Current measurement of emotion often produces cognitive, behavioral, a
nd physiological measures that disagree. Rather than accept the situat
ion as unresolvable the study sought to improve the quality of measure
ment 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 redu
cing the number of false-positive (defined as high cognitive-low behav
ioral fear) identifications from 100 of 228 to 14 of 228, which was si
gnificant at the P < 0.001 level. These reductions were distinctly dif
ferent from those of a simple linear transform and were unique to each
individual. In sum, behavioral and cognitive measures need not disagr
ee when ambiguity is removed.