Estimating Human Error Probabilities from paired Ratios

Abstract

The ratio estimates of subjective human error probabilities obtained from paired comparisons are converted to a best set of relative human error quotients; this subjective ratio scale is then converted into an objectified scale of probabilities using a known objective empirical (anchor) probability. When dependable estimates of human error probabilities are unavailable, the pairwise ratio estimation procedure gives an excellent means of integrating many experts' judgements.