A0928
Title: Hypothesis testing of varying correlations
Authors: Zhaoxia Yu - University of California Irvine (United States) [presenting]
Abstract: Understanding how correlations between two variables evolve in response to a third moderating variable remains a critical yet underexplored challenge across disciplines. While existing approaches (e.g., mediation analysis) address related questions, they do not directly quantify or statistically test hypotheses about constant correlations. The purpose is to introduce a methodological framework to rigorously characterize and evaluate how correlations vary conditional on a third variable. The approach integrates estimation techniques for dynamic correlation patterns with hypothesis-testing procedures to assess their statistical significance.
