Loftus & Palmer: Eyewitness Memory

All 7 Terms & Definitions

Reconstructive memory

The view that memories are not stored as complete, faithful replicas of events, but as fragmentary traces that are actively reconstructed at retrieval using available information — including the original encoding, schema-based expectations, context at retrieval, and post-event information. First articulated systematically by Bartlett (1932), reconstructive memory theory predicts that errors, distortions, and false details should be common features of recall — not anomalies requiring special explanation. Loftus & Palmer's findings are canonical evidence for this view.

The misinformation effect

The finding that post-event information — presented in questions, narratives, overheard conversations, or media coverage — can become incorporated into an eyewitness's memory of the original event, producing confident but false memories of details that were never present. The verb manipulation in Loftus & Palmer (1974) is the foundational demonstration. Subsequent work showed the effect with more dramatic false details (a non-existent stop sign, a man with a moustache, broken glass) and with entirely fabricated events implanted through repeated suggestion.

Leading questions

Questions that presuppose information, suggest an expected answer, or use language that implies a particular version of events. "How fast were the cars going when they smashed into each other?" presupposes high speed and serious impact. "Did you see the broken glass?" presupposes broken glass existed. Leading questions are common in police interviews, courtroom cross-examination, and therapy sessions, and Loftus's research showed they can significantly alter what witnesses subsequently report remembering — with serious consequences for the justice system.

False memory implantation

A more extreme extension of the misinformation effect: the implantation of entirely false autobiographical memories through suggestion, repeated questioning, or the testimony of trusted others. Loftus & Pickrell's "Lost in the Mall" study (1995) demonstrated that a significant proportion of participants came to "remember" a fabricated childhood episode of being lost in a shopping mall after a trusted family member falsely described the event. This research became centrally relevant to debates about "recovered memories" of childhood abuse in therapy and their admissibility in court.

Overwriting vs. coexistence accounts

Two competing explanations for how post-event misinformation affects memory. The overwriting account (Loftus's original position) holds that misleading post-event information replaces the original memory trace, which is permanently lost. The coexistence account holds that both the original and the post-event traces persist, but the misinformation trace is more accessible at retrieval — original memories can sometimes be recovered with appropriate retrieval cues. Current evidence favours coexistence, with source monitoring errors (incorrectly attributing the false detail to the original experience) as the primary mechanism.

Forensic implications and interview reform

Loftus's research directly influenced reforms in forensic interviewing and eyewitness identification procedures. The Cognitive Interview (Fisher & Geiselman, 1992) uses open-ended questioning, mental context reinstatement, and encouragement to report everything to maximise accurate recall without suggestion. The PEACE framework (UK police) similarly emphasises non-leading questions. Sequential line-up procedures (replacing simultaneous presentation) reduce misidentification. Expert testimony on memory's fallibility has become common in cases where eyewitness evidence is contested.

Ecological validity and limitations

Key limitations of Loftus & Palmer's experiments include: (1) Participants were students watching filmed accidents in a laboratory — the stress, arousal, and emotional salience of real accidents (where weapon focus and physiological activation affect encoding) differ substantially. (2) Filmed accidents are less memorable and personally significant than real ones, which may produce greater susceptibility to post-event suggestion. (3) The between-subjects design means individual baseline accuracy could not be tracked. (4) The short delays between event and questioning (immediate for Experiment 1, one week for Experiment 2) do not capture the much longer delays common in real legal cases.