Abstract
Changes in distance bisections as a function of practice were studied, two practice variables being manipulated in a 2 × 3 factorial design. One was the feedback variable: correction vs noncorrection; the other was type of practice: fractionation, bisection with shifting angle of regard, and control (continued bisection practice under standard conditions). Learning effects were measured in terms of judgments made on a set of pretests and posttests, involving bisections of the training distance as well as two new distances, and size matches. Ss were 48 undergraduates. It was found that correction produced a largely ephemeral attenuation of the overconstancy bias which marked all judgments; its effects transferred only to a limited extent to the new distances, and not at all to the size judgments. Type of practice did not result in any significant effects. A non-perceptual, judgmental explanation for the changes produced, and the overconstancy bias itself, is suggested.