We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Economics > Econometrics

Title: Interpreting OLS Estimands When Treatment Effects Are Heterogeneous: Smaller Groups Get Larger Weights

Abstract: Applied work often studies the effect of a binary variable ("treatment") using linear models with additive effects. I study the interpretation of the OLS estimands in such models when treatment effects are heterogeneous. I show that the treatment coefficient is a convex combination of two parameters, which under certain conditions can be interpreted as the average treatment effects on the treated and untreated. The weights on these parameters are inversely related to the proportion of observations in each group. Reliance on these implicit weights can have serious consequences for applied work, as I illustrate with two well-known applications. I develop simple diagnostic tools that empirical researchers can use to avoid potential biases. Software for implementing these methods is available in R and Stata. In an important special case, my diagnostics only require the knowledge of the proportion of treated units.
Subjects: Econometrics (econ.EM); Applications (stat.AP); Methodology (stat.ME)
Cite as: arXiv:1810.01576 [econ.EM]
  (or arXiv:1810.01576v3 [econ.EM] for this version)

Submission history

From: Tymon Sloczynski [view email]
[v1] Wed, 3 Oct 2018 04:01:27 GMT (104kb,D)
[v2] Sun, 9 Dec 2018 17:39:19 GMT (103kb,D)
[v3] Tue, 19 May 2020 22:04:18 GMT (5578kb,D)

Link back to: arXiv, form interface, contact.