Comparison of Scaling Methods: Paired Comparisons versus Constant-Sum

Abstract

15 stimuli were scaled by 100 Ss using the paired-comparison procedure; another 100 Ss scaled the same stimuli using the constant-sum procedure; and finally the data obtained from the constant-sum procedure were reanalyzed in terms of the paired-comparisons method. Three data plots of scale values were made: (a) paired-comparison—constant-sum, (b) paired-comparison—converted constant-sum, and (c) converted constant-sum—constant-sum. All 3 plots indicated a near linear relationship between scales. The findings were interpreted as supporting the notion that qualitative (metathetic) stimuli result in linear functions while quantitative (prothetic) stimuli result in logarithmic functions. It was concluded that the differences between the 2 methods were not sufficiently great to merit the use of the more demanding and time consuming constant-sum technique when scaling “general” stimuli.