Methodological concerns: the feeling-of-knowing task affects resolution

In traditional feeling-of-knowing procedures, participants make judgments on unrecalled items only (e.g. Hart 1965). However, many researchers elicit feeling-of-knowing judgments (FOKs) on all items. When FOKs are made on all items, participants may use recall as a basis for judgments, leading to higher magnitude judgments for recalled items, but causing a relative floor effect for judgments for unrecalled items. We suspected that resolution (relative accuracy) would be better when FOKs are made on all items than when they are made on unrecalled items only. We examined the issue by comparing across studies, reanalyzing data from another experiment, and by conducting an original experiment. In the literature review, we included 83 conditions across 52 studies. We found that feeling-of-knowing judgments made on all items showed higher resolution than feeling-of-knowing judgments made on unrecalled items. This was replicated in the reanalysis of existing data of a single study that used both methods. In the original experiment, we collected feeling-of-knowing judgments for general-information questions. The experiment confirmed that resolution for predicting recognition was higher when feeling-of-knowing judgments were made on all items than when they were made only on unrecalled items. We discuss both methodological and theoretical implications of these data.