A linear alternative to assimilation-contrast: A reply to Manis
Abstract
the relative goodness-of-fit between Upshaw's t condition data (obtained when his judges were exposed to items across the entire attitude continuum), and an assimilation and contrast model for social judgment; the main evidence cited is the fact that for subjects who showed both assimilation and contrast, assimilation was obtained on items relatively close to the judge's own stand, while contrast appeared on items that were farther removed. It is obvious, however, that this formulation is distressingly vague in its failure to specify the range of discrepancies within which assimilation should occur, and beyond which contrast effects are to be expected. More work is needed here; inspection of Upshaw's data does not support the suggestion that this cutoff point coincides with the judge's latitude of acceptance (Hovland et al., 1957). It should be noted that the present considerations do not affect Upshaw's important conclusions concerning the interaction of own position and item context as determinants of social judgment. That is, the issues raised here are largely irrelevant to the finding that judgments remain relatively unchanged when the stimulus series is aborted at the subject's favored end of the continuum, while there is a displacement toward the "missing" items, if