Two studies of visual search for word targets with controlled rates of presentation*The two studies reported involve the visual search of word lists for a target item when the rate of presentation is controlled and the words are presented tachistoscopically. In the first study, the target is differentiated physically from the filler items by being capitalized. When the target is the last item in a list, it is readily identified at all presentation rates, but when it is the first word or is embedded in a list, recognition accuracy is inversely related to presentation rate. In the second study, the differentiation between target and filler items is in terms of the presence or absence of category membership. All Ss at all presentation rates do significantly better on lists with an animal word as a target and a set of unrelated words as filler items than on the converse arrangement.The two exploratory studies reported here stem directly from the work of Neisserand Beller(1965) and from Neisser, Novich, and Lazar (1963) on searching through letter and word lists for a target item. They found that relatively well-trained Ss could scan a column of items, each consisting of a row of six letters, at a rate of approximately 10 items/second when looking for a specified letter target. The error rate, i.e., the failure to detect the target letter, averaged about 23%. The scanning rate was reduced to 6 items/second and the error rate averaged 24% when the column of items consisted of words three to six letters in length and the target was a word from a designated category, such as animals.The interest of Neisserand his associates in these tasks was concerned primarily with the information-processing procedures involved. However, the performances observed raise additional interesting questions: (l) What is the maximum rate at which the various types of visual material can be processed, and (2) what are the role of various factors, such as masking, in limiting the rate of information handling? The procedure employed by Neisser is not appropriate for studying these questions, because of the self-pacednature of his task and the possible limit imposed on search rate by the time required for successive saccades and fixations in scanning a vertical list. A recently developed n-channel stroboscopic tachistoscope overcomes these limitations. It permits the use of reasonably long lists of items, controls the presentation rate over a relatively wide range of values, and eliminates the need for successive saccades.
This study was suggested by a number of incidental observations in the experimental literature. These observations seemingly indicate that a difficult discrimination is more easily established if the subjects are first trained on an easy discrimination of the same type than if all the training is given directly on the difficult discrimination (3,4,7). The greatest efficiency seems to result, moreover, if the transition from the easy to the difficult discrimination is gradual rather than abrupt. If it is assumed that these results still hold when controlled for the total number of practice trials or reinforcements involved, they conflict with the usual expectations as regards the transfer of training. Normally it is assumed that for a fixed amount of practice on a training discrimination, the transfer effect to a test discrimination will be greatest when the training and test discriminations are identical. The above observations are in conflict with this assumption in that transfer is greatest between nonidentical discriminations. Consequently, if they are correct, they place limitations on the various theories of discrimination learning, especially those formulated in terms of generalization gradients. The present experiment is designed to check these observations under the condition of controlled number of training trials.
METHOD SubjectsThe subjects in this experiment were 46 male, albino rats approximately 60 days old at the beginning of training. Six of these were discarded during the adaptation trials because of timidity or excessive weight loss due to the restricted diet.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.