Psychometric properties of dichotic words tests

Abstract

The current study examined the reliability of a measure of ear advantage and the total scores derived from the dichotic listening task. The scores for both measures were found: (1) not to differ from the normal distribution, and (2) to be reliable over items, tests, and time. However, the total scores were dependent upon stimulus materials and experience. In addition, a psychometric classification model was proposed and it was found that the majority, at least 51.6% of the subjects, did not exhibit a truly measureable ear advantage. Finally, it was argued that total scores as well as ear advantage should be considered in the analysis of dichotic listening studies.