Abstract

This study compared the results of unfolding analyses, a technique to visualize judges and judged objects in a single figure, using ranking and rating data. Multidimensional scaling models, such as the unfolding model, are relatively free from statistical assumptions and can be advantageous tools for researchers interested in small groups or individual differences. The data used in the analysis were from 37 participants who were asked to measure 33 behaviors on the blameworthiness dimension, and the data were collected in two ways for each of the 33 behaviors. The data from paired comparison method was converted into ranked data, and the rated data was used as raw data. To understand the relationship between the ranking and rating data, we calculated the rank correlation coefficient (rho) between the aggregated statistics of the two data, and compared the correlation matrices of the 33 behaviors from the two data. The results showed that the two data were not related to each other, and the correlation matrices were not similar. As a result of the unfolding analysis, the placement of the behaviors was more distributed when using the ranking data than the rating data. Therefore, the author concluded that ranking data would be more useful to visually understand the relationship between the judgment objects and the judges. This study showed that the results of the unfolding analysis of ranking and rating data from the same judges are different. The current study suggested that, when applying the unfolding model, researchers should consider which type of data can be useful for the individual difference study.