A Second Opinion Correlation Coefficient

An important decision in most fields of research is whether two variables are related. Pearson's correlation coefficient usually answers this question, but it lacks robustness and depends on normality. Thus, inappropriate decisions can be made because of a few data points skewing the conclusion. This leads either to handling complex data by the dubious process of throwing out a selected set of the data or using other correlation coefficients. Spearman or Kendall can be used but neither seems to be as robust as the Greatest Deviation Correlation Coefficient, which gives a reliable "second opinion, " as illustrated by the example presented. 1