Verbal Discrimination Task Research Articles

Orthographic Distinctiveness of Homonyms and the Feature-Tagging Hypothesis

Errors in verbal-discrimination learning varied inversely with degree of orthographic distinctiveness between intrapair wrong and right items (homonyms scaled for orthographic distinctiveness) in both mixed and unmixed lists. The inverse relationship supports the hypothesis that the difficulty of intrapair discriminations increases as the degree of feature sharing between intrapair items increases. In addition, the operation of a selective-strategy process with the mixed lists served to magnify the effect of distinctiveness. In single-item recognition learning, a number of words are exposed individually during a study phase. During a subsequent test phase, these 'old' words are again presented individually, together with a number of 'new' words that had not been included in the prior study phase. The subject must decide for each test word whether it is old or new. A false recognition occurs when a new word is misidentified as being old. By contrast, an individual item presented for study in the verbal-discrimination task consists of two (or more) words, one of which has been arbitrarily designated by the experimenter as 'right' (R) and the other(s) as 'wrong' (W). Thus, rather than having to make a simple old/new discrimination during test phases, the subject must discriminate between two old words. A false recognition occurs when a W word is misidentified as being an R word. For single words, several current models of recognition learning (e.g., Kintsch, 1970) presume that the stored representation of each word receives a recency tag during the study phase. Identification of a given word as old or new during the test phase is then contingent upon the presence or absence of a recency tag stored with the subject's representation of that word. However, recency tags are likely to be inadequate as a basis for recognizing R words in a verbal-discrimination task, in that both words of each W-R pair may carry such tags. A plausible alternative is that the subject stores a frequency-ofresponding tag with the stored representation of every W and R word in

Internal vs external locus of control and the use of self-reinforcement.

All Ss received training on a verbal discrimination task with E reinforcing correct responses. After training to a 60% criterion E ceased administering reinforcement and Ss were instructed to reinforce themselves after correct responses. Ss were divided into internal and external locus of control groups; during the self-reinforcement phase, half of each group received the same stimuli as during training and half received a generalization set. There were no differences between internals and externals in total self-reinforcement used or in its correctness. Internals made fewer correct responses and gave more incorrect self-reinforcement despite a lack of differences in initial learning. Results and implications were discussed in terms of the identification and establishment of reinforcement contingencies by the two locus of control types.

Imagery and meaningfulness of right and wrong items in verbal discrimination learning

The effects of variations in the rated imagery (I) and the meaningfulness (m) of the right word (R) and the wrong word (W) on a verbal discrimination (VD) task were examined by factorial manipulations of the I and m values of the R and W words within VD pairs. Results indicated that I was a more potent variable than m in the VD task.

Association by contiguity: Clustering in free recall

Unrelated words were randomly paired then presented in a verbal discrimination (VD) task. After VD learning, the 24 words were used in a typical multitrial free recall (FR) situation. Analysis of clustering in FR revealed that contiguous presentation of words during VD learning induced Ss to recall the words together. Association by contiguity appears to be one determinant of clustering (organization) of output order in FR.

Developmental trends in the intentional and incidental learning components of a verbal discrimination task

Two experiments investigated developmental trends in both the intentional and the incidental learning components of a verbal discrimination task. An inverse relationship between age and intentional verbal discrimination learning was found from the fourth to sixth grade levels in Experiment I and from the third to sixth grade levels in Experiment II, in agreement with Weber's law presumed by frequency theory to be operative in verbal discrimination learning. On the other hand, speed of intentional learning increased from the kindergarten to third grade level (Experiment II) and from the sixth to eighth (Experiment I) and ninth (Experiment II) grade levels. In both experiments there was a direct relationship between age and amount of incidental learning, and there was no evidence for the post-sixth-grade decrement in Type 2 incidental learning that has been reported in earlier studies.

The effect of a coding cue on verbal discrimination learning

This study investigated the effect of the oral presentation of an English word as a coding cue for the correct CVC in a verbal discrimination task. The Ss were 22 undergraduates, 11 to a group. It was found that the group given the English words similar to the corresponding nonsense syllables made significantly fewer errors on the test trial.

Performance in a verbal discrimination task with items differing in reinforcement probability.

Performance in a verbal discrimination task with items differing in reinforcement probability.

Response probabilities in a verbal discrimination task with probabilistic reinforcement

Forty Ss received 160 presentations of each of four word-pairs in a variation of the noncontingent binary-choice probability learning paradigm. Successive presentations of any given word-pair were separated by the interpolated presentations of other word-pairs. For each pair, one of the words was designated as correct on a randomly-chosen 70% of the presentations while the other word was correct on the remaining 30% of the presentations. Inspection of asymptotic response probabilities, which approximated the reinforcement probabilities, and first-order conditional probabilities indicated that this method of stimulus presentation may provide a less ambiguous test of the assumptions of certain mathematical models than is afforded by the standard probability learning paradigm.

Verbal discrimination reversal as a function of overlearning and percentage of items reversed: An extension

The Ss received either 0 or 50% postcriterial trials on a verbal discrimination task. Following original training 25, 50, 75, or 100% of the items received a reversed reinforcement contingency. Main effects of degree of training and of percentage of items reversed were significant. The speed with which 100% reversal was obtained suggested that Ss utilized a “concept of reversal” in this task.

Item recall in verbal-discrimination learning as related to pronunciation and degree of practice

Incidental components, wrong (W) item and right (R) item free recall, of a verbal-discrimination (VD) learning task were related to type of item (W vs. R), the presence-absence of overt pronounciation during informative feedback, and degree of practice on the VD task. R recall exceeded W recall, but only in the absence of overt verbalization. Recall of both W and R items increased with degree of practice. Pronounciation was also accompanied by a decrease in proficiency of learning the VD task.

Associative recall in verbal-discrimination learning as related to pronunciation and degree of practice

Modified free recall (MFR) of associations between intrapair wrong (W) and right (R) items learned incidentally during practice on an intentional verbal-discrimination (VD) learning task was related to direction of recall (W-R versus R-W), the presence-absence of overt pronunciation of W and R items during the informative feedback exposures of VD practice, and degree of practice (five versus 10 trials) on the VD task. MFR for both W-R and R-W associations increased with overt verbalization and with greater practice, but W-R and R-W recall did not differ in amount. Associative matching of intrapair W and R items also increased with overt verbalization and degree of practice. Pronunciation was found to be accompanied by a decrease in proficiency of learning the intentional VD task.

Differential accrual of frequency in verbal-discrimination learning

The notion that Ss build up a different number of frequency units for correct and incorrect items of a verbal-discrimination task was tested. Following verbal-discrimination learning, Ss were given a recognition retention test. It was found that they correctly identified more of the correct than incorrect members of pairs. This is in support of the tested notion.

The effects of verbal reinforcement combination on the performance of a four-alternative discrimination task

Groups of medical patients from a Veterans Administration Hospital uninformed about the reinforcement procedures were given a 4-alternative verbal discrimination task under a Right- blank or Wrong- blank reinforcement condition. Although responses to a post-experimental interview suggested that a majority of the Right- blank S s misinterpreted the information value of blank , treating it as though it sometimes or always meant “Right”, their performance was significantly superior ( p blank group, in opposition to the results of prior studies employing problem-solving or 2-alternative discrimination tasks. In a replication of the experiment with college students, a number of Right- blank S s exhibited a similar but weaker tendency to treat blank as “Right,” However, no significant differences in performance related to reinforcement condition were found. Hypotheses concerning differences in S s' approach to discrimination tasks with two and more than two alternatives as related to S s' intellectual ability were discussed.

Knowledge of correct and incorrect items in verbal discrimination

This study shows that Ss learn more about correct than incorrect members during acquisition of a 2-choice verbal-discrimination (V-D) task. The degree of this superiority of knowledge of correct items does not seem to be related to the type of verbal KOR Ss receive (RW, Rb, or Wb), at least when KOR itself does not produce differential rates of V-D list acquisition.

Variables Affecting Incidence of Self Reinforcement

Self reinforcement (sr), the administration of reinforcing stimuli by S to himself without external constraints, was investigated with regard to the effect of task and type of reinforcing stimulus. Each S received the same five tasks during each of two sessions. The tasks all involved a response followed by an sr (3 groups: with sr light, sr chips, sr rating scale). Two of the tasks were relatively unstructured; a second pair were less ambiguous and included a model's response to guide S; and the fifth was a verbal discrimination task learned to a 50% criterion. The results indicated: no consistent differences in rate of sr for the three types of reinforcing stimuli; increases in sr in Session 2 for tasks which showed a practice effect; rate of sr related to task with more sr when S agreed with the model's responses or the model was correct; and significant inter-session correlations on rate of sr for all tasks.

The precriterion phase in verbal discrimination learning

Summary The purpose of the present study was to determine if increments in correct response occur in verbal discrimination learning during the precriterion trials. Previous studies of this issue have involved comparisons of different Ss and different items on successive trials, as fast learners and easy items reached criterion and were dropped from the analysis. In the present study, the precriterion data for each item was Vincentized into three blocks of trials; thus each item was equally weighted in all blocks of trials. Fifty-nine Ss learned a 15-item verbal discrimination task in which each item consisted of five CVCs, and S was to learn which CVC was correct for a given item. The Ss were divided into two groups which differed only in which CVC in an item was considered correct. After each item, S was told the correct response. Ss were given 20 trials. Precriterion trials were those preceding 80% correct performance on an item.Results indicated a progressive increment in correct responses during the precriterion trials. These results suggest that learning is incremental, rather than all-or-none, and is thus inconsistent with Bower's (1962) position. An alternative possibility, that Ss learn to eliminate incorrect responses, thus increasing the probability of guessing the correct response, is also inconsistent with Bower's (1962) position.

Performance on a four-alternative task under different reinforcement combinations

In opposition to what has been found with tasks with 2-response alternatives, performance on a 4-alternative verbal discrimination task was shown to be significantly better (p <.05) under a Right-blank verbal reinforcement combination than under a Wrong-blank combination. Probability analyses suggested that these results could be attributed to b1ank in the Wrong-blank condition having less reinforcement value than “Right” in the Right-blank condition.