Michael Baumann, Graeme Moffat, Larry E Roberts, and Lawrence M Ward (2004)
Constrained scaling: achieving quantitative convergence across laboratories
In: Fechner Day 2004, ed. by A. Oliveira, pp. 304-309, International Society for Psychophysics, Coimbra, Portugal.
It is well known that, although psychophysical scaling produces good qualitative agreement between experiments using ostensively the same methods but run in different laboratories, quantitative agreement is much more difficult to achieve. Constrained scaling, in which observers first learn a standard scale and then make magnitude judgments of other stimuli using the learned scale, has achieved excellent quantitative agreement between individual observers’ psychophysical functions and could theoretically do the same for across-laboratory comparisons. We report two experiments that were replicated, using constrained scaling, in two different laboratories as examples of the level of agreement achievable with this technique. In general, we found acrosslaboratory agreement superior to that typically obtained with magnitude estimation.