Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

Re: st: Propensity score matching and multiple imputation

From (Brendan Halpin)
Subject   Re: st: Propensity score matching and multiple imputation
Date   Thu, 17 May 2012 13:07:26 +0100

On Thu, May 17 2012, Michela Coppola wrote:

> Now the combined standard error using the Rubin's rule is smaller than
> the standard errors in each of the 5 implicates! I expected bigger
> standard errors not smaller ones (as the variability introduced through
> the imputation is now taken into account).
> How is it that possible (I checked several times the formulas, so I'm
> sure that there are no errors when combining in excel the standard
> errors)?

If I understand the relevant formula correctly (see e.g., p237 of
Patrick Royston's 2004 Stata Journal article) the variance of the
imputation is the average of the variance of the imputed data sets (the
"within" part), plus the "between" part, which is strictly non-negative.

I suspect Excel is not doing what you think it is doing. And given
the nature of Excel, it is a little hard for you to show us what it is

Show us the figures, perhaps?

Brendan Halpin,   Department of Sociology,   University of Limerick,   Ireland
Tel: w +353-61-213147  f +353-61-202569  h +353-61-338562;  Room F1-009 x 3147    ULSociology on Facebook:         twitter:@ULSociology
*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index