Are outcome variables always bad controls? The case of mediating variables

Qiang Chen, Ji Qi and Guanpeng Yan

Applied Economics Letters, 2025, vol. 32, issue 14, 2066-2070

Abstract: In an influential book, Angrist and Pischke (2009) claim that outcome variables are always bad controls. However, mediating variables are routinely included in the mediation analysis, despite their being outcome variables by definition. Angrist and Pischke (2009) demonstrate the potential issue of using outcome variables as controls by decomposing the estimand into a causal effect and a selection bias term. However, we show that the selection bias term vanishes under an exogeneity condition in the potential outcome framework and the regression framework. We also discuss a special type of outcome variable that is always a bad control, i.e. an outcome of the dependent variable.

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/13504851.2024.2332533 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:apeclt:v:32:y:2025:i:14:p:2066-2070

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/RAEL20

DOI: 10.1080/13504851.2024.2332533

Access Statistics for this article

Applied Economics Letters is currently edited by Anita Phillips

More articles in Applied Economics Letters from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().