Rethinking the probative value of evidence: Base rates, intuitive profiling, and the "postdiction" of behavior.

Abstract

It is argued that American courts may be routinely admitting evidence with little to no probative value and great potential for prejudicial impact. This may be particularly likely with regard to what is essentially "intuitive profiling" or "stereotype" related evidence, defined herein as evidence suggesting that the defendant (or other party), or his (her) behavior, fits intuitive "profiles" (or stereotypes) of the type of person likely to commit the crime or behavior in question. In other words, "intuitive profiling" evidence is admitted to "postdict" behavior Formal empirically based "profiling" evidence (testimony regarding the fit of a defendants characteristics or behaviors to formal or scientific profiles of the typical perpetrator of the crime in question) for use to prove guilt is inadmissible in American courts. However, we suggest that everyday use of informal intuitive profiles underlies both judicial determinations of probative value (diagnosticity), and thus admissibility, of evidence, and jurors' use of the evidence in determining guilt. Demonstrations of the use of base rate information to evaluate the probative value of such intuitive profiling evidence both as evidence of guilt and as evidence of innocence are provided. Demonstrations of both how to evaluate the actual probative value of evidence (when all necessary values are known), and the theoretical limit of its probative value (in circumstances where some values are not known) are provided. It is argued that such evaluations may provide the basis for (1) support of motions to either admit or to exclude evidence, (2) testimony to the jury to help them weigh or interpret evidence, (3) exculpatory profiling (profiling evidence of innocence), (4) pretrial research to establish probative versus prejudicial value of evidence, and (5) sufficiency analyses to determine maximum likelihood of guilt, given multiple items of evidence. Among these, the first two are considered most important, as it can be demonstrated that many "profiling" characteristics currently admitted in trial (such as evidence of battery to support a murder charge) are not probative of guilt.

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 — 44 citations exist upstream.

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