Contaminated record · genuine or implanted?
False Memory
A recollection of an event or detail that did not occur, or did not occur as remembered, yet is experienced with confidence and vividness comparable to genuine memories.
Exhibit · the account as remembered
Elizabeth Loftus and John Palmer established the reconstructive nature of eyewitness memory in 1974 by showing that post-event leading questions ('smashed' vs 'hit') altered speed estimates and produced false recall of broken glass. Henry Roediger and Kathleen McDermott extended this in 1995 with the DRM paradigm, demonstrating that 40-55% of participants falsely recalled and over 80% falsely recognized non-presented associated words. The effect has been replicated across dozens of languages and age groups (Gallo, 2010). Loftus and Pickrell (1995) later showed that 25% of adults could be led to believe they were once lost in a shopping mall as a child. ← reconstructed, not recorded
How the contamination enters
False memories arise through multiple converging processes. In the DRM paradigm, spreading activation in semantic networks (activation/monitoring theory) causes non-presented associates to feel familiar. Source monitoring failure then leads people to misattribute internally generated items to external experience (Johnson, Hashtroudi, & Lindsay, 1993). Fuzzy-trace theory adds that gist extraction produces a semantic representation that overlaps with the critical lure, driving false recognition even when verbatim traces would permit rejection. In eyewitness contexts, post-event misinformation can overwrite or coexist with original traces, a process called the misinformation effect.
Competing account — Activation/monitoring theory emphasizes automatic associative activation plus controlled monitoring failure, while fuzzy-trace theory emphasizes verbatim vs. gist trace competition. Both have received substantial empirical support and are not mutually exclusive.
Seams the examination found
Mean speed estimates ranged from 31.8 mph ('contacted') to 40.5 mph ('smashed'). In Experiment 2, 32% of those asked 'smashed' falsely recalled broken glass (none existed), compared to 14% in the 'hit' condition and 12% in the control.
Loftus & Palmer · 1974 · n = 45 (Experiment 1); n = 150 (Experiment 2)
False recall of critical lures occurred on 40% of lists (12-word) and 55% (15-word lists). False recognition approached the hit rate for studied words. Critically, 72% of false recognitions were accompanied by 'remember' judgments, indicating vivid phenomenal experience.
Roediger & McDermott · 1995 · n = 48 undergraduate students (Experiment 1); n = 24 (Experiment 2)
25% of participants (6 of 24) developed partial or full false memories of being lost in a mall, adding sensory details not in the original script. Five participants failed to identify the mall event as the false one during debriefing.
Loftus & Pickrell · 1995 · n = 24 (ages 18-53)
Where provenance breaks
Watch for situations where someone confidently recalls specific details of an event but those details were only suggested to them after the fact, or where repeated imagination of a scenario begins to feel like a real memory. In eyewitness contexts, note whether interview questions contain presuppositions or leading elements.
Authenticity — robust. The DRM false memory effect is one of the most replicable phenomena in cognitive psychology, having been reproduced across dozens of languages, age groups, and cultures (Gallo, 2010; Coane, 2021). Murphy et al. (2023) replicated the 'lost in the mall' study with a larger sample (N=123) and found a 35% false memory rate. The Loftus & Palmer misinformation effect has been replicated extensively, though a large-scale direct replication with modern methods is currently being planned by Zwaan and colleagues. Warnings can reduce but not eliminate DRM false memories.
Records later found altered
law · 1990
George Franklin murder trial (1990)
George Franklin was convicted of murdering 8-year-old Susan Nason in 1990 based almost entirely on his daughter Eileen's recovered memory of witnessing the crime 20 years earlier. Franklin's conviction was overturned in 1995 when it emerged Eileen had been hypnotized. Elizabeth Loftus testified as an expert witness for the defense, demonstrating how memory can be distorted by suggestion and external information.
law · ongoing since 1989
DNA exoneration cases involving eyewitness misidentification
The Innocence Project has documented over 375 DNA exonerations in the United States, approximately 69% of which involved eyewitness misidentification as a contributing factor. In many cases, the mistaken witnesses had expressed high confidence in their false memories, demonstrating how confident recollection does not guarantee accuracy.
To authenticate the memory
Use the cognitive interview technique: ask open-ended questions without presuppositions, encourage free recall before specific questioning, and have witnesses mentally reinstate the physical and personal context of the event. For DRM-type false memories, warning participants about the illusion can reduce but not eliminate it.
The cognitive interview, developed by Fisher and Geiselman (1992), increases the amount of correct information recalled without increasing false reports. Warnings about the DRM illusion reduce false recall by roughly half (Gallo et al., 2001).
Easily confused with
- Misinformation EffectparentThe misinformation effect (Loftus, 1975) is the specific mechanism by which post-event information alters memory; false memory is the broader umbrella term for any memory of an event that did not happen as remembered.
- Source Monitoring Errormechanistically-linkedSource monitoring theory (Johnson et al., 1993) provides the cognitive mechanism for many false memories: people fail to distinguish internally generated thoughts from externally experienced events.
- Imagination InflationsiblingImagining an event repeatedly can increase confidence that it actually occurred, a related phenomenon studied by Garry, Manning, Loftus, and Sherman (1996).