Inadequacies of posterior probabilities for the assessment of scientific evidence

Abstract

In a recent publication, Koller et al. (Probability Conclusions in Expert Opinions on Handwriting. Substantiation and Standardization of Probability Statements in Expert Opinions. Munchen: Luchterhand, 2004) elaborate recommendations as to how forensic scientists should evaluate their findings in the specific field of comparative handwriting examination. The proposed procedure is intended to standardise the calculation of posterior probabilities of propositions which are of interest to members of the court, e.g. the proposition of a certain individual being the author of a questioned handwritten sample. Koller et al. (2004) also claim that such an approach may analogously be applied in other forensic disciplines. The present paper discusses some of the compromising effects that these recommendations may have on a coherent evaluation of scientific evidence. Drawbacks of posterior probabilities for describing the probative force of evidence are contrasted with insights offered by other evaluative frameworks. Inadequate implications of posterior probabilities are illustrated and discussed in the context of examples that focus on selected aspects of combining evidence. Graphical probability models, i.e. Bayesian networks, are used with the aim of clarifying the rationale underlying the arguments presented.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 36 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.