Current measurement of emotion often produces cognitive, behavioral, and physiological measures that disagree. Rather than accept the situation as unresolvable the study sought to improve the quality of measurement by removing much of the ambiguity inherent in typical measures of fear. Sixty five (sample 1) and 173 (sample 2) subjects were given an automated version of the Snake Anxiety Questionnaire and an extensive series of follow-up questions ending with a behavioral approach test. Follow-ups were asked contingent on the degree of ambiguity and the subject's response using branching logic. This system has been termed “Initial Response Verification”. The process was successful in reducing the number of false-positive (defined as high cognitive-low behavioral fear) identifications from 100 of 228 to 14 of 228, which was significant at the P<0.001 level. These reductions were distinctly different from those of a simple linear transform and were unique to each individual. In sum, behavioral and cognitive measures need not disagree when ambiguity is removed.