The British Journal of Psychiatry 143: 487-489 (1983)
© 1983 The Royal College of Psychiatrists
ME Dewey
Several coefficients have been suggested for the measurement of agreement between raters. Recently Maxwell (1977) has proposed, and Janes (1979) has extended, a new coefficient, which they have termed the random error coefficient (RE). This paper criticizes their proposal and suggests that Cohen's (1960) coefficient kappa (K) is still the measure of choice.