Using Goodman's (1975) notion of quasi-independence as a method of obtaining goodness of fit measures for non-scalable types in a scalogram analysis, archival data sets were examined using available Guttman scaling techniques, recent developments in latent structure analysis, and multidimensional scaling procedures. The Stouffer-Toby (1951) data were reanalyzed using Cloggs' Maximum Likelihood Latent Structure Analysis (MLLSA) and data from Schuman (1970) were reanalyzed using the quasi-independence model. The analysis of data showed that it was possible to extract a two or even higher dimensional solution which can be loosely interpreted. The data failed to produce a coefficient of reproducibility although an alternate goodness of fit measure provided by the model of quasi-independence show these data to define a Guttman scale tolerably well. Results vary widely regarding the Guttman scale depending on what criterion is used to test the data. Given the resultant data discrepancy, it would be unwise to draw conclusions about the adequacy of a Guttman scale on the basis of a single test. Researchers are encouraged towards sequential testing of unidimensionality and convergence of the three methods if Guttman scaling is to remain a meaningful methodological tool. (Author/PN)