Multidimensional scaling of binary data for homogeneous groups / BEBR No. 411
Robert Redinger, Jagdish N. Sheth · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · 1976
A new methodology is proposed for perceptual mapping as an alternative to the nonmetric multidimensional scaling.The new methodology requires binary similarity judgments and utilizes the Ekart-Young decomposition procedure for mapping objects on a multidimensional space.The new technique is tested with respect to mapping of fifteen brands of soft drinks.The same study also collected rank-ordered similarity judgments and utilized the standard nonmetric scaling techniques as a comparative test.As expected, binary judgments were more reliable and produced a more meaningful multidimensional space than the nonmetric procedure.