Abstract
In two experiments subjects judged the loudness of a 1000-Hz tone within a range of three loudnesses. The first used a within-subjects design with ranges of 1—2 sones, 1—4 sones, and 1—8 sones. The second used a between-subjects design with ranges of 1—2 sones, 1—8 sones, and 1—16 sones. Except for two values in the 1- to 2-sone range the two experiments produced comparable results. As range was increased the tendency to round away from the end points increased and the judged mid-point occurred at a higher value. Comparisons are made with rounding patterns that occur in studies of linear scales. The present results are interpreted to support Poulton's (1975) position, that the mid-point of subjective estimates depends on the range of stimuli judged, rather than that of Stevens (1975), that the mid-point on a prothetic scale is assigned to a value that is lower than the mid-point on the subject's subjective scale.
Get full access to this article
View all access options for this article.
