Number and Percentage Estimates in Sequential Tasks

Abstract

4 groups of Ss estimated percentage and absolute number of occurrence of specified symbols in auditively presented sequences. The groups differed in kind of stimuli used (letters vs numerals) and in type of attention (focused vs distributed). Within each group, number of stimulus categories, length of sequence, and frequency of target symbol were systematically varied. In all conditions, the accuracy of estimates was inversely proportional to frequency of occurrence; and percentage estimates were consistently more reliable than estimates of number. The findings suggest a sampling strategy, where both types of estimates are based upon a mental picture of relative frequencies.