Telling us less than what they know: Some ballistics examiners render inconclusive reports when they detect non-matches

Criminal investigations often hinge on the balance of testimony from forensic scientists. Many ofthese forensic scientists report on feature-comparison methods where they microscopicallycompare the features in an evidence sample (found at the crime scene) to the features in areference sample (associated with the suspect) to determine whether they originated from thesame source. One example of a feature-comparison method is cartridge-case comparison. Pastwork shows that examiners are much more likely to render inconclusive reports for actual non-matches than for actual matches. The result is that inconclusive reports have strong exculpatorypower. This is problematic because it is unlikely that defendants, attorneys, judges, or jurorswould interpret an inconclusive report as diagnostic of innocence. We fit a cognitive processmodel to cartridge-case comparison data to determine why examiners are so much more likely torender inconclusive reports for actual non-matches. We found that some forensic examinersrender inconclusive reports even when they correctly perceive that the reference sample does notmatch the evidence sample. By rendering inconclusive responses even when they perceive a non-match these experts might be concealing from innocent persons evidence that they desperatelyneed to establish their innocence. We argue that this biased examiner might result fromadversarial allegiance bias combined with a flawed response scale.

Link: https://www.researchgate.net/publication/373390356_Telling_Us_Less_Than_What_They_Know_Expert_Inconclusive_Reports_Conceal_Exculpatory_Evidence_in_Forensic_Cartridge-Case_Comparisons