Global testing under sparse alternatives: ANOVA, multiple comparisons and the higher criticism

TL;DR

This paper investigates the effectiveness of ANOVA, multiple comparisons, and higher criticism methods for detecting sparse signals in high-dimensional linear models, establishing their optimality under various sparsity regimes.

Contribution

It characterizes the optimality of classical and modern testing procedures across different sparsity levels, highlighting the superiority of higher criticism in moderate to strong sparsity.

Findings

ANOVA is optimal under moderate sparsity ($0 extless\alpha extlessrac{1}{2}$) with certain design conditions.

Multiple comparison procedures are optimal when $ ext{ extgreater}3/4$ sparsity.

Higher criticism is powerful and optimal for the entire range $ extgreater 1/2$ sparsity.

Abstract

Testing for the significance of a subset of regression coefficients in a linear model, a staple of statistical analysis, goes back at least to the work of Fisher who introduced the analysis of variance (ANOVA). We study this problem under the assumption that the coefficient vector is sparse, a common situation in modern high-dimensional settings. Suppose we have $p$ covariates and that under the alternative, the response only depends upon the order of $p^{1-\alpha}$ of those, $0\le\alpha\le1$. Under moderate sparsity levels, that is, $0\le\alpha\le1/2$, we show that ANOVA is essentially optimal under some conditions on the design. This is no longer the case under strong sparsity constraints, that is, $\alpha>1/2$. In such settings, a multiple comparison procedure is often preferred and we establish its optimality when $\alpha\geq3/4$. However, these two very popular methods are suboptimal, and sometimes powerless, under moderately strong sparsity where $1/2<\alpha<3/4$. We suggest a method based on the higher criticism that is powerful in the whole range $\alpha>1/2$. This optimality property is true for a variety of designs, including the classical (balanced) multi-way designs and more modern "$p>n$" designs arising in genetics and signal processing. In addition to the standard fixed effects model, we establish similar results for a random effects model where the nonzero coefficients of the regression vector are normally distributed.