Abstract
ABR waveforms from 50 multi-handicapped children were analysed by nine judges in an investigation of scoring bias. Prior to estimating threshold for each subject, the judges were provided with either true or false ABR thresholds. This preliminary information was true in 25 of the cases and false in the other 25. The same ABR waveforms were evaluated one week later, but true/false biasing thresholds were reversed. Results revealed that while the more experienced judges were more accurate in their assessments, they were not necessarily less biased. The effects of bias were more predominant in those cases where the estimation of threshold was the most difficult.