Time in retrospect : the recall of temporal information in very long term memory.

Abstract

This thesis investigates the recall of temporal information in very long term memory, particularly event duration. Experiments 1 through 4 examine the estimation of public event duration, for example, how long the Falkland's war lasted. Subjects in Experiment 6, Part a, were asked to date events they had personally experienced (autobiographical events) and thus indicate the time elapsed since the event's occurrence. Subjects in this Experiment were also asked to recall other information about events, such as where they occurred: the event aspects 'what', 'where' and 'who' were used both as prompts to recall and as the aspects to be recalled. In Experiment 6, Part b, subjects estimated the duration of two types of autobiographical events: empty and filled duration events. Events typical of these two types of event are respectively, 'How long after you ordered the computer game did it arrive?' and 'How long were you in hospital for?'. Experiment 6 employed a method new to memory research, the undirected-diary method. Diarists were used as subjects in this experiment. Their diary records, which were recorded with no prior knowledge of any possible use in research, were obtained and used as a source of information about events they had experienced, for example, the nature and actual duration of events was obtained for Experiment 6, Part b. Because of the development of a new method, Experiment 5, a survey of diarists, was conducted. The results of this survey are reported in Chapter 3, in which the undirected-diary method is compared to other methods used to investigate autobiographical memory. It is suggested that the undirected-diary method is a useful method, and one which overcomes a number of problems associated with the study of autobiographical memory. Ornstein's (1969) 'storage size' model of duration estimation was examined in Experiment 2 and Experiment 6, Part b. In general, little support was found for this model and an alternative model was developed: the reconstructive model of duration estimation. Predictions made on the basis of this model were tested in Experiment 4, the results of which generally supported the model. Experiment 6, Part a, found results consistent with other studies of event dating and studies which have examined autobiographical memory. Absolute dating error increased and signed dating error varied systematically with retention interval. Recall of event aspect information varied systematically with the cues provided for recall and was interpreted within a uniqueness explanation of autobiographical memory organization.