Paul,

I assume you mean the joint distribution of the estimated between-imputation 
and within-imputation variance components as opposed to variances on Bayesian 
posterior distributions or between-PSU and within-PSU variance components.

Seems to me that the distribution of the estimated within-imputation variance 
component depends on the variance estimation method.  Kirk Wolter's book (An 
Introduction to Variance Estimation) is a good source for information on 
various variance estimation methods.  Also see chapter 4 of Wayne Fuller's 
newer text (Sampling Statistics).

One way to discover more on this topic is to search for "variance of the 
variance."  Here are a few relevant links.

http://www.amstat.org/sections/srms/proceedings/papers/1989_030.pdf
http://www.amstat.org/sections/srms/Proceedings/y2004/files/Jsm2004-000354.pdf
http://www2.math.su.se/~rolfs/Publikationer/Biometrika94.pdf


--Dave Judkins
Abt Associates

From: Impute -- Imputations in Data Analysis 
[mailto:[email protected]] On Behalf Of Paul von Hippel
Sent: Tuesday, October 02, 2012 8:23 AM
To: [email protected]
Subject: Distribution of variance components

Is there any work on the joint distribution of the Between and Within variance 
components that are used in calculating the usual multiple imputation standard 
error? The Between component is Wishart, that's pretty clear, but the marginal 
distribution of the Within component is trickier, and then there's the joint 
distribution.

--
Best wishes,
Paul von Hippel
Assistant Professor
LBJ School of Public Affairs
Sid Richardson Hall 3.251
University of Texas, Austin
2315 Red River, Box Y
Austin, TX  78712
(512) 537-8112


________________________________
This message may contain privileged and confidential information intended 
solely for the addressee. Please do not read, disseminate or copy it unless you 
are the intended recipient. If this message has been received in error, we 
kindly ask that you notify the sender immediately by return email and delete 
all copies of the message from your system.

Reply via email to