The sequential structure of the quantal experiment

A linear model is proposed for the interrelation of successive responses in the quantal experiment. The model is based on a direct mechanism of response linkage rather than on an indirect relation by means of fluctuations in sensitivity, because Neisser's published results implicate the former mechanism rather than the latter. The proportion of increments detected is an unbiased estimator of the detection probability and has a statistical efficiency of only 20%. In view of this unexpectedly low efficiency it is possible that rectilinear psychometric results have occurred by means of a biased, but unwitting, selection of data.