Validation of the Van Overschelde et al. (2004) category norms: Results from five experiments

Van Overschelde, Rawson, and Dunlosky (2004) collected category/exemplar norms from over 600 participants at three American universities. For each of 70 categories, participants generated as many exemplars as possible in 30 seconds. One measure computed from their data was the "TOTAL" statistic - the proportion of participants that generated a particular exemplar given a category. Five experiments in the present investigation tested the ability of the TOTAL statistic to predict Reaction Time (RT) in a category/exemplar verification task with 236 participants. The simple correlation between "TOTAL" and the natural log of RT was approximately -0.20 with the average cost (slope) from high TOTAL (near 1.0) to low TOTAL (0.02) on the order of 250 ms. The results are similar in magnitude to previous research using older Battig and Montague norms which the Van Overschelde et al., norms supersede suggesting that the updated norms are a suitable contemporary replacement.