Back to Publications
Combining DNA Evidence for Greater Match Information
Perlin, M.W. Combining DNA evidence for greater match information. Forensic Science International: Genetics Supplement Series, DOI 10.1016/j.fsigss.2011.09.112, 2011.
Statistical science studies multiple experiments in order to assess their implications and variation. Forensic science is often single experiment, and so cannot employ statistical inference. With DNA, however, there can be multiple items and assays from crime scene evidence. Since STR data have quantitative peak heights obtained from DNA sequencer signals, they can be statistically analyzed in a likelihood function. Moreover, multiple STR experiments can be mathematically combined using a joint likelihood function. Statistics tells us that using more data should yield more information. We find that computing with a joint likelihood function to combine DNA evidence can infer more identification information, as measured by the likelihood ratio match statistic.