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 on April 23, and its replacement, statalist.org is already up and running.


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

re: Re: st: Paired t-test for propensity match cohort


From   "Ariel Linden, DrPH" <ariel.linden@gmail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   re: Re: st: Paired t-test for propensity match cohort
Date   Wed, 20 Jul 2011 14:14:12 -0700

I suggest that you use -somersd-, -cendif- or -censlope- (all programs
written by the esteemed Roger Newson). These do not require you to
reconfigure your data, they allow you to cluster on the matching variable
(to account for the dependency of the matched pair), and provide confidence
intervals. One stop shopping...

I hope this helps

Ariel

Date: Tue, 19 Jul 2011 09:12:18 +0200
From: Lars Folkestad <lfolkestad@health.sdu.dk>
Subject: Re: st: Paired t-test for propensity match cohort

Hmm
Im not sure.
How about using the matching variable as the grouping variable and the pairs
variable as the id variable in the reshape command:

Reshape wide age var2 var3 var4 ... VarN, i(pairs) j(matching_Variable)

That gives you something like this:
Pair age0 age1
1   53     54
2   55      52
3   55     52
4   53      54

If you have a variable that is supposed to be equal within the pairs eg: Sex
You can use the following command:
Assert sex0==sex1

Hope that helps
lars

Den 17/07/11 18.53 skrev "Sripal Kumar" <sripalkumar@gmail.com> følgende:

> Thanks Lars. I dont think this helps.  In essence, what I need is the
> mean age (and the test for significance) in the matching_Variable 0
> when compared with the mean age in matching_Variable 1, stratified by
> matched pairs (Pairs).  The below example is after propensity score
> matching with pairs 1, 1 forming a pair and so forth.  The simple
> solution would be to create age_0 (age for the matching_Variable 0 for
> pair 1) and age_1 (age for the matching_Variable 1 for pair 1), so
> that the row for pair 1 would have 2 columns for age. However, my
> dataset has around 10,000 cases and the above is not easy to do. Any
> thoughts? I hope I have made myself clear.
> 
> Pairs               matching_Variable              Age
> 1                           0                                    53
> 1                           1                                    54
> 2                           0                                    55
> 2                           1                                    52
> 3                           0                                    55
> 3                           1                                    52
> 4                           0                                    53
> 4                           1                                    54
> 5                           0                                    55
> 5                           1                                    52


*
*   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