Abstract
The purpose of this paper is to calculate the upper limit of the correlation between normal and dichotomous variables. An empirically obtained correlation should be evaluated in view of this limit, instead of the usual limit of Pearson correlation.
Get full access to this article
View all access options for this article.