable match. Understating a statistic might free the guilty, while overstatement could wrongfully imprison an innocent man. Testifying can be stressful, with cross-examination questioning interpretation validity.

Interpreting DNA mixtures with a full statistical model, and hours of mathematical computing, can restore analyst confidence. Going into court with thorough and reliable match results, and an understanding of computer interpretation, establishes scientific comfort. The computer can separate out mixture data into component genotypes, and represent uncertainty as probability, neither understating nor overstating the match statistic. A computationally empowered analyst assists the court through objectively derived fact, not subjective opinion.