paper Null condition results reveal interesting issues. This paper demonstrated a tendency for test participants to assume that each response in a discrimination procedure (e.g. left/right, or dimmer/brighter) should be used with equal frequency. This tendency persists even when the choice of reference stimulus meant these responses are not equally relevant. Stimulus frequency bias is sufficient to influence the results of null condition trials and between-stimulus trials.

Fotios SA, Cheal C. The effect of a stimulus frequency bias in side-by-side brightness ranking tests. Lighting Research and Technology 2008; 40(1); 43-54

paper A response to stimulus frequency bias is to have an equal number of conditions who are brighter and dimmer than that which is predicted to be equally bright as the reference. This approach however reinforces central tendencies, and means the assumed equal brightness setting is proven by default. In this article we suggest that an all-possible-pairs approach is used rather than a fixed reference stimulus.

Fotios S, Houser K. Using forced choice discrimination to measure the perceptual response to light of different characteristics. Leukos 2013; 9(4); 245-259