Spatial frequency dependent observer bias in the measurement of contrast sensitivity

A previously reported spatial frequency dependent change in contrast sensitivity, apparently produced by the psychometric method used, was investigated in two experiments. A relative reduction in the measured contrast sensitivity function at higher spatial frequencies was found when using a simple single staircase procedure, even when all external characteristics of the presentation procedure were the same as a more sophisticated and more reliable psychometric method (adaptive probit estimation). The estimated false-negative response rate was found to be greater for the single staircase procedure and was greater at a higher spatial frequency, while the false-positive response rate to interleaved blank presentations did not vary between spatial frequencies and between the two psychometric methods. This supports the suggestion that some subjects, given a knowledge of the psychometric method controlling grating contrast, may adopt different decision criteria at different spatial frequencies.