This study investigated three sources of differences between recommended sets of equal-interval response anchors derived from two scaling techniques. One set, produced by magnitude estimations, is compared with a set produced by Thurstone Case III pair-comparison treatment of complete ranks. Subjects from two universities (N = 95, N = 110) completed magnitude estimations and rankings of 20 frequency expressions. Divergences in the results between the two methods were apparently due to differences between the samples as well as to the tasks and computational algorithms employed in the two approaches. Consensus anchor labels are identified and modifications in the magnitude estimation technique are recommended to improve its usefulness.
ASJC Scopus subject areas
- Developmental and Educational Psychology
- Applied Psychology
- Applied Mathematics