The veracity of long-term memories - Did Bartlett get it right?

Long-term memory can be defined as a store of information gained from past experiences. However, the information stored may not be accurate, or the processes of retrieval may lead to errors in recall. One theory put forward to explain inaccuracies that occur during recall is schema-based reconstruction of information during encoding. To investigate the part played by reconstruction in recall the present study uses the "method of repeated reproduction" described by Bartlett in 1932, substituting a real-life incidental event for the supernatural folk tale used as the stimulus material by Bartlett. Bartlett found accuracy in the initial recall of the tale to be rare, with further loss of accuracy observed with repeated recall. In the study reported here, the use of "real-life" stimulus event produced accuracy of recall that was maintained, regardless of the number of repeated recalls or time between recall sessions. The few changes observed were in the form of descriptions containing pre-event schemata of similar locations and, with repeated recall, post-event schemata developed through subsequent experience. The initial accuracy sustained throughout the time period, together with the relative lack of change over time, suggests very limited use of reconstructive processes. (C) 1998 John Wiley & Sons, Ltd.