Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • HETTREATREG: module to compute diagnostics for linear regression when treatment effects are heterogeneous

    Dear All:

    I would like to announce that hettreatreg, a Stata module to compute diagnostics for linear regression when treatment effects are heterogeneous, is now available in SSC.

    hettreatreg implements several diagnostics for linear regression when the main interest is in the effects of a binary variable ("treatment") and these effects are potentially heterogeneous. hettreatreg represents OLS estimates of the effect of a binary treatment as a weighted average of the average treatment effect on the treated (ATT) and the average treatment effect on the untreated (ATU). The module estimates the OLS weights on these parameters, computes the associated model diagnostics, and reports the implicit OLS estimate of the average treatment effect (ATE). Further details are described in my paper, "Interpreting OLS Estimands When Treatment Effects Are Heterogeneous: Smaller Groups Get Larger Weights," which is available here: http://people.brandeis.edu/~tslocz/S...regression.pdf.

    Please let me know if you experience any problems with this module.

    Best wishes,

    Tymon

  • #2
    Dear Tymon,

    Thank you for contribution. I have found the appendix to your paper on your website and in it a more detailed presentation of the example that is included in the help file; which is really useful.

    In the appendix several diagnostic graphics are included. For educational purposes, could you also provide the code (do files) to create the figures in the appendix (i.e. D1.1-8, D2.1-8)?

    Best,
    Eric
    http://publicationslist.org/eric.melse

    Comment


    • #3
      Dear Eric,

      Thank you very much for your interest in my work. I agree that the appendix might be useful! Just to clarify, in case someone else reads this, I show some examples of possible usage of hettreatreg in Appendix C to my paper: http://people.brandeis.edu/~tslocz/S...regression.pdf.

      Because this paper is currently under resubmission to a journal, I would prefer not to post those additional codes publicly. I will definitely post them publicly if the paper is accepted for publication. Right now, I could send you this replication code via email. Would this work? Would you mind emailing me at [email protected]?

      Thanks again,

      Tymon

      Comment

      Working...
      X