The advantages of the “constant-effect” procedure are perhaps even more clear in sensory experiments.
In a typical experiment on visual dark adaptation the subject’s eyes are first exposed to light for a preliminary interval of time.
Then the light is cut off and the increase in visual sensitivity is measured by determining the intensity of a test flash that the subject is just able to see at various intervals of time after the light is cut off. In this case, we assume that when a test light can barely be detected, it is producing a relatively constant level of responding on the part of the visual system, and accordingly we express the visual sensitivity of the organism as the reciprocal of that amount of light Much less satisfactory would be the “constant-treatment” procedure of using a fixed intensity of test flash throughout the process of dark adaptation and asking the subject to give repeated estimates of its apparent brightness.