Context-effects in judgements of length.

Abstract

Research on judgment has often involved experimental manipulation of the stimulus-context in which simple perceptual judgments are made. It is a commonplace that all judgments are relative; the problem has been to discover what they are relative to-that is, to specify how different features of the context determine the judgments. This has been the central concern of the theory of adaptation-level which asserts that the category-scale for psychophysical judgments is centered at the weighted mean of all stimuli affecting the judgments.l Recent research has indicated, however, that the mean may not be the most useful parameter for characterizing the stimuluscontext. Independent manipulation of either the midpoint (the mean of the two end-stimuli) or the median of the set of stimuli presented for judg, ment had significant effects upon the centering of the scale of judgment, but manipulation of the mean appeared to have no effect when midpoint and median were held constant.2 These findings were interpreted as consistent with the hypothesis that the scale of judgment reflects a compromise between two different tendencies: (1) to divide the range of stimuli presented for judgment into proportionate sub-ranges; and (2) to use the alternative categories of judgment with proportionate frequencies. This interpretation was based upon experiments studying absolute judgments of numerical magnitude. The present research was designed to investigate the same variables with a more clearly perceptual dimension, length of line. In other respects, the basic logic of the research remained unchanged. Thus, Ss made absolute judgments of the lengths of different sets of lines. The sets differed with respect to the physical values of their midpoints, medians, and means. The primary purpose was to determine how the scale of judgment is affected by variation of these parameters of