Familiarity Scale Values for 420 Nouns in Twelve Combinations of Frequency of Occurrence and Conceptual Categorization

Abstract

A category-scaling technique was used to obtain familiarity scale values for 420 nouns. The judges were instructed to base their ratings of familiarity on exposure frequency. The nouns were selected with random numbers from word pools of approximately equal sizes in the Thorndike-Lorge (T-L) source with control of word length and syllabification. 12 sets of words, with 35 words in each set, were scaled. The word sets were formed from words in three T-L frequency categories (.22 < 1, 1–4, and 5–100+ words per million) and four conceptual categories (names of animals, persons, occupations and articles of dress). Correlation coefficients which ranged from .68 to .95 were obtained between scale values having frequency-specific anchors and scale values with anchors which were not specific to the frequency categories. Indices of internal consistency show that empirical cumulative rating proportions were predicted from normal-theory cumulative proportions with an average error of .016.