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, is already up and running.

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

st: Testing Covariance Balance without Matching

From   Guy Grossman <>
Subject   st: Testing Covariance Balance without Matching
Date   Mon, 10 May 2010 20:08:30 -0400

Dear stats users,

I am using Stata 10.1 to analyze results from a lab experiment in
which subjects were randomly assigned to one of three treatment
conditions (T1, T2, T3). I collected some socio-demographic
information on the participants and I wish to test whether covariates
are balanced across treatment groups (some variable are categorical,
some continuous).

I have previously used pscore and psmatch2 packages (user-written) and
found them quite useful. However, it seems that both these
user-written programs allow to test for balance only after matching.
In this study I am not using matching in my analysis.

My questions are:

1. Does Stata have a command or user-written package to test for
covariate balance by treatment status without first running a Matching
algorithm? I am interested especially in randomization inference (i.e.
permutation-based) balance tests (as Sekhon's Match package in R)

2. If so - is there an efficiently way to export this covariates'
balance test to Latex/ breamer?

3. This is slightly more general question - what covariate balance
test is considered today "best practice"? in poli-sci (my discipline)
researchers usually only present t-test on difference of means, but
this seems unsatisfactory.

Your comments and suggestions would be highly appreciated,
*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index