An empirical comparison of the Tucker-Messick and INDSCAL models: Measuring viewpoints about transit attributes

This investigation sought (1) to identify perceptually homogenous respondent groups by two individual differences scaling models initially proposed by Tucker and Messick, and Carroll and Chang, (2) to reveal the significance of group percepts with respect to anticipated satisfactions and socio-economic and activity pattern characteristics of group members, and (3) to test empirically the significance of formal distinctions between the two individual differences models. The study utilized judgments about 12 transit attributes of three innovative urban, public transportation modes from a sample of 243 respondents. It was possible to specify seven perceptually homogenous groups, which were distinct in terms of a qualitative analysis of their perceptual spaces and a quantitative convergent-discriminant validity analysis predicated on the distances between pairs of attributes in their spaces. The perceptual groups were shown to have interpretable links to socio-economic and activity pattern characteristics of the respondents. In addition, it was possible to statistically account for the satisfaction ratings of the respondents by the dimensions of their corresponding spaces. Since the Tucker-Messick model was shown to derive more distinct spaces for the separate groups than the INDSCAL model of Carroll and Chang, the Tucker-Messick spaces more uniquely tied a group's percepts to its corresponding satisfactions.