Correlation coefficient derived from Pearson linear coefficient for ordinal and dichotomic variables

Abstract

As the Correlation Analysis is used in several knowledge areas, it is important to be familiar with the different methods existing to obtain correlation coefficients. Usually, we can find situations where variables are not measured by interval level but according to ordinal and/or dichotomic. The present work shows the correlation coefficient methods derived from Pearson linear coefficient and applied to situations involving variables measured by interval, ordinal and dichotomic levels, such as: bi-serial point, phi and Spearman correlation coefficient, between interval and ordinal variables, and between ordinal e dichotomic variables. Keywords: correlation coefficient; correlation coefficient for ordinal and dichotomic variables.