Home > Publications . Search All . Browse All . Country . Browse PSC Pubs . PSC Report Series

PSC In The News

RSS Feed icon

Almirall says comparing SMART designs will increase treatment quality for children with autism

Thompson says America must "unchoose" policies that have led to mass incarceration

Alter says lack of access to administrative data is "big drag on research"


Knodel honored by Thailand's Chulalongkorn University

Susan Murphy to speak at U-M kickoff for data science initiative, Oct 6, Rackham

Andrew Goodman-Bacon, former trainee, wins 2015 Nevins Prize for best dissertation in economic history

Deirdre Bloome wins ASA award for work on racial inequality and intergenerational transmission

Next Brown Bag

Monday, Oct 12 at noon, 6050 ISR
Joe Grengs: Policy & planning for transportation equity

Yu Xie photo

Estimating Heterogeneous Treatment Effects with Observational Data

Publication Abstract

Download PDF versionXie, Yu, Jennie Brand, and Ben Jann. 2011. "Estimating Heterogeneous Treatment Effects with Observational Data." PSC Research Report No. 11-729. February 2011.

Heterogeneous treatment effects are widely recognized but seldom studied empirically in quantitative sociological research. We suspect that lack of accessible statistical methods is one reason why heterogeneous treatment effects are not routinely assessed and reported. In this paper, we discuss a practical approach to studying heterogeneous treatment effects, under the same assumption commonly underlying regression analysis: ignorability. We specifically describe two methods. For the first method (SM-HTE), we begin by estimating propensity scores for the probability of treatment given a set of observed covariates for each unit and construct balanced propensity score strata; we then estimate propensity score stratum-specific average treatment effects and evaluate a trend across the strata-specific treatment effects. For the second method (MS-HTE), we match control units to treated units based on the propensity score and transform the data into treatment-control comparisons at the most elementary level at which such comparisons can be constructed; we then estimate treatment effects as a function of the propensity score by fitting a non-parametric model as a smoothing device. We illustrate the application of the two methods with a concrete empirical example.

Browse | Search : All Pubs | Next