On the plausibility of the latent ignorability assumption

TL;DR

This paper critically examines the latent ignorability assumption in instrumental variable models, highlighting its behavioral implications and analyzing its impact on causal effect estimates using empirical data from the Job Corps study.

Contribution

It formally discusses the behavioral implications of the latent ignorability assumption and empirically investigates its sensitivity in a real-world study.

Findings

Latent ignorability has strong behavioral implications in standard IV models.

Sensitivity analysis shows how estimates vary with different assumptions about attrition.

Empirical results from the Job Corps study illustrate the impact of LI on causal inference.

Abstract

The estimation of the causal effect of an endogenous treatment based on an instrumental variable (IV) is often complicated by attrition, sample selection, or non-response in the outcome of interest. To tackle the latter problem, the latent ignorability (LI) assumption imposes that attrition/sample selection is independent of the outcome conditional on the treatment compliance type (i.e. how the treatment behaves as a function of the instrument), the instrument, and possibly further observed covariates. As a word of caution, this note formally discusses the strong behavioral implications of LI in rather standard IV models. We also provide an empirical illustration based on the Job Corps experimental study, in which the sensitivity of the estimated program effect to LI and alternative assumptions about outcome attrition is investigated.