The experiment reported in this document extended the study by Milburn and Bell (1969) of English word frequencies. Specifically, this experiment examined the influence of instructional set on the relationship between affect and frequency estimates, with a distinction between personal and nationwide norms. The experiment design provided for two within-subjects variables--natural language frequency and affect (three categories: positive, neutral, and negative)--and one between-subjects variable--instructions (two categories: personal and Kucera-Francis). Sixty-nine subjects completed the two tests. Results showed that there is a curvilinear relationship between affect and estimates of English word frequency; that subjects believe that positive and negative English words occur equally often, but more often than neutral words; that the influence of affect on frequency estimates was strongest in the personal instructions condition and in the lower frequency range; and that subjects tended to make more conservative judgments in the Kucera-Francis condition than in the personal condition. (JM)