Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: N to 1 matching without replacement


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: N to 1 matching without replacement
Date   Wed, 5 May 2010 13:18:58 -0400

Russell Callaghan <Russell_Callaghan@camh.net> :
I take it back--it looks like  -psmatch2- (on SSC) will not do 1 to
many matching w/o replacement but issues the error:
"Matching without replacement is only implemented with 1-to-1
propensity score matching"

I guess the issue with matching N cases without replacement is that
the order of the matches plays a much bigger role in how well the
matches approximate a overall distance-minimizing choice... though
that seems an issue of magnitude not of type (of potential erroneous
matching).

Why don't you just reweight instead of matching?  Easier and more defensible...

On Wed, May 5, 2010 at 12:53 PM, Austin Nichols <austinnichols@gmail.com> wrote:
> Russell Callaghan <Russell_Callaghan@camh.net> :
> -psmatch2- (on SSC) has a noreplacement option; but once you have
> matched n cases onto each of the treatment cases, you can only see the
> id of the first match in the _n1 variable.  I guess you could hack
> into the code, at places like
>
>        st_view(N1,.,"_n1")
>
> to instead create a set of N match id variables.  Make sure you save
> the resulting bastardized program as my_psmatch.ado or somesuch.
>
> On Wed, May 5, 2010 at 12:35 PM, Russell Callaghan
> <Russell_Callaghan@camh.net> wrote:
>> Hello,
>> I want to use psmatch2 (or any other STATA propensity score matching
>> algorithm) to perform an N to 1 matching without replacement. I have
>> looked at the STATAlist extensively and have found no solution to my
>> problem. Although psmatch2 provides 1 to n matching with replacement, I
>> want to do matching without replacement so that I can then do a Cox
>> regression which accounts for the correlated structure of the data by
>> using the VCE(cluster paired) command. Any thoughts about the way ahead?
>> Thanks,
>> Russ
>>
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index