[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Imbalance in control versus treated group, and weights

From   "Austin Nichols" <[email protected]>
To   [email protected]
Subject   Re: st: Imbalance in control versus treated group, and weights
Date   Thu, 9 Oct 2008 17:16:24 -0400

Note that there are references in the help for -rd- (ssc inst rd,
replace) which mention specific examples and details on assumptions.
According to your first post, you have a dummy that increased
Pr(treatment), not a discontinuous jump in Pr(treatment) at some
cutoff in a continuous var, so no RD seems possible, and you expect
your dummy to have a direct impact on outcomes as well, so no
Instrumental Variables either.  It's unclear whether a
matching/reweighting approach would work for you.

On Thu, Oct 9, 2008 at 4:56 PM,  <[email protected]> wrote:
> I have another question. I followed the advice and looked into propensity score reweighting (PSR) and regression discontinuity (RD). Google pointed me to Austins presentation about this topis,
> I have read through the presentation, but I do not understand all the assumptions that underpins RD. My problem pass the first assumption that my treatment is not randomly assigned, though it started out as a randomized controlled trial, just that not all those supposed to have a treatment got one. Further, the assignment variable is based on a observable variable. Or well, it was not supposed to be an assignment variable, but it turned out to be, and consequently contaminated the treated versus the control group.
> However I am uncertain what the second assignment is telling me, quoting Austins presentation
> "The crucial second assumption is that there is a discontinuity at some cutoff value of the assignment variable in the level of treatment."
> My assignment variable do produce a jump in the level of treatment, but I am unsure whether this actually means that I pass assumption 2?
> I also downloaded the RD package from SSC (findit regression discontinuity). However, I am still unclear how I can relate the provided example to my own problem. I am having trouble locating other examples, and any tip would be greatly appreciated.
> Best wishes,
> Alexander Severinsen
*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index