Normative studies were carried out with 8 adult subjects whose averaged evoked responses to auditory stimuli were scored visually and by a number of differently defined quantitative methods (machine scoring). In order to compare different scoring methods a common signal-to-noise ratio measure was introduced based on a model where noise is distributed normally and signal is additive. Visual and machine scoring proved to be approximately equally sensitive, but the latter lends itself to a superior testing procedure which takes only one-third as long for equivalent results and is not contaminated by subjective error. The authors believe that the greatest promise for improving evoked response audiometry lies in the exploration of the new techniques proposed earlier. Progress in proving out one of these techniques, fast periodic stimulation, is reported.