The use of DNA analysis in the criminal justice system has become a staple of investigations from homicides to property crimes. In many cases, the origin of the DNA is undisputed, and questions regarding how and/or when the DNA was deposited are key during court proceedings. Activity Level Evaluation (ALE), Activity Level Reporting (ALR), or Evaluation Given Activity Level Propositions (EGALP) are all terms that describe the process of assessing forensic DNA findings against competing propositions about disputed activities.

The framework for this analysis was initially described by Evett et al. in the 1980s and 1990s.  Analyses using versions of this approach have been performed since the 1990s and numerous guidelines now advocate evaluating forensic observations considering activity level propositions. However, some publications argue against activity level in court, citing it as an immature field, lacking data available to justify these evaluations, lacking validation, or simply being inadmissible in some jurisdictions.

While some of these concerns have been addressed, others persist, leaving DNA analysts in a difficult position when asked activity level questions in court. In this study, we address the validation of ALE, focusing on reproducibility.  Eight participants from six institutions, spanning several continents, were provided identical case background information, description of the forensic analyses performed, DNA analysis results, DNA profiles, and probabilistic genotyping analysis outputs. Evaluation propositions were not provided.  Participants were requested to return their reports and associated casefiles, including supplementary materials, appendices and the Bayesian Network, if applicable. 

The following aspects of the activity level analysis were examined: which aspects of the background information were used, or not, to inform the calculation; the length of time to complete each major step of the analysis; how many activity nodes were created in the calculation; how many and which studies were used for each node; how probabilities were derived from published studies; where the same study was cited by two or more labs, was the same probability derived; were there probabilities used that were not based on a published study; did the laboratory account for lab-to-lab variability and if so, how and what effect did it have; did the laboratory perform a sensitivity calculation and if so, how. 

This study demonstrates many similarities alongside intriguing differences in ALE practices across laboratories that merit further exploration. These findings may guide which aspects of ALE will be most impactful to address when seeking consistency in the field.

Authors

• Todd Bille (US Bureau of Alcohol, Tobacco, Firearms, and Explosives, United States of America)
• Duncan Taylor (Flinders University, Australia)
• Michael Coble (Center for Human Identification, University of North Texas Health, United States of America)
• Mariya Goray (Flinders University, Australia)