Adaptive probit estimation and body size: the evaluation of a new psychophysical technique.

Studies of body size estimation are frequently used to identify body-image disturbances in clinical populations. No clear pattern of results has, however, so far emerged. One reason for this may be that studies confound non-sensory influences on performance deriving from motivational components with the observer's discriminative sensitivity. This study introduces an adaptive version of the method of constant stimuli and probit analysis (APE) to determine independently bias and threshold in the estimation of four body parts for two samples of undergraduate students. In the first study, test-retest reliability coefficients were obtained for body-size estimates, using the traditional technique and the new method. Estimates of subjective body size obtained by the new method were shown to have higher test-retest reliabilities than those obtained using traditional techniques. The stability of sensitivity estimates was in the .8-.9 range. In a second study we attempted to validate the new technique by comparing the body-size estimates of male and female observers. Independent bias and sensitivity estimates were found to be significantly different. Female observers were more likely to overestimate the size of their chest and waist whilst underestimating face size. There was no difference overall between the accuracy of male and female observers. Perceptual sensitivity did not differ significantly overall, but whilst female subjects showed a particularly low threshold for waist size, male observers showed the same specific sensitivity for thigh width. The extension of this method of measuring body-image distortions to relevant clinical populations is recommended.