Understanding the Statistics Behind Forensic Conclusions Workshop Series

Since the dawn of forensic DNA analyses in the 1980s, legal and scientist scholars have pushed for strengthening the logical and statistical foundations of forensic science. In the U.S., this impulsion has resulted in reports from the National Research Council of the National Academy of Sciences (in 2009) and from the President’s Council of Advisors on Science and Technology (in 2016). Both reports urge forensic scientists to shift from opinion-based to data-supported conclusions.

Currently, the legal community is faced with multiple challenges when considering the results of forensic examinations:

1. There is no consensus in the legal, forensic and scientific communities on what type(s) of conclusions should be reported by forensic scientists. The type of statements offered by forensic scientists when asked about the source of a forensic trace may vary widely within a given type of evidence, and between evidence types;
2. Not all sub-disciplines of forensic science have embraced the requested changes and are performing the necessary research to strengthen their scientific foundations at the same pace. Some evidence types have been heavily debated in court and in the literature (e.g., DNA, latent prints) and abundant literature describing their strengths and limitations is available; unfortunately, others have yet to demonstrate their reliability (e.g., firearm, shoe impressions);
3. Increasingly complex statistical techniques are proposed to support forensic conclusions. The debate is evolving from whether analytical techniques were appropriately used, to whether logically correct conclusions were statistically inferred from the data.

This series of 8 modules aims at providing lawyers with the necessary bases to understand, question and challenge the (lack of) logic and statistical support behind forensic conclusions. At first, participants will be introduced to some notions of statistics and probability theory. This necessary step will provide the audience with some knowledge of technical terms used by statisticians and forensic scientists. Following, participants will learn about the logic behind the different types of forensic conclusions, their limitation, and what is required to scientifically justify each one of them. Finally, participants will review past challenges of well understood forensic analytical processes and learn how these challenges can provide a framework to evaluate the reliability and scientific support of less well-discussed evidence types.