Hello people,
I am in need for testing Marginal Homogeneity for sparse (more then 2X2)
matrixes.
After searching, what I found by now is that for more then 2 by 2 matrixes,
one turns to "stuart maxwell tests" that are available in two
packages:
irr - see: stuart.maxwell.mh
coin - see: mh_test
But I couldn't find in the documentation how valid the results are for
sparse matrixes, and if the validity is indeed compromised (as I suspect it
might be), I wasn't able to find an out-of-the-box alternative function to
use.
A search in google gave me a link to this article:
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V8V-45F5P9B-5&_user=1535420&_rdoc=1&_fmt=&_orig=search&_sort=d&_docanchor=&view=c&_searchStrId=1096547725&_rerunOrigin=google&_acct=C000053610&_version=1&_urlVersion=0&_userid=1535420&md5=604554a739739d90ae72b3a60ce81a00
named "Comparing marginal distributions of large, sparse contingency
tables"
Which sounds good, but the abstract suggests a more complex solution then I
know how to code in my current stage (e.g: "*This article describes
alternative approaches that are much more feasible. We recommend a “pseudo
ML” approach that obtains model parameter estimates by treating repeated
responses as independent and uses a jackknife to estimate the covariance
matrix of those estimates. We test marginal homogeneity using a Wald
statistic, or by adapting the efficient score statistic from the
independent-samples case."*)
This leaves me with the following questions:
1) How can one know how "badly" the sparsity hearts the assumptions of
the
stuart maxwell test?
2) Is there a simple alternative for this test? (and how to use it?)
3) Is there a not-so-simple alternative for this test? (and how to use it?)
Many thanks,
Tal
----------------------------------------------
My contact information:
Tal Galili
E-mail: Tal.Galili@gmail.com
Phone number: 972-52-7275845
FaceBook: Tal Galili
My Blogs:
http://www.talgalili.com (Web and general, Hebrew)
http://www.biostatistics.co.il (Statistics, Hebrew)
http://www.r-statistics.com/ (Statistics,R, English)
[[alternative HTML version deleted]]
