Question - Matchit - lists.gking.harvard.edu

List overview All Threads
Download

Question

matchit() silently adds a...

Using matchit for large datasets

georgebaah

5 Apr 2012 5 Apr '12

11:37 a.m.

Hi Everyone, I have a theoretical question. It's been bothering me for a while. I have a dataset from which I know the treatment is stochastic: 0 < Pr(T=1 | X) < 1. Also I know that Y is independent of T given X. However, T is not independent of Y given X. Does this make sense? Does this violate the theory behind the potential outcome model? Thanks. George

Reply

Show replies by date

Olivia Lau

5 Apr 5 Apr

12:50 p.m.

Think of it this way: If T were independent of Y | X, that is like saying that you expect the treatment to have no effect on the outcome Y, conditional on background characteristics. If this were the case, then the treatment T would be considered a failure in the real world, right? :) On Thu, Apr 5, 2012 at 8:37 AM, georgebaah <baah(a)gatech.edu> wrote:

Hi Everyone, I have a theoretical question. It's been bothering me for a while. I have a dataset from which I know the treatment is stochastic: 0 < Pr(T=1 | X) < 1. Also I know that Y is independent of T given X. However, T is not independent of Y given X. Does this make sense? Does this violate the theory behind the potential outcome model? Thanks. George- --- MatchIt mailing list served by HUIT List Address: matchit@lists.gking.harvard.**edu<matchit@lists.gking.harvard.edu> Subscribe/Unsubscribe: http://lists.gking.harvard.** edu/mailman/listinfo/ei<http://lists.gking.harvard.edu/mailman/listinfo/… MatchIt Software and Documentation: http://gking.harvard.edu/**matchit/<http://gking.harvard.edu/matchit/> Browse/Search List Archive: http://lists.gking.harvard.** edu/mailman/private/matchit/<http://lists.gking.harvard.edu/mailman/priv… Matchit mailing list Matchit(a)lists.gking.harvard.**edu <Matchit(a)lists.gking.harvard.edu> https://lists.gking.harvard.**edu/mailman/listinfo/matchit<https://lists…

Reply

georgebaah

1:20 p.m.

Hmmm, interesting. Maybe I am thinking too much. Let me put it this way. Suppose Y takes on two values y1, y2 and 0 < Pr(Y|T) < 1 but say Pr(T=1|Y=y1) = 1.0. Does this violate the potential outcome model? George On Apr 5, 2012, at 12:50 PM, Olivia Lau wrote:

Think of it this way: If T were independent of Y | X, that is like saying that you expect the treatment to have no effect on the outcome Y, conditional on background characteristics. If this were the case, then the treatment T would be considered a failure in the real world, right? :) On Thu, Apr 5, 2012 at 8:37 AM, georgebaah <baah(a)gatech.edu> wrote: Hi Everyone, I have a theoretical question. It's been bothering me for a while. I have a dataset from which I know the treatment is stochastic: 0 < Pr(T=1 | X) < 1. Also I know that Y is independent of T given X. However, T is not independent of Y given X. Does this make sense? Does this violate the theory behind the potential outcome model? Thanks. George- --- MatchIt mailing list served by HUIT List Address: matchit(a)lists.gking.harvard.edu Subscribe/Unsubscribe: http://lists.gking.harvard.edu/mailman/listinfo/ei MatchIt Software and Documentation: http://gking.harvard.edu/matchit/ Browse/Search List Archive: http://lists.gking.harvard.edu/mailman/private/matchit/ Matchit mailing list Matchit(a)lists.gking.harvard.edu https://lists.gking.harvard.edu/mailman/listinfo/matchit

Reply

Kosuke Imai

6 Apr 6 Apr

11:22 a.m.

Not at all. You don't want to condition on the observed outcome, which is different from potential outcomes. Kosuke Imai Princeton University http://imai.princeton.edu Sent from my iPhone On Apr 5, 2012, at 1:20 PM, georgebaah <baah(a)gatech.edu> wrote:

Hmmm, interesting. Maybe I am thinking too much. Let me put it this way. Suppose Y takes on two values y1, y2 and 0 < Pr(Y|T) < 1 but say Pr(T=1|Y=y1) = 1.0. Does this violate the potential outcome model? George On Apr 5, 2012, at 12:50 PM, Olivia Lau wrote:

Think of it this way: If T were independent of Y | X, that is like saying that you expect the treatment to have no effect on the outcome Y, conditional on background characteristics. If this were the case, then the treatment T would be considered a failure in the real world, right? :) On Thu, Apr 5, 2012 at 8:37 AM, georgebaah <baah(a)gatech.edu> wrote: Hi Everyone, I have a theoretical question. It's been bothering me for a while. I have a dataset from which I know the treatment is stochastic: 0 < Pr(T=1 | X) < 1. Also I know that Y is independent of T given X. However, T is not independent of Y given X. Does this make sense? Does this violate the theory behind the potential outcome model? Thanks. George- --- MatchIt mailing list served by HUIT List Address: matchit(a)lists.gking.harvard.edu Subscribe/Unsubscribe: http://lists.gking.harvard.edu/mailman/listinfo/ei MatchIt Software and Documentation: http://gking.harvard.edu/matchit/ Browse/Search List Archive: http://lists.gking.harvard.edu/mailman/private/matchit/ Matchit mailing list Matchit(a)lists.gking.harvard.edu https://lists.gking.harvard.edu/mailman/listinfo/matchit

- --- MatchIt mailing list served by HUIT List Address: matchit(a)lists.gking.harvard.edu Subscribe/Unsubscribe: http://lists.gking.harvard.edu/mailman/listinfo/ei MatchIt Software and Documentation: http://gking.harvard.edu/matchit/ Browse/Search List Archive: http://lists.gking.harvard.edu/mailman/private/matchit/ Matchit mailing list Matchit(a)lists.gking.harvard.edu https://lists.gking.harvard.edu/mailman/listinfo/matchit

Reply

4425

days inactive

4426

days old

matchit@lists.gking.harvard.edu

Manage subscription

3 comments

3 participants

tags (0)

participants (3)

georgebaah
Kosuke Imai
Olivia Lau