Nomogram for the Tetrachoric Correlation Coefficient

Abstract

One of the most frequently used measures of association in psychology, and even in other biological sciences, is the tetrachoric correlation coefficient. Unfortunately, the computing of r_tet from the standard formula is extremely laborious, and the value of some aid to this end has been recognized from the earliest odays. This has become more important than ever, when correlations are prepared in hundreds for one factor analysis. The earliest of the devices for computing measures of association is the ABAC of Burt, which consists of a family of curves on one graph¹. Vernon has four such graphs². The best known of such aids are the computing diagrams of Thurstone³. There are forty-six such diagrams, each consisting of a family of curves. They are quite simple to read when working to two decimal places; but this can lead to considerable inaccuracy, and three places involves interpolating between two diagrams, with four readings each, a very time-consuming procedure.