Comparison of Correlation Measures for Nominal Data

Abstract

In social sciences, a plethora of studies utilize nominal data to establish the relationship between the variables. This, in turn, requires the correct use of correlation technique. The choice of correlation technique depends upon the underlying assumptions and power of the test of significance. The objective of the research is to explore the best measure of association for nominal data in terms of size, power and bias in estimation. Monte Carlo simulations reveal that the Phi and Pearson correlation statistics performs equally well in terms of size, power, and bias for naturally dichotomous variables. When both variables are artificially dichotomized, the Tetrachoric statistic has an edge in terms of bias to Pearson correlation statistic. If one variable is continuous and other is artificially dichotomized, the Biserial correlation measure turns out to be less biased as compared to Pearson statistic although both statistics exhibit similar power and size properties. If one variable is continuous and other is naturally dichotomized, it is hard to choose between the Point Biserial and Pearson correlation measures. Finally, if one variable is naturally dichotomous and other is artificially dichotomized, correlation coefficient V is compared with Pearson, Phi and Tetrachoric correlation techniques in terms of bias in estimate. The results indicate that the Tetrachoric statistic considerably overestimates the correlation value against non-normal distributions. Pearson and Phi correlation slightly underestimate the correlation value. In contrast, the correlation statistic V perform well.

Identification and ordering of safety performance indicators using fuzzy TOPSIS: a case study in Indian construction company

“…Spearman's rank correlation coefficient (ρ) measures the correlation between the ranks assigned using two different algorithms. It is less sensitive to the bias induced by the outliers (Islam and Rizwan, 2020). The formula used for the calculation of ρ is given by…”

Item-Level Psychometric Analysis of the Psychosocial Processes at Work Scale (PROPSIT) in Workers

“…The MIC provides a value between 0.0 (complete statistical independence) and 1.0 (complete dependence), indicating the strength of the relationship for some type of bivariate function relationship (not necessarily linear). Because it can be compared to the standardized covariance percentage (r 2 ) and measures of mutual information tend to be inefficient for detecting linear relationships [17,24], it was used in conjunction with the Spearman monotonic correlation to detect monotonic covariation. We used the R program testforDEP [73].…”

“…In exploratory contexts where the evaluation of statistical dependence between variables is performed, it is especially useful to detect the type of bivariate functional relationship that exists [24,25], without necessarily starting from a predefined hypothesis about some type of association, as usually occurs when starting with the search for linear relationships [16,24,26,27]. This point implies that the search for bivariate relationships should start by detecting some kind of existing association and serve as a condition for performing a second step where more specific or more complex relationships are tested (e.g., multivariate analysis), based on either a linear or nonlinear model.…”