The examination of genetic patterns found in the DNA of biological specimens at crime scenes has become an important and widespread tool for forensic analysis. An important component of this technology is the assessment of how likely it is for the genetic patterns in two different samples to match when in fact they are not from the same person. The techniques for estimating these statistics currently employed by essentially all forensic laboratories in the United States is to assume statistical independence among multiple gene loci and to ignore the possibility that laboratory errors may give rise to erroneous matches. A recent report by the National Research Council (NRC) has made many recommendations which generally support the current practices in forensic laboratories and contradict recommendations of an earlier report by the NRC on this same topic. I present an assessment of the recommendations of the new NRC report. In particular I address what is known about false positives in forensic laboratories, the extent to which forensic laboratories have followed recommendations in actual case work and what I consider rational solutions to these problems.

Back to Workshop Schedule