The effects of shape, size, heterogeneity, and instructional set on the judgment of visual number.

This early work was extended by Dallenbach and his students and Fernberger and his students in the 1920's, and by Taves and by Saltzman and Garner in the 1940's, and more recently by Reese and his colleagues.! During the course of this work, three separate processes in the discrimination of number have been studied: counting, estimating, and subitizing.' In the words of Reese and his colleagues, "Estimating is what occurs when the stimulus-number is greater than six; subitizing is what occurs when the stimulus-number is less than six." Subitizing is considerably more accurate, more rapid, and more 'confident' than estimating. With regard to the estimation of the number of objects, previous investigators report errors increase rapidly as the number of objects increases up to 100. Beyond this point, though errors increase, they do so at a slower rate. Among the factors found to influence the estimation of number are illumination," instructional set," differential reinforcement,' variations in the size of the area occupied by the pre-