Using Nonbinding Experiments to Estimate Causal Effects With Instrumental-Variables Regression: A Tutorial

Read full text Summarise with VIRAT
Authors: Sanford R Student
Cited by: 0
References: 54

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Often, causal effects of interest in psychological research involve causes that cannot be randomly assigned. Instrumental-variables (IV) regression, a cornerstone of the econometric literature on causal inference with nonrandom causes, allows for unbiased causal inference in the face of several threats to the validity of causal inference using ordinary-least-squares regression. IV regression is enabled by an instrument—a third variable that affects the outcome only through its influence on the hypothesized cause. One way to study causal effects in psychology is thus the random assignment of an intervention that drives change in the cause of interest but is otherwise unrelated to the outcome, followed by IV analysis to produce a valid causal effect. Moreover, IV regression models can be estimated straightforwardly as structural equation models (SEMs). Based on open data from a recently published study in the economics literature, in this tutorial, I outline the use of SEMs to estimate the causal effect of a behavior of interest, exercise, on an outcome of interest, academic performance, enabled by an instrument—a randomly distributed free gym card—that constitutes a random, nonbinding manipulation of exercise. The role of covariates in IV regression is outlined, and R code is provided for all analyses.

Fields of Study & Subjects

Advanced Causal Inference Techniques Statistical Methods and Bayesian Inference Psychometric Methodologies and Testing

Publication details

Journal / source
Advances in Methods and Practices in Psychological Science
Volume
9
Issue
3
Publication date
1 July 2026
Language
EN