False memory
≈ 3 MIN · 15 WORDS + 13 CARDS · Memory
What this is about
Memory does not store, it builds: fifteen neighbours of a word that never appears are enough for many to have “seen” it afterwards. Deese found this in 1959; Roediger and McDermott made it a paradigm in 1995 — the laboratory evidence that remembering is reconstruction.
In the original, participants falsely recognised the lure words at rates close to genuinely shown words — in the second study on about three quarters of the lists, often judging that they distinctly remembered them. That confidence is the point.
How it runs
- Fifteen words appear one after another, 1.5 seconds each (fixed associative order — shuffling would weaken the effect).
- Twenty seconds of checked backwards counting.
- Recognition over 13 cards: old (key 1) or new (key 2) — six old, six neutral new, one lure.
Runs entirely in your browser — nothing is transmitted or stored.
You can stop the run at any time.
Your finding
What it means
Limits of this experiment
- One list, one lure — the original averaged over many lists.
- Recognition without a sure/unsure judgment; the original collected that too.
- Knowing the paradigm lets you guess the lure — the finding says so.
→ Case file 009 · LOFTUS→ Topic: Measurement and Method→ Glossary: Reconstructive memory→ Glossary: Misinformation effect