Loftus & Palmer 1974 Reconstructive Memory Study Replication

Abstract

This study replicated Loftus and Palmer’s (1974) classic experiment on reconstructive memory and misinformation effect, examining whether the verb used in a question, “smashed” vs “hit”, affects speed estimates and false memory for broken glass. Sixty high school students from Hangzhou Basis International School, aged 16‑18, with 35 being males and 25 being females, watched a short YouTube video of a car accident and were randomly assigned to answer a question about the car’s speed in the accident containing either “smashed” or “hit.” Immediately afterward, they gave their estimation for the car’s speed in km/h. One week later, they were asked whether they had seen broken glass, which is not present in the actual video. Results showed that participants in the “smashed” condition gave statistically significantly higher speed estimates, with a mean of 87.93 km/h, than those in the “hit” condition, with a mean of 54.16 km/h, as t= 3.82 and p < 0.05 in the t-test. Additionally, participants who were asked with “smashed” were significantly more likely to falsely report seeing broken glass, 69.0% in “smashed” group, compared to those asked with “hit”, 32.3% in “hit” group, as  χ² (1) = 8.08 and p < 0.05 in the chi-square test. These findings support hypotheses and the fact that memory can be reconstructed. It demonstrates that subtle wording changes can distort eyewitness memory for both numerical estimates and details. The experiment applies to real-world settings, reminding people that neutral, non-leading questioning in legal settings, especially in eyewitness testimony is essential.