In two experiments, we asked college students to judge the ease of learning word pairs known to differ reliably in actual ease of learning. The independent variable in both experiments was the rate of presentation (fast and slow) of to-be-judged items. The results of both experiments revealed only small differences in decision accuracy as a function of presentation rate. This suggests that metacognitive judgments of this kind rely on rather fundamental cognitive processes, which in our view are similar to those used to discover meaning in everyday verbal messages. Consider the following metacognitive task related to associative learning: A subject is presented a list of pairs of items (e.g., words), with the instruction to judge the ease oflearning an association between members of each pair. No learning is requested of the subject; rather, the subject simply predicts (e.g., via a rating scale) how easy it would be to learn an association between members of each pair iflearning were requested. These so-called ease of learning (EL) judgments reliably predict associative learning (see, e.g., Arbuckle & Cuddy, 1969; Kearney & Zechmeister, 1999), and thus they presumably provide information on the basis of which a learner can plan efficient study. For example, Nelson and Leonesio (1988) showed that more self-paced study time was allocated to items previously identified as more difficult on the basis of EL judgments. Kearney and Zechmeister (1989) speculated that the ability to make EL judgments would be related to associative learning ability. They found, however, only a slight relationship between performance on an associative learning task and performance on an EL task, although both tasks involved the same kind of items. Such a finding seems paradoxical, if it is assumed that effective associative learning is related to efficient use of study time based on judgments of item difficulty. A dissociation between performance on these tasks is also unexpected when viewed in terms of some interpretations of the associative learning process. Wang (1983), for example, suggests that good learners generate more effective elaborators than do poor learners. A dissociation between learning and EL measures would imply that the ability to produce effective elaborators is independent of the ability to recognize the items that can be most effectively elaborated .We wish to acknowledge support provided by a small grant from the Loyola University Research Services department. Thanks are also due John J. Shaughnessy for his critical reading of a draft of this manuscript and James W. Hall for many helpful discussions on this topic. Requests for reprints should be addressed to Eugene B. Zechmeister, Psychology Department, Loyola University of Chicago, 6525 N. Sheridan Rd., Chicago, IL 60626.Copyright 1991 Psychonomic Society, Inc. 36In the present study, we report the results of two experiments in which we asked college students to make EL judgments of word pairs known to differ reliably in learning difficulty. The independent variabl...