Skip to main content

ETHNOS_APP

Home • Search • Journals • List 0

Reducing Model Misspecification and Bias in the Estimation of Interactions

Bibliographic Data

ID6331978
AuthorsMatthew Blackwell (0000-0002-3689-9527, Harvard University, corresponding author), Michael P Olson (0000-0002-8023-1815, Washington University in St. Louis)
Year2022
Volume30
Issue4
Pages495-514
Publication date2022-10-01
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenuePolitical Analysis (JOURNAL)
Journal identifiersISSN: 1047-1987 • E-ISSN: 1476-4989
PublisherCambridge University Press (CUP) (PUBLISHER)
DOI10.1017/pan.2021.19
OpenAlexW3184944157
LanguageEN
Citations received38
References cited29

Analyzing variation in treatment effects across subsets of the population is an important way for social scientists to evaluate theoretical arguments. A common strategy in assessing such treatment effect heterogeneity is to include a multiplicative interaction term between the treatment and a hypothesized effect modifier in a regression model. Unfortunately, this approach can result in biased inferences due to unmodeled interactions between the effect modifier and other covariates, and including these interactions can lead to unstable estimates due to overfitting. In this paper, we explore the usefulness of machine learning algorithms for stabilizing these estimates and show how many off-the-shelf adaptive methods lead to two forms of bias: direct and indirect regularization bias. To overcome these issues, we use a post-double selection approach that utilizes several lasso estimators to select the interactions to include in the final model. We extend this approach to estimate uncertainty for both interaction and marginal effects. Simulation evidence shows that this approach has better performance than competing methods, even when the number of covariates is large. We show in two empirical examples that the choice of method leads to dramatically different conclusions about effect heterogeneity

Artificial neural network · Covariate · Econometrics · Estimator · Interaction · Lasso (programming language · Machine learning · Model selection · Multiplicative function · Overfitting · Regression · Regularization (linguistics · Statistics · Advanced Causal Inference Techniques · Computer Science · Mathematics · Statistical Methods and Bayesian Inference · Statistical Methods and Inference · Artificial Intelligence

  • Occupation-Specific Skills, Labor Market Context, and Preferences for Redistribution

    Josep Serrano-Serrat•The Journal of Politics•2025

  • Reactive Cosmopolitanism

    Open Access•Andreas Genoni, N Geurts•Global Networks•2026

  • Shouldering childhood

    Open Access•Maria Rubio-Cabañez, Josep Serrano-Serrat•European Sociological Review•2025

  • Quantities of Interest for Interactions and the Pitfalls of Assuming Linear Treatment Effects

    Open Access•Josep Serrano-Serrat•Sociological Science•2026

  • The Dual Impact of Organic Farming Support

    Open Access•Varacca Alessandro, Alessandro Varacca et al.•Business Strategy and the…•2026

  • Occupational labor markets, skills, and immigration concerns

    Open Access•Josep Serrano-Serrat•European Journal of Political…•2026

  • A representative survey experiment of motivated climate change denial

    Open Access•Lasse S Stoetzer, Florian Zimmermann•Nature Climate Change•2024

  • Legislative Organization and Political Representation

    Michael P Olson, Jon C Rogowski•Quarterly Journal of Political…•2024

  • Causal Machine Learning for Moderation Effects

    Open Access•Nora Bearth, Michael Lechner•Journal of Business and Economic…•2025

  • Impacts of Social Assistance During the Covid‐19 Pandemic in Poor Remote Communities

    Open Access•Yoshito Takasaki, Oliver T Coomes et al.•Journal of International…•2026

  • Gender Gaps in Immigrants' Political Participation within and across Borders

    Antoine Bilodeau, Colin Scott•Journal of Immigrant & Refugee…•2025

  • Inequality and Total Effect Summary Measures for Nominal and Ordinal Variables

    Open Access•Trenton D Mize, Bing Han•Sociological Science•2025

  • Framing electricity plan choices to enhance green energy usage

    Open Access•Robert Neumann, Guido Mehlkop•Energy Research & Social Science•2020

  • Generalizing Trimming Bounds for Endogenously Missing Outcome Data Using Random Forests

    Open Access•Cyrus Samii, Ye Wang et al.•Political Analysis•2025

  • Ethnic politics and sovereign credit risk

    Open Access•Kathleen J Brown, Matthew Digiuseppe et al.•Review of International Political…•2023

  • Automated decision-making in public administration

    Open Access•Aya Rizk, Ida Lindgren•Government Information Quarterly•2025

  • How information, social norms, and experience with novel meat substitutes can create positive political feedback and demand-side policy change

    Open Access•Lukas Paul Fesenfeld, Maiken Maier et al.•Food Policy•2023

  • Geography, development, and power

    Open Access•John Cruzatti C, Christian Bjørnskov et al.•World Development•2024

  • Race, legislative speech, and symbolic representation in Congress

    Open Access•Arjun Vishwanath•American Journal of Political…•2025

  • Do Political Preferences Affect Policy Learning and Uptake? Evidence from a Field Experiment with Informal Entrepreneurs

    Cesar Zucco, Anna-Katharina Lenz et al.•The Journal of Politics•2023

  • Americans’ Responses to Covid-19 and the Conditional Role of Dispositional Needs for Security

    Open Access•Adam R Panish, Trent Ollerenshaw et al.•Public Opinion Quarterly•2026

  • The Heterogeneous Associations of Rural Consciousness and Political Preferences

    Open Access•Trent Ollerenshaw•Political Behavior•2026

  • Capturing the Fourth Estate

    Open Access•Ruilin Lai•British Journal of Political…•2025

  • Incorporating Machine Learning into Sociological Model-Building

    Open Access•Megan Verhagen•Sociological Methodology•2024

  • The Consequences of Model Misspecification for the Estimation of Nonlinear Interaction Effects

    Open Access•Janina Beiser-Mcgrath, Liam F Beiser-Mcgrath•Political Analysis•2023

  • Longitudinal attenuation in political polarization

    Open Access•Rudolf Fürst, Rafael Goldszmidt et al.•Social Science & Medicine•2024

  • Group Ties amid Industrial Change

    Open Access•Noah Zucker•World Politics•2022

  • Sources of Partisan Change

    Alexander F Gazmararian•The Journal of Politics•2025

  • Household education gaps and gender role attitudes

    Open Access•Marco Giani, Diane Hope et al.•Political Science Research and…•2022

  • Technology Threats to Employment, Issues, and Candidate and Party Preferences in the United States

    Open Access•Tobias Heinrich, Christopher Witko•Political Research Quarterly•2022

  • Affect, Not Ideology

    Open Access•Sam Fuller, Nicolás De La Cerda et al.•Political Behavior•2025

  • Reducing Model Misspecification and Bias in the Estimation of Interactions

    Open Access•Matthew Blackwell, Michael P Olson•Political Analysis•2022

  • Carbon inequality and support for carbon taxation

    Open Access•Liam F Beiser-Mcgrath, Marius R Busemeyer•European Journal of Political…•2023

  • Informed or Overwhelmed? Disentangling the Effects of Cognitive Ability and Information on Public Opinion

    Open Access•Adam R Panish•British Journal of Political…•2025

  • Why Anxious People Lean to the Left on Economic Policy

    Open Access•Adam R Panish, Andrew W Delton•British Journal of Political…•2025

  • When expectations backfire

    Open Access•Andreas Genoni, Didier Ruedin•Social Forces•2025

  • Relationship between English proficiency and socioeconomic status in Asia

    Open Access•Takunori Terasawa•World Englishes•2024

  • Does taking parental leave make you a more likeable worker? Evidence from a survey experiment

    R J Petts, Trenton D Mize et al.•Community Work & Family•2025

  • The American Direct Primary

    Open Access•Alan Ware•American Direct Primary•2002

  • Bart

    Hugh A Chipman, Edward I George et al.•The Annals of Applied Statistics•2010

  • The Adaptive Lasso and Its Oracle Properties

    Hui Zou•Journal of the American…•2006

  • Some heteroskedasticity-consistent covariance matrix estimators with improved finite sample properties

    Open Access•James G MacKinnon, Halbert White•Journal of Econometrics•1985

  • Inference on Treatment Effects after Selection among High-Dimensional Controls

    Alexandre Belloni, Victor Chernozhukov et al.•The Review of Economic Studies•2014

  • Estimating treatment effect heterogeneity in randomized program evaluation

    Katsushi Imai, Kosuke Imai et al.•The Annals of Applied Statistics•2013

  • Regression Shrinkage and Selection Via the Lasso

    Open Access•Robert Tibshirani•Journal of the Royal Statistical…•1996

  • Inference in High-Dimensional Panel Models With an Application to Gun Control

    Alexandre Belloni, Victor Chernozhukov et al.•Journal of Business and Economic…•2016

  • Nonparametric Estimation of Average Treatment Effects Under Exogeneity

    Guido W Imbens•The Review of Economics and…•2004

  • The Decline of Third-Party Voting in the United States

    Sedi Hirano, Shigeo Hirano et al.•The Journal of Politics•2007

  • Problems with products? Control strategies for models with interaction and quadratic effects

    Open Access•Janina Beiser-Mcgrath, Liam F Beiser-Mcgrath•Political Science Research and…•2020

  • The causal interpretation of estimated associations in regression models

    Open Access•Luke Keele, Randolph T Stevenson et al.•Political Science Research and…•2020

  • At the Nexus of Observational and Experimental Research

    Open Access•Cindy D Kam, Marc Trussler et al.•Political Behavior•2017

  • Reducing Model Misspecification and Bias in the Estimation of Interactions

    Open Access•Matthew Blackwell, Michael P Olson•Political Analysis•2022

  • Sparse Estimation and Uncertainty with Application to Subgroup Analysis

    Open Access•Marc Ratkovic, Dustin Tingley•Political Analysis•2017

  • Understanding Interaction Models

    Open Access•Thomas Brambor, William R Clark et al.•Political Analysis•2006

  • How Much Should We Trust Estimates from Multiplicative Interaction Models? Simple Tools to Improve Empirical Practice

    Open Access•Jens Hainmueller, Jonathan Mummolo et al.•Political Analysis•2019

  • Kernel Regularized Least Squares

    Open Access•Jens Hainmueller, Chad Hazlett•Political Analysis•2014

  • Hypothesis Testing and Multiplicative Interaction Terms

    Bear F Braumoeller•International Organization•2004

  • Marginal Effects in Interaction Models

    Open Access•Justin E Esarey, Justin Esarey et al.•Comparative Political Studies•2018

  • Remittances and Protest in Dictatorships

    Open Access•Abel Escribà‐folch, Covadonga Meseguer et al.•American Journal of Political…•2018

  • Testing for Interaction in Binary Logit and Probit Models

    Open Access•William D Berry, Jacqueline Hr Demeritt et al.•American Journal of Political…•2010

  • Modeling and Interpreting Interactive Hypotheses in Regression Analysis

    Robert J Franzese, Robert Franzese et al.•Modeling and interpreting…•2007

Unique citing works38
Citations per year6,33
Citation span2020 - 2026 (7)
Citation velocitycurrent
Highly citedNo
Citation typesNeutral: 37
Ethnos_APP • Open Source Project • MIT License • Frontend v2.0.0 • Privacy and Cookies • API Documentation: api.ethnos.app/docs • API Source Code: GitHub • DOI: 10.5281/zenodo.17049435 • Frontend Source Code: GitHub • DOI: 10.5281/zenodo.17050053 • cruz.rio.br • Expectantes Misericordiae