ABSTRACT
This chapter discusses how unfolding methods can be used to analyze rating scale preferential choice data. These types of data are arranged as rectangular matrices with the individuals on the rows and the stimuli on the columns. The chapter presents a general discussion of the thermometers problem and proceed to an exposition of the MLSMU6 and SMACOF least squares metric unfolding methods. The smacofRect() function in the smacof package uses the SMACOF optimization procedure to perform metric unfolding on rectangular matrices. The chapter shows the estimated two-dimensional configuration of Danish voters and parties from the propensity to vote ratings. Since the MLSMU6 and SMACOF procedures for metric unfolding are both least squares optimization methods, it is appropriate to ask whether they produce meaningfully different point estimates. Both procedures perform identically well in the recovery of the true locations of the interest groups and legislators when there is no missing data.
