Abstract

There has been continuing debate within the psychological community as to the nature of the relationship between mood and memory (Beck & Clark, 1988; Bower, 1981; Williams, Watts, MacLeod, & Mathews, 1997). The current research investigated the relationship between mood and memory and tested hypotheses related to the structure of memory for threatening and non-threatening information. The current set of experiments involved groups of students who had less than clinical level of symptoms of anxiety and/or depression, those with clinically significant symptoms of anxiety and/or depression, and those individuals who had been induced into a negative mood state. Participants completed inclusion and exclusion stem completion tasks in a process dissociation procedure (Bellezza, 2003; Bodner, Masson, & Caldwell, 2000; Jacoby, Yonelinas, Jennings, 1997). Two types of models with differing assumptions related to the dependence and independence of automatic and conscious processing were fit to the data. The models included high and non-high threshold versions of generate source and direct retrieval types (Belleza, 2003; Bodner et al., 2000). Examination of the fits produced by multinomial models indicate that data were best fit by non-high threshold direct retrieval models for individuals in the mood induced, clinical, and control groups. The high and non-high threshold models for a generate source type did not provide good fits to the data. The results of Experiment 1 did not indicate a shift in mood as a result of participation in the experimental procedure. The results of Experiment 2 did indicate that there were differences in the processing of threatening and non-threatening information between control participants and their high negative affect counterparts. The results of Experiment 3 indicated (similar to Experiment 2) that individuals who experienced a negative mood induction procedure also processed threatening and non-threatening information differently than the control group. The results of Experiment 3 indicate support for the model of a non-high threshold direct retrieval model as being appropriate in the evaluation of data for this group. The results indicate that individuals who are either in a negative mood or experience a negative mood induction procedure generated non-threatening types of words to complete new word stems as opposed to the control group which generated threatening words. The significance of the results is discussed. The results of the current experiments provide empirical support for the writings of Williams et al. (1997) as well as Beck (1986) in that clinical populations do appear to differ in the manner of their processing of threatening information from non-clinical samples.