This paper investigates two hypotheses of the screening phases of two-staged multiattribute choice procedures: the editing phase in which dominated choice alternatives are eliminated and the compatibility test of image theory. As the importance weights and rejection thresholds of the compatibility test are recalcitrant to elicitation, we develop a new method to examine the compatibility phase treating importance weights and rejection thresholds as latent parameters. Our method allows a checkup of the application of a compatibility test without any knowledge of subjects' importance weights and rejection thresholds. Our experimental findings show that the compatibility test enjoys consistency rates of some 70% as compared to consistency rates of about 15% for the editing hypothesis of the elimination of dominated choice alternatives. This result is supported by applying Selten's measure of predictive success.