Some comments on Me use of Pearson’s coefficientasan index of interobserver agreernent.

Abstract

The aun of this paper was to discuss tite use of Pearson’s corre/a tion coefficient to evaluate sorne psychornetric properties of assessrnent techniques used in behavioral sta¬dies. Tite characteristic of these techniques are described first, and titen tite use of Pear¬Sons coefficient as a cøntparison statistic of standard scores, and as a linear ntodel of tite forrn ‘~‘t = a ± ¡3 X~ is discussed.It is suggested titat tite ,neasurernent of interobserver agreernent is only a special case of tite linear regression ntodel, and that tite use of this tecitnique in assessing inter¬observer agreernent is notjustsfied. Tite need for developing newpsychontetric tecitniques for behavioral studies is st re ssed.