Computer selection of verbal research materials

Abstract

Most studies of verbal learning, memory, and perception require the selection of stimulus material the characteristics of which are chosen to study, or to control for, one or more of the attributes on whichverbal material may vary. For example, in a study of the effects of written word frequency on recall, the researcher would not only be required to select words on the basis of frequency, but also to control for their length in number of letters, degree of imageability, and perhaps even the age at which the words are typically acquired. A number of reports have been published that provide norm ratings for a wide variety of word attributes, (e.g., Gilhooly & Logie, 1980; Paivio, Yuille,& Madigan, 1968; Toglia & Battig, 1978). However, the process of selecting materials from such sets of norms can be extremely time consuming and is greatly eased if much of the selection is carried out by computer. Coltheart (1981) and others are developing just such a facility, using sets of norms from a variety of sources with a total data base of 98,538 words, and providing a potentially invaluable resource.However, not all attributes are given for each word, and since differingsets of norms are involved, the measures for each attribute are not wholly consistent. In addition, at present, the size of the data base and the wide variety of parameters available require the researcher to submit requests for verbal material by post, in a specialized, although relatively straightforward, access language. This entails a necessary delay between submission of jobs and receipt of the sets of stimuli. Furthermore, the initial run on the system may not be wholly successful, and the job may have to be resubmitted. One solution to this would be on-line, interactive computer selection with a program and data base that are sufficiently small to be implemented on the computing facilities readily available in most psychology departments. The program to be described aims to provide such a facility. Data Base. The program WORDSAMP uses as its data base the Gilhooly and Logie (1980) measures of age of acquisition, imagery, concreteness, familiarity, and ambiguity for 1,944 words. Thorndike and Lorge (1944) frequencies are also included for all words, all of which can be used as nouns. The programalso allows sampling on the basis of length in number of letters, and on the