Francis J. Di Vesta

The distribution of modifiers used by children in a word-association task

Summary A standardized word-association test was administered to 100 Ss in each of the grades 2–6 inclusive. The responses were restricted to modifiers. All inflected parts of speech used as modifiers were excluded from the analysis. The frequency for all other unique responses was tabulated. The number of words occupying the same frequency category was determined and a distribution made. The frequency distribution was found to conform reasonably to the lognormal variant of Zipfs law. The distribution of the total occurrences comprises the first moment of the distribution of occurrence-frequency categories. The two distributions were parallel and separated by the distance demanded if the conditions of lognormality were to be met. The transformation provides an interesting description of the language behavior of the children. The estimated parameters indicated a trend toward preference in the use of qualifier types common to the majority of respondents in the later grades, i.e., stereotypy in the kinds of words used. There was a significant progressive increase over age levels in the diversity of words used and in the number of words used in the categories of lower frequency of occurrence.

A simplex analysis of changes with age in responses to a restricted word-association task

Summary Modifiers were elicited to each of 100 substantives from 100 children in each of the grades 2–6 inclusive. Each modifier was then described by the following indices: frequency of occurrence, diversity of substantives to which it was applied, and partial H, an index combining frequency (or density) and diversity. The substantives were described by the H-index only. The hypothesis that these indices would undergo patterned change over age levels was tested against the simplex model with a least-squares solution. The empirical and theoretical matrices were compared by use of q2 for determining goodness of fit. In each case the hypothesis was supported adequately. The over-all changes over age groups were greatest for conceptual differentiation and least for language usage involving modes of qualifying. The implications of these data for measurement of semantic differentiation were indicated.