Memory conformity between co-witnesses : the effects of discussion on subsequent memory accuracy
Errors in eyewitness accounts can occur when a witness comes into contact with post-event information. A common way to encounter this is for witnesses to discuss their memories with one another. The current research addresses this issue, by investigating 'memory conformity' between individuals who witness and then discuss a crime-event or pictures. For this research, a novel procedure was developed and employed whereby dyad members each encode a slightly different version of the same basic stimuli, which is then discussed prior to a free-recall test that is completed individually. Experiment 1 found a significant memory conformity effect between witnesses who had discussed an event prior to recall, in comparison to a control group with no discussion. At test, 71% of witnesses errantly reported at least one unseen detail that had been acquired during the co-witness discussion. No age-related differences in susceptibility to memory conformity were found between younger (18- 30 years) and older (60-80 years) adults. Experiment 2 found that misleading post-event information acquired during a discussion with a co-witness was a more powerful means of influencing memory reports than misleading narratives that are commonly employed in eyewitness research. This was true for both younger (17-33 years) and older (58-80 years) adults. Following these findings, Experiments 3, 4 and 5 investigated possible factors underlying memory conformity, as well as incorporating a source-monitoring test. Individual differences in personality and memory ability were not reliably associated with susceptibility to memory conformity. However, a consistent finding when analysing the co-witness discussions was a relationship between 'response order' and memory conformity. Specifically, the first witness to mention a critical (experimentally manipulated) detail that they had seen was the most influential dyad member, and the most resistant to influence, even when their memory was disputed by a co-witness. Applied and theoretical implications of the main findings are discussed.