The impact of forensic delay: facilitating facial composite construction using an early-recall retrieval technique

Abstract

Memory for facial features deteriorates over time, diminishing one's ability to construct an accurate visual likeness of a face (i.e., a facial composite). In Experiment 1, we investigated how delay affects composite construction. Participants recalled an unfamiliar face during a Cognitive Interview (CI) and constructed a feature composite across four post-encoding retention intervals. Correct composite naming declined sharply after a 3–4 hour retention interval, remained stable at two days, and dropped to floor-level after one week. Experiments 2–4 examined how composite effectiveness was influenced by the incorporation of two factors: (a) a novel, self-administered written face-recall attempt, conducted 3-4 hours after encoding, and (b) a standard or modified holistic recall elicited immediately before construction. Participant-witnesses created more identifiable likenesses when early recall was invited, suggesting that this intervention consolidated and enhanced access to facial-feature information. The addition of a character-based interview further improved both feature and holistic composites.