Topics of the day:
1. Likelihood Ratio-how test it?
2. help on calibrate estimators
3. Silicon analysis of breast and capsular tissue/thanks SA & M
4. Distance Learning in Statistics
5. Sample size estimation for non-parametric tests
6. Fw: VIRUS WARNING
7. biostat graduate programs
8. confirm stat-l
9. Hessian Approximation of the Covariance Matrix
10. Eliminate your phone bill.......
11. the Budweiser Frogs HOAX
12. on usage of Multidimensional Scaling
13. Exact logistic regression
----------------------------------------------------------------------
Date: Sat, 19 Sep 1998 11:08:12 GMT
From: petereikwall@MY-DEJANEWS.COM
Subject: Likelihood Ratio-how test it?
How can I test if a likelihood ratio is significant >1 or/and <>1 ?
Are there any test statistics like a t-test?
P.E
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum
------------------------------
Date: Mon, 21 Sep 1998 10:38:08 -0100
From: Rosignoli Stefano <rosignol@STAT.DS.UNIFI.IT>
Subject: help on calibrate estimators
Hi, I am writing my thesis in survey sampling teory and I ask some
information and some references about calibrate estimation.
Please, if you have same information about them write me.
Stefano Rosignoli
e-mail: rosignol@stat.ds.unifi.it
------------------------------
Date: Mon, 21 Sep 1998 03:50:41 GMT
From: Brian Sandle <bsandle@SOUTHERN.CO.NZ>
Subject: Re: Silicon analysis of breast and capsular tissue/thanks SA & M
SusanS3733 <susans3733@aol.com> wrote:
: In article <906306909.842928@mnementh.southern.co.nz>, Brian Sandle
: <bsandle@southern.co.nz> writes:
:>I have been searching
:>http://www.ncbi.nlm.nih.gov/PubMed/index.html
:>for combinations of such terms as
:>silicon, adverse, adequate, toxicity, selenium, atheroma.
: You might try searching for "silicone"
Sorry, and breast, of course, and maybe liver or brain.
Partial Correlation.
Once a fuel economy test of cars had perplexing result. It turned
out that a manufacturer in one country was using greater tyre
pressures. To find out the real performance of the different motors
it was necessary to partial out the effect of the tyre pressures.
So with silicone breast implants, health and forms of silicon in
the body. It is not certain whether the silicon is from the implant
or from other sources. It is known that it can improve health at
some levels.
"Nonparametric Statistics for the Behavioral Sciences
Siegel, S & Castellan, N J (Second edition) 1988," deals with
partial correlations, including those where the distribution is not
normal, and even gives a significance figure.
I have a VMS program to do it, but have not yet got around to learning
the Fortram NAG linking.
I repeat the abstract you posted. Things are bit more difficult here as
silicon starts to have an opposite effect perhaps at higher levels -
as in lung disease from exposure in the air.
Does anyone on sci.stat.consult have any suggestions? How do breast
implants affect health when the beneficial effects of some (forms
conversion?) silicon are allowed for?
Dodo Dolphin.
TITLE: Silicon analysis of breast and capsular
tissue from
patients with saline or silicone gel breast implants: II. Correlation with
connective- tissue disease.
AUTHOR: Weinzweig J; Schnur PL; McConnell JP; Harris JB; Petty
PM; Moyer TP; Nixon D
AUTHOR AFFILIATION: Department of Plastic Surgery at the Brown
University School of Medicine, Providence, RI 02905, USA.
SOURCE: Plast Reconstr Surg 1998 Jun;101(7):1836-41
CITATION IDS: PMID: 9623824 UI: 98285282
ABSTRACT: The silicone breast implant controversy rages on. Recent
work has demonstrated that normal or baseline breast tissue silicon levels
in women who had had no prior exposure to any type of breast implant may be
as high as 446 microg/gm of tissue. These data ranged from 4 to 446
microg/gm of tissue, with a median of 27.0 microg/gm of tissue. In addition,
numerous other epidemiologic and rheumatologic studies have demonstrated no
association between silicone breast implants and any connective-tissue
diseases. Despite these reports, the use of silicone implants remains
restricted. The present study measured breast and capsular tissue silicon
levels from 23 breasts in 14 patients with saline implants, and from 42
breasts in 29 patients with silicone implants. No patient in the saline
implant group presented with signs or symptoms of connective-tissue disease.
Patients with silicone implants, however, were divided into three groups
based on the presence or absence of signs or symptoms of connective-tissue
disease: group I, no symptoms or signs; group II, + symptoms, no signs; and
group III, + symptoms, + signs. Six patients in group III were diagnosed
with a specific connective-tissue disease, including systemic lupus
erythematosus, rheumatoid arthritis, or scleroderma. The most common
indications for implant removal or exchange were capsular contracture and
implant rupture, although 41 percent of patients with silicone implants
expressed media-related concern over the implant issue. The most common
symptoms described by patients in groups II and III were joint pain and
stiffness, arm pain and numbness, and fatigue. In all groups, capsular
tissue silicon levels were significantly greater than breast tissue levels.
This finding may indicate that the capsule serves as a barrier to the
distribution of silicone from the implant into adjacent breast tissue.
Although breast tissue silicon levels in patients with silicone implants
were not significantly greater than those in patients with saline implants
(p = 0.48), capsular tissue levels in patients with silicone implants were,
indeed, significantly greater than those in patients with saline implants (p
< 0.001). However, no statistically significant differences in tissue
silicon levels were observed with relation to the presence or absence of
connective-tissue disease signs or symptoms in patients with silicone
implants (groups I to III). Therefore, these data strengthen the conclusion
that there is no association between tissue silicon levels and
connective-tissue disease.
MAIN MESH HEADINGS: Breast/*chemistry
*Breast Implants
Connective Tissue Diseases/*metabolism
Silicon/*analysis
*Silicones
*Sodium Chloride
ADDITIONAL MESH HEADINGS: Breast Implants/adverse effects
Connective Tissue Diseases/etiology
Female
Human
Prosthesis Failure
Spectrum Analysis
PUBLICATION TYPES: JOURNAL ARTICLE
REGISTRY NUMBERS: 0 (Silicones)
7440-21-3 (Silicon)
7647-14-5 (Sodium Chloride)
LANGUAGE: Eng
------------------------------
Date: Mon, 21 Sep 1998 15:44:24 +0200
From: Haiko Luepsen <Luepsen@RRZ.UNI-KOELN.DE>
Subject: Re: Distance Learning in Statistics
>
> Can someone point me to an article, website, etc.
> which addresses what's been and being done in distance
> learning of statistics courses. I suppose this includes
> anything from following a book and taking exams to
> participating in live (chat) virtual classrooms.
> Are organizations like ASA or RSS active in this
> field?
>
> Thanks, Tim O'Brien
>
A list of links to such resources you will find on
http://www.uni-koeln.de/themen/Statistik/onlinebooks.html
Another good list offers Mike Talbot:
http://www.bioss.sari.ac.uk/~mike/webtra.htm
Haiko Luepsen
University of Cologne, Computing Centre
Germany
luepsen@rrz.uni-koeln.de
------------------------------
Date: Mon, 21 Sep 1998 16:03:13 +0100
From: PETER HUDSON <PETER.HUDSON@WMH-TR.WALES.NHS.UK>
Subject: Sample size estimation for non-parametric tests
Could anyone give me any pointers to methods for sample size estimation
for non-parametric tests? We have results of a standard questionnaire
on a set of patients which shows a skewed distribution. Before a trial
of a treatment regime which aims to improve scores, we need to estimate
how many new patients and controls to recruit. The usual calculations
suggest a sample size of 100 per group when applied to the untransformed
data and 10 when applied to the log transformed data. I'm looking for a
'3rd way' to improve my confidence in the figures!
Peter Hudson
Principal Biochemist
Wrexham Maelor Hospital
Wales, UK
peter.hudson@wmh-tr.wales.nhs.uk
------------------------------
Date: Mon, 21 Sep 1998 16:29:00 +0100
From: Debbie Law <dlaw@BRIDGEWAY.CO.UK>
Subject: Fw: VIRUS WARNING
----------
From: Stephen McClune
To: DLAW; "nt-uk"; "windows95-uk"
Subject: Fw: VIRUS WARNING
Date: 21 September 1998 16:23
As Ian says, this warning might be useful
Regards
Stephen
-----
Stephen McClune Phone: +44 (0)1232 323418
GCAS Advertising Ltd Fax: +44 (0)1232 245741
Belfast
-----Original Message-----
From: Baillie, Ian <ian.baillie@stenaline.com>
To: 'Jeff Knox - University of Ulster' <JW.Knox@ulst.ac.uk>; 'Philip
McCullagh' <philip.mc@virgin.net>; 'Jonathan Smith'
<jonathan.smith.deni@nics.gov.uk>; 'GCAS Advertising (Studio)'
<Studio@gcasgroup.com>
Date: 21 September 1998 10:43
Subject: FW: VIRUS WARNING
>Thought you might find this warning useful.
>
>Regards
>Ian
>
>
>>----------
>>From: S=F6dermyr, Anders
>>Sent: 21 September 1998 07:19
>>To: *All Mail Users
>>Subject: VB: VIRUS WARNING
>>
>>
>>
>>>----------
>>>Fr=E5n: Thesslund, Peter
>>>Skickat: den 20 september 1998 08:22
>>>Till: *All Mail Users; Christensen, Elisabeth
>>>Ang=E5ende: SV: VIRUS WARNING
>>>
>>>
>>>
>>>----------
>>>Fr=E5n: Christensen, Elisabeth
>>>Skickat: den 16 september 1998 11:44
>>>Till: *All Mail Users
>>>Ang=E5ende: VIRUS WARNING
>>>
>>>
>>>
>>>> Someone is sending out a screensaver : the Budweiser Frogs. But if you
>>>
>>>> download it you will lose everything!!!!
>>>
>>>> Your hard drive will crash!!! DON'T DOWNLOAD THIS UNDER ANY
>>>
>>>> CIRCUMSTANCES!! IT WENT INTO CIRCULATION ON THE 09/09/98, DISTRIBUTE
>>>
>>>> THIS TO AS MANY PEOPLE AS POSSIBLE.
>>>
>>>> If you download, some jerk from the internet will get your screen name
>>>
>>>> and password! Please send this to any names you can think of and
>>>
>>>> remember never download BUDDYLST.ZIP. This is a new, very malicious
>>>
>>>> virus and not many people know about it. This information was announce=
d
>>>
>>>> yesterday morning from Microsoft. Please share it with everyone that
>>>
>>>> might access the internet.
>>>
>>>> Once again, pass along to everyone in your address book so that this
may
>>>
>>>> be stopped. Also do not open or even look at any mail that says
>>>
>>>> "RETURNED OR UNABLE TO DELIVER". This virus will attach itself to your
>>>
>>>> computer components and render them useless. Immediately delete and
mail
>>>
>>>> items that says this. AOL has said that this is a very dangerous virus
>>>
>>>> and that there is NO remedy for it at this time. Please practice
>>>
>>>> cautionary measures and forward this to all on line friends ASAP.
>>>
>>>
>>>
>>>--
>>>
>>>Fiona MacKenzie
>>>
>>>
>>>
>>>Africa in Style Safari Operator
>>>
>>>
>>>
>>>T: +27+11+8848200
>>>
>>>
>>>
>>>F: +27+11+8848201
>>>
>>>
>>>
>>>http://www.africa-in-style.co.za
>>>
>>>
>>
>
------------------------------
Date: Sat, 19 Sep 1998 17:44:55 GMT
From: danwong@MY-DEJANEWS.COM
Subject: biostat graduate programs
Hi, I am wondering if anyone knows the ranking for biostat graduate programs,
or anywhere I can look for it?
Thanks.
Daniel
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum
------------------------------
Date: Mon, 21 Sep 1998 10:48:50 -0500
From: "Babinec, Tony" <tony@SPSS.COM>
Subject: confirm stat-l
CONFIRM STAT-L
------------------------------
Date: Mon, 21 Sep 1998 09:37:01 +0100
From: Carlos Romero <c.romero@IC.AC.UK>
Subject: Hessian Approximation of the Covariance Matrix
The inverse of the matrix of partial derivatives of the log-likelihood
with respect to the parameters (Hessian) is used as an approximation to
the negative of the asymptotic variance matrix of the parameters. My
objective is to sample from a multivariate normal distribution using the
Cholesky decomposition.
Can anybody tell me why the Hessian approximation is valid and under
what conditions?, or at least, suggest me a book to look at (different
than Arnold 1990 which I can not find)
Any help would be highly appreciated,
Carlos.
------------------------------
Date: Mon, 21 Sep 1998 12:12:03 EST
From: livecoverage@MAILEXCITE.COM
Subject: Eliminate your phone bill.......
A FREE Nokia 6160 digital PCS phone that covers the 50 United States
including Alaska and Hawaii and 600 FREE minutes of airtime utilizing
the AT&T wireless network with No roaming or Long Distance charges!
Eliminate your current phone bill forever as you continue to receive
free airtime.
Nokia 6160 digital PCS phone sent via Federal Express direct from AT&T.
Listen to our 3 Minute recorded Presentation regarding this Free Nokia Phone
offer at 1-888-834-5017 or 1-888-248-6031
Pull up the Fax-on-Demand at 1-716-720-2299
Not only can you get a FREE Nokia 6160 phone, you can receive residual
income that allows the sky to be the limit as far as your future income is
concerned.
Need more information:
listen to our 24hr recorded conference call at 1-212-796-6870
To get started after listening to the 3 minute message, you will need the
following information:
Sponsor ID: 100108281
Sponsor Phone: 1-800-636-6773 ext 4533
Let me know if you decide to join so I can help you receive your FREE phone
quick.
Good Luck.
This mailing has been forwarded to you as a person interested in
seeking additional sources of income. If this reached you by error,
please hit the Reply button and write Remove in the Subject line.
We sincerely apologize for the possible inconvenience.
This message is sent in compliance of the new e-mail bill: SECTION
301. Per Section 301, Paragraph (a)(2)(C) of S. 1618,
------------------------------
Date: Mon, 21 Sep 1998 12:48:04 -0400
From: "Wuensch, Karl L." <WUENSCHK@MAIL.ECU.EDU>
Subject: the Budweiser Frogs HOAX
Debbie Law [SMTP:dlaw@BRIDGEWAY.CO.UK]
<mailto:[SMTP:dlaw@BRIDGEWAY.CO.UK]> sent us a copy of the Budwiser Frogs
urban legend. Probably most STAT-L subscribers already know that this is a
hoax, that the virus resides not in any a screen saver but rather the hoax
letter itself when sent to others, especially to a listserv list -- that is,
when YOU forward mail like this, you are a virus.
+++++++++++++++++++++++++++
Karl L. Wuensch, Department of Psychology,
East Carolina University, Greenville NC 27858-4353
Voice: 252-328-4102 Fax: 252-328-6283
WUENSCHK@mail.ecu.edu <mailto:WUENSCHK@mail.ecu.edu>
http://ecuvax.cis.ecu.edu/~pswuensc/klw.htm
<http://ecuvax.cis.ecu.edu/~pswuensc/klw.htm>
------------------------------
Date: Mon, 21 Sep 1998 17:11:40 +0100
From: Raphael HALIMI <raphael.halimi@GEP.FR>
Subject: on usage of Multidimensional Scaling
I' ve just come to the use of Multidimensional Scaling, with one article
of Stephen P.Borgatti.
1- I have a matrix of non-euclidian distances (or dissimilarities);
2- I want to cluster my population;
3- I think that any factor analysis or clustering need a matrix of
euclidian distances;
4- Therfore I entend to use MDS to transform non-euclidian distances
into euclidian ones to do the clustering.
Is the 3rd point correct? What do you think of this method ? It seems
this usage of MDS is really far from its purpose.
Thanks
------------------------------
Date: Tue, 22 Sep 1998 10:13:49 +1000
From: Mahshid Nasseri <mahshid@GIZMO.MATHS.MONASH.EDU.AU>
Subject: Exact logistic regression
Hi,
I would like to find out if there is any difference between the Exact
Logistic Regression and the Logistic Regression. If so, how can I
find out more about the Exact method? Can anyone please help me?
Mahshid
------------------------------
End of STAT-L Digest - 20 Sep 1998 to 21 Sep 1998
*************************************************