Word Frequency and the Recall-Recognition Paradox
AdvisorEakin, Deborah K.
CommitteeGiesen, J. Martin
When people predict recognition performance, they wrongly predict that high frequency words will produce better recognition than low frequency words. To examine whether familiarity was the heuristic behind these inaccurate predictions, participants saw some words prior to study to increase their familiarity. We found that familiarity influences predictions, but word frequency has the greater influence. Research has shown that these inaccurate predictions can be corrected with test experience. Subsequent predictions are more accurate, but it is unclear whether participants learn that low frequency words are always better for memory or that participants had learned that low frequency words are only better for recognition and high frequency words are better for recall. We resolved this issue by giving a forced-choice recognition test after the single-item recognition test to determine what participants learned after the first test, and we found that participants learned that low frequency words facilitate recognition but not recall.