Abstract

ABSTRACT. Unlike Andrew Kyngdon, I think the issue he has addressed is most informatively considered outside the confines of the representational theory of measurement. Then it becomes clear that while the theory of con-joint measurement is about situations like that treated by the Rasch model, the former isolates a different feature of those situations to the latter. But, if the relevant attributes are already presumed to be quantitative, the perceived differences are minimized and the Rasch model might seem to be a version of conjoint measurement. It is on this basis that Rasch modellers pursue their paradoxical quest for measurement. However, because the relevant attributes are not actually known to be quantitative, use of the Rasch model to measure psychological attributes remains logically dependent upon the outcome of research involving the theory of conjoint measurement or some-thing very similar. KEY WORDS: conjoint measurement, psychometrics, Rasch model, represen-tational theory The socio-economic conditions sustaining modern psychometrics discourage critical attitudes and it now stands suspended in a post-critical bubble. As a science, it desperately needs deflating. So it would be churlish to let my dis-agreement with the detail of Andrew Kyngdon’s (2008) argument prevent me applauding his critical spirit by offering further critical thoughts supplement-ing his on the relationship between the Rasch model and conjoint measure-ment. The former is, as he says, sometimes promoted as the latter. He favours a narrow focus, concentrating upon formal matters within the framework of representational theory. But a wider lens may be helpful, for the Rasch model does not easily fit the representational schema, as Kyngdon found, and the theory of conjoint measurement transcends its narrow frame.