No description available.

This dissertation is based on the Kida-Smith (1995) model of "The encoding and retrievability of numerical data." It is concerned with the variable conditions under which a positive affective response (i.e., a decision or opinion that results in a positive valence) on previously viewed accounting information may and may not influence current decision-making. An affective response to accounting numbers may adversely influence decisions made based on those numbers. Prior research has found that individuals recall information that is consistent with prior decisions more readily than they recall inconsistent information. Research has also shown that current judgements are biased toward prior decisions or judgements. These biases may cause current decisions to be suboptimal or dysfunctional. Two 2x2 experiments were conducted to examine four hypotheses. These hypotheses concerned (1) the influence of an affective response on an investment decision when the differences between two sets of accounting numbers are small and when the differences are large, (2) the influence of an affective response on the recall of numerical data, (3) the influence of time on the recall of numerical data given an affective response, and (4) the influence of an affective response on an investment decision when the level of cognitive processing at the time the affective response is produced is low and when the level of processing is high. The first experiment used graduate students in an accounting course to investigate the influence of differences between numerical amounts on decision making. It also investigated the influence of time between the encoding and retrieval on recall of numerical amounts. The second experiment used accounting practitioners to investigate the influence of differences between numerical amounts on decision making, and to examine the influence of different levels of cognitive processing at the time of encoding on decision making. Results indicate that an affective response does produce suboptimal decisions. In the case of accounting practitioners, however, the influence of the affective response is mitigated when the magnitude of the difference between the accounting numbers previously viewed and those undergoing current examination is large rather than small. The affective response did not significantly influence the recall of numerical amounts. There was no significant change in the influence of the affective response on recalled amounts with increased time between encoding and retrieval. Also, there were no significant changes in decision-making with increased processing at the time of encoding. / Ph. D.