On measuring psychometric functions: a comparison of the constant-stimulus and adaptive up-down methods.

Psychometric functions were obtained using the conventional constant-stimulus method and an adaptive up-down method, for both computer-simulated and human observers. Except when the stimuli are closely placed, psychometric functions obtained with the adaptive method are as accurate as those obtained with the constant-stimulus method. Empirically, the adaptive method has some potential advantages owing to its ability to automatically concentrate the trials within the dynamic range of the psychometric function. It needs no pilot measurements for setting the signals as required by the constant-stimulus method. Furthermore, following a marked change in the underlying psychometric function, the distribution of the trials is automatically readjusted. Thus, on the basis of empirical considerations, the adaptive method is a better choice than the constant-stimulus method for measuring psychometric functions.