Development of a 'bipolar' R-index

A new 'bipolar' R-index analysis was proposed and evaluated. Eighteen judges evaluated red color in eight wine samples by comparing each sample with the control. Judges indicated whether the sample had 'more', the 'same', or 'less' red color than the control, and whether they were sure or unsure of their decision. Three computational methods were used to examine the results: the 'traditional' R-index, the 'bipolar' R-index (R-more or R-less) and the 'weighted-bipolar' R-index. While all three methods provided consistent results, the 'bipolar' R-indices reflected bidirectional differences among the samples thus providing more information. A refinement to the computation ('weighted-bipolar' R-index) was an approach for eliminating the bias associated with overestimation of the sample size and accordingly changed some of the significance levels. Further research is currently underway to expand the scope and application of this method.