Improved Methods for Making Inferences About Multiple Skipped Correlations

TL;DR

This paper introduces improved methods for making inferences about multiple skipped correlations, addressing limitations of existing techniques and enhancing robustness against outliers in multivariate data analysis.

Contribution

It proposes alternative approaches for skipped correlation inference that overcome limitations of current methods, improving robustness and error control in multivariate data analysis.

Findings

New methods better handle outliers in correlation testing

Enhanced control of Type I error rates

Applicable to high-dimensional data scenarios

Abstract

A skipped correlation has the advantage of dealing with outliers in a manner that takes into account the overall structure of the data cloud. For p-variate data, $p \ge 2$, there is an extant method for testing the hypothesis of a zero correlation for each pair of variables that is designed to control the probability of one or more Type I errors. And there are methods for the related situation where the focus is on the association between a dependent variable and $p$ explanatory variables. However, there are limitations and several concerns with extant techniques. The paper describes alternative approaches that deal with these issues.