Higher Moments for Optimal Balance Weighting in Causal Estimation June 24, 2022
We explore the potential benefit of using higher-order moments in the balancing conditions for covariate-balancing propensity scores and entropy balance.
Research Using Propensity Scores
Observational research projects that include multiple study groups may rely on propensity scores to help researchers balance the study groups and draw the appropriate causal conclusions. A propensity score is the probability that, based on certain characteristics, a study participant would be assigned to a specific treatment group. Here we list a selection of RAND publications that describe research using this method.
-
-
A Tutorial Comparing Different Covariate Balancing Methods with an Application Evaluating the Causal Effects of Substance Use Treatment Programs for Adolescents June 24, 2022
A state-of-the-art tutorial on how best to implement propensity score weighting analyses.
-
Volunteer Recovery Support for Adolescents: Using Propensity Score Based Methods to Understand Dosage Effects Within a Randomized Controlled Trial June 24, 2022
The study used statistical methods designed to approximate RCTs when comparing more than two nonequivalent groups that include an assessment of the potential impact of omitted variables in order to address potential dosage effects for a commonly used evidence based substance use treatment program for adolescents.
-
Brief Tutorial for twangContinuous March 10, 2021
This document provides a brief tutorial on using the twangContinous package to estimate causal effects for continuous exposure variables using generalized propensity scores estimated via generalized boosted models.
-
Toolkit for Weighting and Analysis of Nonequivalent Groups: A Tutorial on the TWANG Shiny Application for Time-Varying Treatments May 14, 2021
This tutorial demonstrates the use of the Toolkit for Weighting and Analysis of Nonequivalent Groups (TWANG) Shiny application for time-varying treatments to estimate the inverse probability of treatment weights (IPTW) for time-varying treatments when using observational data. The tutorial also shows how to calculate treatment effect estimates using those IPTW.
-
Quantifying the Bias Due to Observed Individual Confounders in Causal Treatment Effect Estimates: A Tutorial for the Selection Bias Decomposition (SBdecomp) Package December 29, 2020
In this tool, the authors explain the methodology behind the primary function of the selection bias decomposition (SBdecomp) package; describe its features, syntax and how to implement the function; and illustrate its use with an example. The package allows researchers to identify the most important observed confounder(s) in accounting for observed selection bias.
-
Optimally Balanced Gaussian Process Propensity Scores for Estimating Treatment Effects September 19, 2019
This paper introduces a new approach to estimating the propensity score using Gaussian processes and optimizing hyperparameters with respect to covariate balance.
-
Addressing Geographic Confounding Through Spatial Propensity Scores: A Study of Racial Disparities in Diabetes December 14, 2017
A spatial "doubly robust" estimator can minimize geographically related risk difference among racial/ethnic groups in health disparities studies.
-
Optimizing Variance-Bias Trade-Off in the TWANG Package for Estimation of Propensity Scores December 28, 2017
This discussion highlights potentially meaningful ways to optimize propensity score machine learning methods to allow for minimal bias and less variability.
-
Propensity Scores for Multiple Treatments: A Tutorial on the MNPS Command for Stata Users May 5, 2017
This tutorial explains the syntax and features related to the implementation of the MNPS commands in the Stata TWANG series.
-
Propensity Scores for Repeated Treatments: A Tutorial for the iptw Function in the TWANG Package September 14, 2017
This tutorial describes the use of the TWANG package in R to estimate inverse probability of treatment weights (IPTWs) when one has time varying treatments or sequences of treatments over time.
-
A Propensity-Score-Weighted Population-Based Study of the Health Benefits of Dogs and Cats for Children August 29, 2017
Counter to widely held beliefs, pet ownership does not improve children's general or psychological health.
-
Propensity Scores for Multiple Treatments: A Tutorial for the MNPS Macro in the TWANG SAS Macros March 13, 2015
This tutorial explains the syntax and features related to the implementation of the MNPS function in the SAS TWANG macros. The MNPS function, which stands for multinomial propensity scores, estimates propensity score weights for studies involving more than two treatment or exposure groups. The SAS TWANG macros were developed to support the use of the TWANG tools without requiring analysts to learn R.
-
A Tutorial on Propensity Score Estimation for Multiple Treatments Using Generalized Boosted Models January 1, 2013
The use of propensity scores to control for pretreatment imbalances on observed variables in non-randomized or observational studies examining the causal effects of treatments or interventions has become widespread over the past decade.
-
Bias and Variance Trade-Offs When Combining Propensity Score Weighting and Regression: With an Application to HIV Status and Homeless Men June 1, 2012
There is a bias-variance tradeoff at work in propensity score estimation; every step toward better balance usually means an increase in variance and at some point a marginal decrease in bias may not be worth the associated increase in variance.
-
Assessing the Effect of Race Bias in Post-Traffic Stop Outcomes Using Propensity Scores January 1, 2006
The present study proposes a propensity score technique to determine the extent to which race bias affects citation rates, search rates, and the duration of the stop. Adjusting for confounding variables using the propensity score offers an alternative to multivariate regression that is more interpretable, less prone to errors in model assumptions, and ultimately easier to present to stakeholders in policing practices.
-
Assessing the Effect of Race Bias in Post-traffic Stop Outcomes Using Propensity Scores January 1, 2006
Addresses the role race plays a role in officers' use of discretion in traffic stops by proposing a technique to determine the extent to which race bias affects citation rates, search rates, and the duration of the stop.
-
Propensity Score Estimation with Boosted Regression for Evaluating Causal Effects in Observational Studies January 1, 2004
Propensity score weights estimated using boosting eliminate most pretreatment group differences and substantially alter the apparent relative effects of adolescent substance abuse treatment.
