ANOVA for longitudinal data with missing values

TL;DR

This paper develops a flexible, model-robust ANOVA method for longitudinal data with missing values, using empirical likelihood and semiparametric modeling to compare treatments effectively.

Contribution

It introduces a nonparametric ANOVA testing framework that accommodates various data types and models, including missing data, with a flexible semiparametric approach.

Findings

Effective treatment comparison in longitudinal studies with missing data.

Flexible framework applicable to parametric, semiparametric, and nonparametric models.

Robust nonparametric tests for treatment effects and interactions.

Abstract

We carry out ANOVA comparisons of multiple treatments for longitudinal studies with missing values. The treatment effects are modeled semiparametrically via a partially linear regression which is flexible in quantifying the time effects of treatments. The empirical likelihood is employed to formulate model-robust nonparametric ANOVA tests for treatment effects with respect to covariates, the nonparametric time-effect functions and interactions between covariates and time. The proposed tests can be readily modified for a variety of data and model combinations, that encompasses parametric, semiparametric and nonparametric regression models; cross-sectional and longitudinal data, and with or without missing values.