Eyewitness Memory Distortion Following Co-Witness Discussion: A Replication of Garry, French, Kinzett, and Mori (2008) in Ten Countries

• Published in: Journal of applied research in memory and cognition Vol. 8; no. 1; pp. 68 - 77
• Main Authors: Ito, Hiroshi, Barzykowski, Krystian, Grzesik, Magdalena, Gülgöz, Sami, Gürdere, Ceren, Janssen, Steve M. J., Khor, Jessie, Rowthorn, Harriet, Wade, Kimberley A., Luna, Karlos, Albuquerque, Pedro B., Kumar, Devvarta, Singh, Arman Deep, Cecconello, William Weber, Cadavid, Sara, Laird, Nicole C., Baldassari, Mario J., Lindsay, D. Stephen, Mori, Kazuo
• Format: Journal Article
• Language: English
• Published: Washigton Elsevier Science 01-03-2019; Society for Applied Research in Memory and Cognition
• Subjects: Collaboration; Conformity (Personality); Conversation; Female; Human; Male; Memory; Suggestibility; Witnesses; Post-event conversation; Multi-lab replication project; Co-witness suggestibility effect; Memory conformity; Eyewitness memory
• ISSN: 2211-3681; 2211-369X
• DOI: 10.1037/h0101833

We examined the replicability of the co-witness suggestibility effect originally reported by Garry et al. (2008) by testing participants from 10 countries (Brazil, Canada, Colombia, India, Japan, Malaysia, Poland, Portugal, Turkey, and the United Kingdom; total N = 486). Pairs of participants sat beside each other, viewing different versions of the same movie while believing that they viewed the same version. Later, participant pairs answered questions collaboratively, which guided them to discuss conflicting details. Finally, participants took a recognition test individually. Each of the 10 samples replicated the Garry et al. finding: Participants often reported on the final test a non-witnessed answer that their co-witness had stated during the collaboration phase. Such co-witness suggestibility errors were especially likely when the witness had not disputed the co-witness's report during the collaboration phase. The results demonstrate the replicability and generalizability of the co-witness suggestibility effect.
General Audience Summary
Often in criminal cases, multiple witnesses observe the crime. Co-witnesses might remember the details of the event differently, due to differing viewpoints, differences in arousal or attention, or mistakes due to the fallibility of memory. Co-witnesses often talk amongst themselves before being interviewed by police. This raises the possibility that witnesses' subsequent statements may be distorted by misinformation from other witnesses.
To explore this possibility, Kanematsu et al. (1996/2003) and Garry et al. (2008) exposed co-witnesses to subtly different versions of an event by using polarized video projectors. Co-witnesses sat next to each other, viewing the same screen but seeing (unbeknownst to them) slightly different versions of a movie (e.g., one witness saw a man wearing a blue baseball cap, whereas for the other witness the cap was black). After a short delay, co-witness pairs worked together to answer questions about the event, including questions about details that differed between the two movie versions. Finally, participants individually completed a memory test about what they had observed. People sometimes reported seeing details they had heard about from their co-witness.
The present study replicated the Garry et al. (2008) experiment using the same procedure and materials and similar instructions in 10 countries: Brazil, Canada, Colombia, India, Japan, Malaysia, Poland, Portugal, Turkey, and the United Kingdom. The effect was observed in every country, demonstrating that this co-witness suggestibility effect is robust and common to many cultures. Participants conformed to their partner mostly when they had not contradicted their partner's report during the discussion. Further research is needed to reveal the conditions under which discussion among co-witnesses is more versus less likely to lead to false reports and to explore for cultural differences in co-witness dynamics.