Abstract:Partial association refers to the relationship between variables <math>Y1,Y2,…,YK</math> while adjusting for a set of covariates <math>X={X1,…,Xp}</math>. To assess such an association when Yk's are recorded on ordinal scales, a classical approach is to use partial correlation between the latent continuous variables. This so-called polychoric correlation is inadequate, as it requires multivariate normality and it only reflects a linear association. We propose a new framework for studying ordinal-ordinal partial association by using Liu-Zhang's surrogate residuals. We justify that conditional on <math>X</math>, Yk, and Yl are independent if and only if their corresponding surrogate residual variables are independent. Based on this result, we develop a general measure <math>ϕ</math> to quantify association strength. As opposed to polychoric correlation, <math>ϕ</math> does not rely on normality or models with the probit link, but instead it broadly applies to models with any link functions. It can capture a nonlinear or even nonmonotonic association. Moreover, the measure <math>ϕ</math> gives rise to a general procedure for testing the hypothesis of partial independence. Our framework also permits visualization tools, such as partial regression plots and three-dimensional P-P plots, to examine the association structure, which is otherwise unfeasible for ordinal data. We stress that the whole set of tools (measures, p-values, and graphics) is developed within a single unified framework, which allows a coherent inference. The analyses of the National Election Study (K = 5) and Big Five Personality Traits (K = 50) demonstrate that our framework leads to a much fuller assessment of partial association and yields deeper insights for domain researchers. <a class="ext-link" href="https://doi.org/10.1080/01621459.2020.1796394">Supplementary materials</a> for this article are available online.

Relative Importance Analysis With Multicategory Dependent Variables:

Beyond step-down analysis: A new test for decomposing the importance of dependent variables in MANOVA.

Relative Use of Univariate and Multivariate Analysis of Variance in Developmental Psychology

Inferential Tools for Assessing Dependence Across Response Categories in Multinomial Models with Discrete Random Effects

Variable Importance Assessments and Backward Variable Selection for High-Dimensional Data

Assessing Variable Importance for Best Subset Selection

A Comparative Study of Imputation Methods for Multivariate Ordinal Data

The Importance of Variable Importance

One step at a time: A statistical approach for distinguishing mediators, confounders, and colliders using direction dependence analysis.

Difference in Differences and Ratio in Ratios for Limited Dependent Variables

RWA Web: A Free, Comprehensive, Web-Based, and User-Friendly Tool for Relative Weight Analyses

Testing Dependence Among Serially Correlated Multicategory Variables

A novel dependency-oriented mixed-attribute data classification method

The Analysis of Repeated Measures Designs Involving Multiple Dependent Variables

Multivariate differential association analysis

Variable importance in regression models

Multivariate Almost Stochastic Dominance: Transfer Characterizations and Sufficient Conditions Under Dependence Uncertainty

Variable importance: Comparison of selectivity ratio and significance multivariate correlation for interpretation of latent‐variable regression models

MDA for random forests: inconsistency, and a practical solution via the Sobol-MDA

Averaged Chain Substitution Method – Applicability, Advantages, and Disadvantages

Assessing Partial Association Between Ordinal Variables: Quantification, Visualization, and Hypothesis Testing