Values and Estimation of Word Frequencies

Abstract

It was hypothesized that people scoring high on a value or need such as evaluative dependence (measured by the Marlowe-Crowne Social Desirability Scale, M-C SD) would be more sensitive to words relevant to this need than would low scoring Ss and thus would tend to overestimate the frequency of occurrence of the need-relevant words. Words chosen for frequency ratings varied on a good-bad dimension pertaining to human characteristics or behaviors, and appropriate word classification was determined by the evaluative ratings of an initial sample of Ss. Words within each category were also matched for actual frequency of occurrence according to Thorndike-Lorge counts. The results did not support the hypothesis since high and low M-C SD scorers did not differ in their frequency ratings. The major finding, replicated in a second experiment, was that all Ss rated evaluative words as occurring more frequently than neutral words, even though the actual frequency of occurrence of the word types was the same.