Treatment interaction with non-experimental data in Stata

Graham K Brown, Thanos Mergoupis

Research output: Contribution to journalArticle

Abstract

Treatment effects may vary with the observed characteristics of the treated, often with important implications. In the context of experimental data, a growing literature deals with the problem of specifying treatment interaction terms that most effectively capture this variation. Some of the results of this literature are now implemented in Stata. With non-experimental (observational) data, and in particular when selection into treatment depends on unmeasured factors, treatment effects can be estimated using Stata's treatreg command. Although not originally designed for this purpose, treatreg can be used to consistently estimate treatment interactions parameters. In the presence of interactions, however, adjustments are required to generate predicted values and estimate the Average Treatment Effect (ATE). This paper introduces commands that perform this adjustment for the case of multiplicative interactions and shows the adjustment that is required for more complicated interactions.
Original languageEnglish
Pages (from-to)545-555
JournalStata Journal
Volume11
Issue number4
Publication statusPublished - Sep 2011

Fingerprint

Adjustment
Interaction
Treatment Effects
Average Treatment Effect
Estimate
Multiplicative
Experimental Data
Vary
Term
Context

Cite this

Treatment interaction with non-experimental data in Stata. / Brown, Graham K; Mergoupis, Thanos.

In: Stata Journal, Vol. 11, No. 4, 09.2011, p. 545-555.

Research output: Contribution to journalArticle

Brown, Graham K ; Mergoupis, Thanos. / Treatment interaction with non-experimental data in Stata. In: Stata Journal. 2011 ; Vol. 11, No. 4. pp. 545-555.
@article{748527e2019540409a839f5733ec65e0,
title = "Treatment interaction with non-experimental data in Stata",
abstract = "Treatment effects may vary with the observed characteristics of the treated, often with important implications. In the context of experimental data, a growing literature deals with the problem of specifying treatment interaction terms that most effectively capture this variation. Some of the results of this literature are now implemented in Stata. With non-experimental (observational) data, and in particular when selection into treatment depends on unmeasured factors, treatment effects can be estimated using Stata's treatreg command. Although not originally designed for this purpose, treatreg can be used to consistently estimate treatment interactions parameters. In the presence of interactions, however, adjustments are required to generate predicted values and estimate the Average Treatment Effect (ATE). This paper introduces commands that perform this adjustment for the case of multiplicative interactions and shows the adjustment that is required for more complicated interactions.",
author = "Brown, {Graham K} and Thanos Mergoupis",
year = "2011",
month = "9",
language = "English",
volume = "11",
pages = "545--555",
journal = "Stata Journal",
issn = "1536-867X",
publisher = "DPC Nederland",
number = "4",

}

TY - JOUR

T1 - Treatment interaction with non-experimental data in Stata

AU - Brown, Graham K

AU - Mergoupis, Thanos

PY - 2011/9

Y1 - 2011/9

N2 - Treatment effects may vary with the observed characteristics of the treated, often with important implications. In the context of experimental data, a growing literature deals with the problem of specifying treatment interaction terms that most effectively capture this variation. Some of the results of this literature are now implemented in Stata. With non-experimental (observational) data, and in particular when selection into treatment depends on unmeasured factors, treatment effects can be estimated using Stata's treatreg command. Although not originally designed for this purpose, treatreg can be used to consistently estimate treatment interactions parameters. In the presence of interactions, however, adjustments are required to generate predicted values and estimate the Average Treatment Effect (ATE). This paper introduces commands that perform this adjustment for the case of multiplicative interactions and shows the adjustment that is required for more complicated interactions.

AB - Treatment effects may vary with the observed characteristics of the treated, often with important implications. In the context of experimental data, a growing literature deals with the problem of specifying treatment interaction terms that most effectively capture this variation. Some of the results of this literature are now implemented in Stata. With non-experimental (observational) data, and in particular when selection into treatment depends on unmeasured factors, treatment effects can be estimated using Stata's treatreg command. Although not originally designed for this purpose, treatreg can be used to consistently estimate treatment interactions parameters. In the presence of interactions, however, adjustments are required to generate predicted values and estimate the Average Treatment Effect (ATE). This paper introduces commands that perform this adjustment for the case of multiplicative interactions and shows the adjustment that is required for more complicated interactions.

UR - http://people.bath.ac.uk/gkb22/resources.html

UR - http://www.stata-journal.com/article.html?article=st0240

M3 - Article

VL - 11

SP - 545

EP - 555

JO - Stata Journal

JF - Stata Journal

SN - 1536-867X

IS - 4

ER -