The roles of typicality, instance dominance, and category dominance in verifying category membership.

Four experiments examined the relationships between category verification reaction time (RT), exemplar typicality, instance dominance, category selection time (an RT measure of category dominance), and three production measures of category dominance. The category dominance measures were obtained in two experiments for use in two nearly identical category verification experiments. In one category verification experiment, the category name was presented 800 ms before the exemplar name; in the other, the exemplar name preceded the category name by 800 ms. Multiple regression analyses of the yes and no category verification RTs indicated that category dominance produced large and highly significant effects in all conditions, whereas instance dominance produced marginally significant and selective effects. Typicality had no effect beyond its shared effect with the dominance measures except as a possible suppressor variable. Category selection time was the best predictor, although all of the category dominance measures were better predictors than typicality or instance dominance. It is concluded that typicality may be an inappropriate independent variable for RT tasks studying the memory representation of natural-world knowledge.