A Note of Extreme Caution on the Use of Guttman Scales
Carmi Schooler · American Journal of Sociology · 1968
Since the change occurence of the various scale types is a function of the joint probability of the occurrence of their constituent responses, it is possible that Scalograms meeting currently acceptable levels of Scalabity and Reproducibility can occur by chance alone. Even Chilton's method, which determines whether the observed Reproducibility is significantly better than chance, does not guarantee that a scale is homogeneous; significant Reproductibility can occur even when three of six items are unrelated. Although very strict adherence to all of Ford's criteria might eliminate most chance-derived scales, the rule-of-thumb origins of these criteria limit their usefulness as rational proof that Guttman Scales are homogeneous.