The advantages provided to memory by the distribution of multiple practice or study opportunities are among the most powerful effects in memory research. In this paper, we critically review the class of theories that presume contextual or encoding variability as the sole basis for the advantages of distributed practice, and recommend an alternative approach based on the idea that some study events remind learners of other study events. Encoding variability theory encounters serious challenges in two important phenomena that we review here: superadditivity and nonmonotonicity. The bottleneck in such theories lies in the assumption that mnemonic benefits arise from the increasing independence, rather than interdependence, of study opportunities. The reminding model accounts for many basic results in the literature on distributed practice, readily handles data that are problematic for encoding variability theories, including superadditivity and nonmonotonicity, and provides a unified theoretical framework for understanding the effects of repetition and the effects of associative relationships on memory.Recent years have seen a resurgence of interest in the advantages provided to memory by the distribution of multiple practice or study events. This interest owes in part to the ramifications such results have for developing effective training, educational, and athletic regimens, and to the impressive success researchers have had in importing interventions out of the laboratory and into training settings, the classroom, and the practice field. The effects of distributing practice are extremely robust and cross-cutting-the advantages are evident in basic memory tasks using words (Cepeda, Pashler, Vul, Wixted, & Rohrer, 2006;Janiszewski, Noel, & Sawyer, 2003) and pictures (Hintzman & Rogers, 1973), in motor skill acquisition (Lee & Genovese, 1988), and with more complex educationally relevant materials (Krug, Davis, & Glover, 1990). Almost every major textbook on cognitive psychology and on human memory discusses the advantages of distributed learning, and such effects are among the keystone results that contending theoretical models of learning and memory must dutifully explain.The results from experiments on distributed practice have informed memory theory principally insofar as those theories successfully handle the difference between massed and spaced practice conditions (the spacing effect; Ebbinghaus, 1885) or between differentially spaced conditions (the lag effect; Melton, 1967), as well as how they account for differences arising from intentionality of learning during study or from different memory tests (Challis, 1993;Greene, 1989;Russo, Parkin, Taylor, & Wilks, 1998). Here we make the case that the superadditivity effect (Begg & Green, 1988;Watkins & Kerkar, 1985;Glanzer, 1969;Waugh, 1963), which concerns the relationship between memory for single events and memory for multiple events, and the nonmonotonicity effect, which concerns the shape of the lag function, are empirical

Metacognitive monitoring and control must be accurate and efficient in order to allow self-guided learners to improve their performance. Yet few examples exist in which allowing learners to control learning produces higher levels of performance than restricting learners’ control. Here we investigate the consequences of allowing learners to self-pace study of a list of words on later recognition, and show that learners with control of study-time allocation significantly outperformed subjects with no control, even when the total study time was equated between groups (Experiments 1 and 2). The self-pacing group also outperformed a group for which study time was automatically allocated as a function of normative item difficulty (Experiment 2). The advantage of self-pacing was apparent only in subjects who utilized a discrepancy reduction strategy—that is, who allocated more study time to normatively difficult items. Self-pacing can improve memory performance, but only when appropriate allocation strategies are used.

If the mnemonic benefits of testing are to be widely realized in real-world learning circumstances, people must appreciate the value of testing and choose to utilize testing during self-guided learning. Yet metacognitive judgments do not appear to reflect the enhancement provided by testing Karpicke & Roediger (Science 319:966-968, 2008).

The ability to take a different perspective is central to a tremendous variety of higher level cognitive skills. To communicate effectively, we must adopt the perspective of another person both while speaking and listening. To ensure the successful retrieval of critical information in the future, we must adopt the perspective of our own future self and construct cues that will survive the passage of time. Here we explore the cognitive underpinnings of perspective-taking across a set of tasks that involve communication and memory, with an eye toward evaluating the proposal that perspective-taking is domain-general (e.g., Wardlow, 2013). We measured participants' perspective-taking ability in a language production task, a language comprehension task, and a memory task in which people generated their own cues for the future. Surprisingly, there was little variance common to the 3 tasks, a result that suggests that perspective-taking is not domain-general. Performance in the language production task was predicted by a measure of working memory, whereas performance in the cue-generation memory task was predicted by a combination of working memory and long-term memory measures. These results indicate that perspective-taking relies on differing cognitive capacities in different situations.

Predicting what others know is vital to countless social and educational interactions. For example, the ability of teachers to accurately estimate what knowledge students have has been identified as a crucial component of effective teaching. I propose the knowledge estimation as cue-utilization framework, in which judges use a variety of available and salient metacognitive cues to estimate what others know. In three experiments, I tested three hypotheses of this framework: namely, that participants do not automatically ground estimates of others' knowledge in their own knowledge, that judgment conditions shift how participants weight different cues, and that participants differentially weight cues based upon their diagnosticity. Predictions of others' knowledge were dynamically generated by judges who weighed a variety of available and salient cues. Just as the accuracy of metacognitive monitoring of one's own learning depends upon the conditions under which judgments of self are elicited, the bases and accuracy of metacognitive judgments for others depends upon the conditions under which they are elicited.