Abstract:Researchers in many fields endeavor to estimate treatment effects by regressing outcome data (Y) on a treatment (D) and observed confounders (X). Even absent unobserved confounding, the regression coefficient on the treatment reports a weighted average of strata-specific treatment effects (Angrist, 1998). Where heterogeneous treatment effects cannot be ruled out, the resulting coefficient is thus not generally equal to the average treatment effect (ATE), and is unlikely to be the quantity of direct scientific or policy interest. The difference between the coefficient and the ATE has led researchers to propose various interpretational, bounding, and diagnostic aids (Humphreys, 2009; Aronow and Samii, 2016; Sloczynski, 2022; Chattopadhyay and Zubizarreta, 2023). We note that the linear regression of Y on D and X can be misspecified when the treatment effect is heterogeneous in X. The "weights of regression", for which we provide a new (more general) expression, simply characterize how the OLS coefficient will depart from the ATE under the misspecification resulting from unmodeled treatment effect heterogeneity. Consequently, a natural alternative to suffering these weights is to address the misspecification that gives rise to them. For investigators committed to linear approaches, we propose relying on the slightly weaker assumption that the potential outcomes are linear in X. Numerous well-known estimators are unbiased for the ATE under this assumption, namely regression-imputation/g-computation/T-learner, regression with an interaction of the treatment and covariates (Lin, 2013), and balancing weights. Any of these approaches avoid the apparent weighting problem of the misspecified linear regression, at an efficiency cost that will be small when there are few covariates relative to sample size. We demonstrate these lessons using simulations in observational and experimental settings.

Exact expressions for the weights used in least-squares regression estimation for the log-logistic and Weibull distribution

A compromise criterion for weighted least squares estimates

New results and regression model for the exponentiated odd log-logistic Weibull family of distributions with applications

Resurrecting Weighted Least Squares

Exact Solutions in Log-Concave Maximum Likelihood Estimation

Robust explicit estimation of the log-logistic distribution with applications

A REPRESENTATION FOR THE REGRESSION COEFFICIENTS IN WEIGHTED LEAST SQUARES

A weighted least squares procedure to approximate least absolute deviation estimation in time series with specific reference to infinite variance unit root problems

Understanding and avoiding the "weights of regression": Heterogeneous effects, misspecification, and longstanding solutions

New results and regression model for the exponentiated odd log-logistic family with applications

On the implied weights of linear regression for causal inference

Asymptotic Theory of Least Distances Estimate in Multivariate Linear Models

On the correspondence of deviances and maximum likelihood and interval estimates from log-linear to logistic regression modelling

Tukey’s biweight estimation for uncertain regression model with imprecise observations

Weighted asymmetric least squares regression with fixed-effects

On comparison of estimators for proportional error nonlinear regression models in the limit of small measurement error

Estimating the failure rate of the log-logistic distribution by smooth adaptive and bias-correction methods

Weighted Lasso Estimates for Sparse Logistic Regression: Non-Asymptotic Properties with Measurement Errors

The log-weighted exponential regression model: alternative to the beta regression model

Essential formulae for restricted maximum likelihood and its derivatives associated with the linear mixed models

Bias-Corrected Maximum Likelihood Estimation of the Parameters of the Weighted Lindley Distribution