Interpreting TSLS Estimators in Information Provision Experiments

TL;DR

This paper analyzes the causal interpretation of TSLS estimators in information provision experiments, highlighting conditions under which they can be reliably used and providing guidance for empirical researchers.

Contribution

It characterizes TSLS estimators as weighted averages of causal effects and discusses their interpretation under various belief updating models, including passive and active control designs.

Findings

Some passive control estimators have negative weights, affecting causal interpretation

The framework applies to both passive and active control designs

Guidance is provided for practical implementation and interpretation

Abstract

To estimate the causal effects of beliefs on actions, researchers often run information provision experiments. We consider the causal interpretation of two-stage least squares (TSLS) estimators in these experiments. We characterize common TSLS estimators as weighted averages of causal effects, and interpret these weights under general belief updating conditions that nest parametric models from the literature. Our framework accommodates TSLS estimators for both passive and active control designs. Notably, we find that some passive control estimators allow for negative weights, which compromises their causal interpretation. We give practical guidance on such issues, and illustrate our results in two empirical applications.