Two-alternative versus three-alternative procedures for threshold estimation

Abstract

The relative efficiency of two- and three-alternative psychophysical methods to measure auditory thresholds in naive listeners was assessed for both adaptive staircase and maximumlikelihood procedures. Threshold measures were obtained for a 1000-Hz sinusoid presented in noise across six blocks of 50 trials each. Ten independent subjects were tested by each of four methods: the two-alternative adaptive staircase, the two-alternative maximum-likelihood, the three-alternative adaptive staircase, and the three-alternative maximum-likelihood procedure. The four methods obtained comparable threshold estimates. Differences were found between procedures on measures of the variability of threshold values, both between and within subjects. In general, the three-alternative adaptive staircase is recommended as the most efficient procedure in most standard psychophysical measurement situations, but there are instances in which one of the other procedures would be preferred. © 1984 Psychonomic Society, Inc.