| Popis: |
Kindergarten children needed more trials to learn a list of similar than dissimilar words but made fewer over-generalization errors on subsequent transfer tasks. A third group of children, trained on the dissimilar list, was given overlearning trials to make up for the additional training trials needed by the similar group. While the overlearning and similar groups did not differ on over-generalization errors, both made fewer errors than were made by the dissimilar group. The authors conclude that over-generalization errors fade with over-learning and should not unduly alarm teachers. Over-generalization errors under whole-word reading instruction have been studied by McCutcheon and McDowell (1969), Otto and Pizillo (1970). and Samuels and Jeffrey (1966). These investigators found that children needed more trials to learn lists of similar than dissimilar words, but make fewer over-generalization errors on subsequent transfer tasks. These studies indicate that paired-associate training which forces children to attend to each letter in a word is less likely to lead to subse- quent over-generalization errors than training which permits children to discriminate words on the basis of fractional cues (stimulus selection). It is known, however, that adult learners engage in stimulus selection until a list is learned. Then during further overlearning trials, stimulus selection may be relaxed and additional parts of stimuli may be learned (James & Greeno, 1967). It is possible that giving overlearning trials on dissimilar lists to equal the additional trials required to learn similar lists would produce comparable over-generalization errors on transfer tasks. This hypothesis was tested with children by Gilbert, Spring, and Sassenrath (1977), with equivocal results. The investigators noted that their training procedure, patterned on previous studies, inadvertently resulted in their similar group receiving overlearning trials. This was because subjects received twelve preliminary fixed- order trials before scoring was commenced on random-order trials and some subjects in the similar group mastered the task before completing preliminary trials. In the present study, therefore, preliminary trials were reduced. |
