Diversity training initiatives are an increasingly large part of many organizations' diversity management portfolio. Little is known, however, about the effectiveness of such initiatives. In this article, we demonstrate how increased adherence to the principles of established social psychological theory can guide and make more coherent the development of diversity initiatives. Likewise, outcomes of diversity training can inform and make more practical social psychological theory and research. In short, both diversity trainers and academics would benefit from greater dialogue, as well as grappling with the tensions that naturally arise when theory and practice collide.

A series of experiments relevant to the functional value of accessible attitudes is reported. Experiment 1 established that diastolic blood pressure was sensitive to the task demands involved in subjects' expression of a preference between pairs of abstract paintings. In Experiments 2 and 3, subjects who had developed and rehearsed attitudes toward the individual paintings displayed a smaller elevation in diastolic blood pressure while performing this pairwise preference task than control subjects, suggesting that attitudes can ease decision making. Experiment 4 provided converging evidence for this hypothesis in that faster decision latencies were observed during the pairwise preference task when subjects had previously developed accessible attitudes toward the individual paintings. Evidence was also obtained suggesting that accessible attitudes enhanced the quality of the decisions made during the pairwise preference task.

Two experiments investigated the organization in memory of expectancy-congruent and expectancy-incongruent information pertaining to multiple trait concepts in an impression-formation task. In Experiment 1, when multiple trait concepts were represented in the information describing the target person, both congruent and incongruent items reflecting the same trait concept were stored together and were directly associated in memory, and both types of items were recalled equally well. In Experiment 2, when only one trait concept was represented in the information, incongruent items were recalled with higher probability than congruent items, and the latter were not directly associated in memory. Results suggest that with increasing categorical complexity of stimulus information, processes are invoked that do not occur in simpler impression-formation contexts. Implications for theoretical models of person memory are discussed.

The effect of uncertainty orientation on the processing of expectancy-relevant information was investigated. Uncertainty orientation concerns the extent to which people seek out and integrate information about themselves, others, and their environment. Uncertainty-oriented persons use this information to attain clarity and understanding, whereas certainty-oriented persons use such information to maintain preexisting belief systems. Consistent with previous findings, overall subjects recalled a higher proportion of expectancy-incongruent than expectancy-congruent information. However, the strength of this effect was a function of subjects' uncertainty orientation. As predicted, the advantage in recall for incongruent information occurred for uncertainty-oriented, but not certainty-oriented, people. Establishing a point of interface between the information-processing and individual difference perspectives, these results indicate that personality variables can reflect natural variation in cognitive processes that mediate social phenomena, thereby modifying effects assumed to be characteristic of people in general.

Two experiments demonstrated the impact of passively activated categories on the use of prime-relevant versus prime-irrelevant dimensions in later evaluations and preference judgments. In both experiments, concepts relevant to one of two dimensions of judgment were initially activated by requiring subjects to rehearse words in the course of a memory test. In supposedly unrelated subsequent tasks, subjects were asked to evaluate (in the first experiment) and choose between (in the second experiment) targets described as having positive attributes on one dimension and negative attributes on the other dimension. As predicted, subjects' evaluations and preferences reflected the greater impact of the information relevant to the primed dimension. Recall for the information presented in Experiment 2 provided some evidence for differential attention to the prime-relevant information as the mediating mechanism for the effect.