ABSTRACT
Individual Differences Scaling is a form of weighted multidimensional scaling. Its basic assumption is that each individual responds to all the dimensions of the stimuli but may utilize the dimensions in varying degrees. The initial individual differences methodology was developed by Carroll and Chang and promoted through the effective FORTRAN program sindscal. A SINDSCAL analysis is performed on the scalar product distances between objects. Carroll and Chang obtain convergence with an algorithm called CANDECOMP for Canonical Decomposition similar to singular value decomposition (SVD). A canonical form of a matrix is the most convenient and usually simplest form to which a square matrix can be reduced by a certain type of transformation. Young developed ALSCAL which stands for Alternating Least Squares Analysis which is used by SAS for classical MDS as well as weighted MDS. AlSCAL is a repetitive analysis procedure.
