A verbal discrimination (VD) study phase was followed by a single-item (i.e., one item at a time) test phase. Prior right (R) items from the VD list were found to have higher old-new hit rates on the test phase than prior wrong (W) items. However, prior W items had higher hit rates than prior R items in terms of recognition of their intrapair function. These results were interpreted in terms of a high criterion model whereby an old test item must prossess a large number of frequency units before it is identified as being R. In addition, false recognition effects were demonstrated for new test items that were homophones of prior R items, but not for new test items that were homophones of prior W items. These results were interpreted in terms of a feature analytic extension of frequency theory.Both Erlebacher, Hill, and Wallace (1967) and Radtke and Foxman (1969) compared postcriterial recognition for wrong (W) and right (R) items on a verbal discrimination (VD) task, and found recognition to be substantially higher for R than for Witems (e.g., 99%vs 87% in the Erlebacher et al study). Recognition in these studies simply required old-new discriminations for old items (R and W) and new items. A correlated, but unexplored, question concerns S's further ability to discriminate between old items on the basis of their intrapair function (i.e., R or W). The present study introduced a methodology for testing the accuracy of functional, as well as old-new, discriminations. The Ss were given a standard study phase on a VD list and then a test phase in which items were exposed individually. Some of the test items were new (i.e., were not in the prior VD list) while the other items were either old R items or old W items. The S then decided whether they had functioned previously as R or W in the VD list.Predictions from frequency theory (Underwood, 1972) are clear regarding the effect of prior intrapair function on old item recognition. Old-new discriminations are based on the presence of situational response frequencies to old items, but not to new. Since the average number of units accrued to R items is greater than that for W items, discriminability from zero situational frequency should be greater for R than for W items. The earlier item recognition studies agree with this prediction, and there is no reason to predict otherwise in the single trial situation.

Individual items were tested for old-new and right-wrong identifications following one, two, or four study trials on a multiple-item recognition learning task. The hit rate for old item identifications was greater for right items than for wrong items, and the hit rate for both kinds of items increased progressively with increasing study trials. However, the expected interaction between item type (right vs. wrong) and study trials was not found. The hit rate for identifying the prior correctness of right items increased progressively across trials, and the false-alarm rate for misidentifying prior wrong items as right decreased progressively across trials. The pattern found for functional identifications suggests that frequency cues may be supplemented by other kinds of cues that enhance identifications of items in terms of their prior study list functions.

We compared an alternating task modules (ATM) protocol—which alternated sessions on Space Fortress, Phoenix (video-game-like tasks), and algebra word problems—with a massed protocol, which blocked sessions on the tasks. The protocols were isochronal (equal in duration), each entailing one week of training and testing. ATM gave the men and women an advantage in learning and retention on Space Fortress and algebra word problems; it did not change their performance on Phoenix. Arousal, not confidence or motivation, was implicated in the ATM advantage. The practical implication of demonstrating and describing ATM benefits is that real-world training often requires minimal durations. This pragmatic constraint favors ATM over frequently investigated anisochronal schedules, in which spacing practice over long durations is more effective than massing practice in short durations. Evaluations of retention, arousal, motivation, and confidence began the process of testing theories that might be expanded to explain and enhance ATM benefits.

Sets of pairs for a multiple-item recognition learning task varied in their number of exposures (one, two, and four) during a single extended study trial. The test phase required frequency of exposure judgments of individual items (i.e., prior right and wrong items). In agreement with frequency analyses of multiple-item recognition learning, frequency judgments of right items were greater than judgments of wrong items at each exposure level. Judgments of both item types increased as the number of pretest exposures increased, and the disparity between item types also increased with increases in exposure. Individual differences in frequency judgments were greater for wrong items than for right items.Implicit in frequency analyses of multiple-item recognition, or verbal discrimination, learning (Ekstrand, Wallace, & Underwood, 1966; Kausler, Pavur, & Yadrick, 1975) are the assumptions that variation in pretest exposure produces variation in frequency values for both right and wrong items and that magnitude of these frequency values for right items is greater than the magnitude for wrong items at each level of pretest exposure. In addition, the disparity favoring right items is assumed to increase progressively as the number of pretest exposures increases. Nevertheless, evidence regarding the nature of the covariations between frequency of exposure and frequency values in multipleitem recognition learning is quite limited. Wallace, Murphy, and Sawyer (1973) obtained subjective frequency ratings for right and wrong items after three successive study trials on a verbal discrimination list, and both Rowe (1975) and Wallace, Sawyer, Murphy, and Robertson (1976) recorded overt rehearsals of items during study trials that alternated with pairwise test trials on a verbal discrimination list. However, none of these studies attempted to determine the covariation between item exposure and the frequency values of right vs. wrong items. When a study list is composed of individual items that vary in frequency of exposure, an effective means of determining the nature of the covariation between frequency of occurrence and frequency value is to require subjects to make absolute frequency judgments of estimates (e.g., Begg, 1974; Rose & Rowe, 1976). Such frequency judgments presumably reflect the underlying frequency values of items at each level of pretest exposure. A procedure employed in a recent study by Kausler, Dalezman, and Yadrick (in press) is readily adaptable for use in a frequency judgment task This study was supported in part by National Science Foundation Grant BMS75-05007 awarded to the nISt author. 487in which both right and wrong items vary in their frequency of occurrence. This adaptation was employed in the present experiment as a means of detennining the covariation between frequency of exposure and frequency values for right vs. wrong items.The procedure calls for an extended study trial in which pairs are varied in their number of pretest exposures (one, two, or four). That is, subjects receive a heteroge...

Sets of pairs for a multiple-item recognition (verbal discrimination) learning task varied in their number of presentations during a single extended study trial. The test phase required old-new and right-wrong (functional) identifications of individual items. Old-item identification increased with increasing exposure for both right and wrong items. However, an increase in functional identification with increasing exposures occurred for right items but not for wrong items. This pattern of results suggests that identifications of prior wrong items are mediated by frequency cues alone, whereas identifications of right items are supplemented by nonfrequency cues that are associated with right items through their elaborative rehearsal during the study phase.The ability of subjects to identify both the oldness and the prior function (i.e., whether they are right or wrong) of individual items following study trials on a multiple-item recognition, or verbal discrimination, task has been demonstrated in several recent studies (e.g., Kausler, Pavur, & Yadrick, 1975;Kausler & Yadrick, 1977). The means by which response-based frequency units may mediate such identifications have been described by Kausler et al. (1975) in terms of a signal detection model. Briefly, separate distributions of frequency units are presumed to be generated for right and wrong items during practice. Although these distributions are expected to be partially overlapping, the mean of the right-item distribution should exceed the mean of the wrong-item distribution, beginning with the first study trial (see Figure 1). The disparity between distributions is attribut-