A Covariate-Adjusted Homogeneity Test with Application to Facial Recognition Accuracy Assessment

TL;DR

This paper introduces a covariate-adjusted homogeneity test to compare accuracy across multiple rater groups, accounting for covariates, with theoretical validation and application to facial recognition data.

Contribution

The paper develops a new statistical test for rater accuracy comparison that adjusts for covariates, with theoretical derivation and empirical validation.

Findings

The test effectively detects differences in rater accuracy.

Simulation studies confirm good finite sample performance.

Applied to face recognition data, it identified significant group differences.

Abstract

Ordinal scores occur commonly in medical imaging studies and in black-box forensic studies \citep{Phillips:2018}. To assess the accuracy of raters in the studies, one needs to estimate the receiver operating characteristic (ROC) curve while accounting for covariates of raters. In this paper, we propose a covariate-adjusted homogeneity test to determine differences in accuracy among multiple rater groups. We derived the theoretical results of the proposed test and conducted extensive simulation studies to evaluate the finite sample performance of the proposed test. Our proposed test is applied to a face recognition study to identify statistically significant differences among five participant groups.