Accuracy of distinguishing truth from lie

Abstract

Judging another’s veracity is a complex and difficult, albeit second-nature, undertaking. Humans ascertaining the truthfulness of another’s communication without the benefit of any special training or instrumentation average only 54% overall accuracy. Experts often perform no better than laypeople when judging the brief, decontextualized, low-stakes video and audio clips that dominate the meta-analytic literature. However, judges attain higher accuracy when they are trained and experienced, engage in longer interactions, have access to context information or baseline behavior, and adopt strategic questioning strategies. Judgment accuracy is a function not only of characteristics of the judge but also the sender, sender–receiver relationship, signal, communication genre (e.g., interview), modality (e.g., face-to-face, audio), and context. Behavioral observation techniques used by trained human coders, software for automating behavioral observation, and instruments for sensing and tracking behavior are being developed at a rapid rate, enabling improved accuracy by both experts and lay people.

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 — 43 citations exist upstream.

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