Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: factor variables and qreg?
From
"Roger B. Newson" <[email protected]>
To
[email protected]
Subject
Re: st: factor variables and qreg?
Date
Fri, 24 Aug 2012 20:27:40 +0100
Yes, I do mean robust variances, assuming that, by "robust variances",
you mean "Huber variances". No statistical method is just "robust",
because there are tradeoffs, depending on what we wanted the variances
to be "robust" to. The traditional Wedderburn quasi-likelihood variance,
for example, was designed to be "robust" to tiny subsample numbers, as
in comparing a large number of groups of 3, which people frequently do,
in Rothamsted and elsewhere.
I think the existing -qreg- variances are already Huber variances, but
currently they seem only to allow sampling-probability weighting, not
clustering.
Best wishes
Roger
Roger B Newson BSc MSc DPhil
Lecturer in Medical Statistics
Respiratory Epidemiology and Public Health Group
National Heart and Lung Institute
Imperial College London
Royal Brompton Campus
Room 33, Emmanuel Kaye Building
1B Manresa Road
London SW3 6LR
UNITED KINGDOM
Tel: +44 (0)20 7352 8121 ext 3381
Fax: +44 (0)20 7351 8322
Email: [email protected]
Web page: http://www.imperial.ac.uk/nhli/r.newson/
Departmental Web page:
http://www1.imperial.ac.uk/medicine/about/divisions/nhli/respiration/popgenetics/reph/
Opinions expressed are those of the author, not of the institution.
On 24/08/2012 20:03, Feiveson, Alan H. (JSC-SK311) wrote:
Roger - Hope you're correct. I assume you mean robust variance estimation and not some sort of random-effects model - although I would like to see that also.
Al
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Roger B. Newson
Sent: Friday, August 24, 2012 1:59 PM
To: [email protected]
Subject: Re: st: factor variables and qreg?
Thanks to Bill for this information. Would I be right in guessing that
StataCorp are also working on clustered variances for -qreg-?
Best wishes
Roger
Roger B Newson BSc MSc DPhil
Lecturer in Medical Statistics
Respiratory Epidemiology and Public Health Group
National Heart and Lung Institute
Imperial College London
Royal Brompton Campus
Room 33, Emmanuel Kaye Building
1B Manresa Road
London SW3 6LR
UNITED KINGDOM
Tel: +44 (0)20 7352 8121 ext 3381
Fax: +44 (0)20 7351 8322
Email: [email protected]
Web page: http://www.imperial.ac.uk/nhli/r.newson/
Departmental Web page:
http://www1.imperial.ac.uk/medicine/about/divisions/nhli/respiration/popgenetics/reph/
Opinions expressed are those of the author, not of the institution.
On 24/08/2012 19:01, William Gould, StataCorp LP wrote:
Ariel Linden <[email protected]> wrote,
I wonder if any StataCorp people on the list could respond to this
query [why qreg does not allow factor variables] with a definitive
answer?
We at StataCOrp are working on making -qreg- with factor variables.
Maarten Buis <[email protected]> speculated
-qreg- seems to have its own build-in maximization algorithm, which
does not surprise me as it is a maximization problem with its own
peculiar set of difficulties. So it is a lot more work to implement
it.
Maartin is correct. We're working on it now, however, and have been
working on it since before this thread arose.
-- Bill
[email protected]
*
* 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/
*
* 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/
*
* 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/
*
* 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/