Judgement of direction of gaze: an interpretation of discrepant results.

Recent research on non-verbal behaviour in face-to-face social interaction has suggested that direction of gaze may be an important variable. However, investigators have paid little direct attention to the reliability and validity problems of this measure, relying largely on results by Gibson & Pick (1963) and Cline (1967) which showed good discrimination in an artificial experimental setting. Some recent studies have thrown doubt on the assumption of high validity of observer judgements of direction of gaze. This in turn throws doubt on results of studies using direction of gaze as a variable. Analysis of reliability and validity studies suggests that the discrepant results may be reconciled. It is argued that in natural interactions it may be sufficient if observers can accurately discriminate between gazes at the face of another and gazes directed away from the face, an easier task than in several laboratory studies. Nevertheless, in view of the possible bias effects involved, caution is needed in interpreting results involving assessments of gaze directions.