Statisticians_group : Messages : 2878-2907 of 4104

this is virus dont use it

wahjuda


--- mgsayyad@... wrote:

> Mail transaction failed. Partial message is
> available.
>
>

Reply | Forward | Messages in this Topic (3)

#2906

From: mgsayyad@...
Date:: Mon Sep 29, 2008 3:18 am
Subject:: Hello

mgsayyad@...

Mail transaction failed. Partial message is available.

Reply | Forward | Messages in this Topic (3)

#2905

From: "A Kemal Erdemoglu" <akerdemoglu@...>
Date:: Sun Sep 28, 2008 3:33 pm
Subject:: Reliability Analyses

akerdemoglu

Online

Dear Friends,
I need your help for reliability analyses of a questionnaire.
As I read it from the book or reliability sheets, it made up new
questions about the test.
I have 7 questions for the 100 patients. How will I use the data? Will
it be 100 or 7 item in the reliabilty analyses?
How will I design it? I designed as 7 variable and 100 cases in SPSS
format. Is it correct or do I have to transpose the variable with cases?
Waiting your emails
AKE

#2904

From: farah dhiab <farahdhiab@...>
Date:: Sun Sep 28, 2008 1:44 am
Subject:: Re: time series

farahdhiab

Offline

Hi, try this http://www.statsoft.com/textbook/sttimser.html#1general

Regards,

--- On Tue, 9/23/08, hilal a�fffffe�fffffdk <asikhilal@...> wrote:

From: hilal a�fffffe�fffffdk <asikhilal@...>
Subject: [Statisticians_group] time series
To: statisticians_group@...
Date: Tuesday, September 23, 2008, 10:31 AM

Hi all,
I need help about application of time series in biostatistics. I have data about canser between the years 1980-2008. and I hope to find a significant change in some years. and for this data � want to use spss packet program. bu I do not know how.
thanks

#2903

From: "gaetanlion" <gaetanlion@...>
Date:: Thu Sep 25, 2008 11:31 pm
Subject:: Re: Sample size

gaetanlion

Offline

--- In Statisticians_group@..., prashant rai
<pk_stats@...> wrote:
>
> Hi
>
> How to determine about the significant sample size for developing a
logistic regression model.
> What if the one category of dependent variable has few observations.
> Best wishes
> Prashant
>
>
>       Unlimited freedom, unlimited storage. Get it now, on
http://help.yahoo.com/l/in/yahoo/mail/yahoomail/tools/tools-08.html/
>

Obviously, you are limited by your dependent variable's number of
observations.  Given that, go ahead with what you've got.  And, check
the output of your logit regression.  Watch for the p values
associated with your independent variables' regression coefficients.
If they are less than 0.33, I would say that's pretty good.  If one
jumps out above 0.50 but the other variables have low p values, you
can rerun your model by taking this one suspiciously weak independent
variable.

Logit regression does not have a specific sample size threshold to
work.  Typically, if your sample size is greater than 10 or 15 and
your independent variables are strong, the model can work just fine.

Hope that helps.

Guy

Reply | Forward | Messages in this Topic (36)

#2902

From: prashant rai <pk_stats@...>
Date:: Thu Sep 25, 2008 12:15 pm
Subject:: Sample size

pk_stats

Offline

Hi

How to determine about the significant sample size for developing a logistic
regression model.
What if the one category of dependent variable has few observations.
Best wishes
Prashant


       Unlimited freedom, unlimited storage. Get it now, on
http://help.yahoo.com/l/in/yahoo/mail/yahoomail/tools/tools-08.html/

Reply | Forward | Messages in this Topic (36)

#2901

From: farah dhiab <farahdhiab@...>
Date:: Wed Sep 24, 2008 11:19 pm
Subject:: Re: A DOE Question

farahdhiab

Offline

Hi,

If you are talking about Mr.Pradipta problem. Yes Spss files will do.

Regards

--- On Wed, 9/24/08, kishore reddy <kishore_pingili@...> wrote:

From: kishore reddy <kishore_pingili@...>
Subject: Re: [Statisticians_group] A DOE Question
To: Statisticians_group@...
Date: Wednesday, September 24, 2008, 8:15 AM

Hi
better to send SPSS File

   Rgds,
kshr

--- On Tue, 23/9/08, Zidan M. Houmani <houmani67@yahoo. com> wrote:

From: Zidan M. Houmani <houmani67@yahoo. com>
Subject: Re: [Statisticians_ group] A DOE Question
To: Statisticians_ group@yahoogroup s.co.in
Date: Tuesday, 23 September, 2008, 1:41 PM

i have all of my results data in SPSS FILES and word files which one you want me to send to you

You�re truly

Mr.Zidan Mohamed Houmani

Nuclear Science Building
School of Applied Physics
Faculty of Science and Technology
University of Kebangsaan Malaysia
43600 UKM Bangi Selangor D.E,
Malaysia.

Tel (H): 00603-87344616
H/P    : 006016-3103526
E-mail: houmani67@yahoo. com & houmani67@gmail. com

--- On Thu, 9/18/08, farah dhiab <farahdhiab@yahoo. com> wrote:

From: farah dhiab <farahdhiab@yahoo. com>
Subject: Re: [Statisticians_ group] A DOE Question
To: Statisticians_ group@yahoogroup s.co.in
Date: Thursday, September 18, 2008, 3:17 PM

I have some questions considering your experiment, how many levels did u set for each variable x? And can u send your design plan so I can have better idea of your experiment?
Also, have you thought of using robust parameter design?? If you think that idea might help u to get better optimum results, email me back and we will discuss it in focus.
Regards,
Farah

--- On Thu, 9/18/08, Pradipta Sarkar <sarkarp2002@ yahoo.com> wrote:

From: Pradipta Sarkar <sarkarp2002@ yahoo.com>
Subject: [Statisticians_ group] A DOE Question
To: Statisticians_ group@yahoogroup s.co.in
Date: Thursday, September 18, 2008, 6:49 AM

I have a real life problem .... would appreciate any help you can provide.

Let us say we have k variables (which are ingredients of a product, k is about 20). We want to study some technical response Y as a function of the k variables X_1, X_2, ..., X_k. The variable X_i can take value between 0 and a_i

So we can create a response surface design using some optimality criteria such as D-Optimality or I-Optimality. Ater running the experiment and collecting data we can fit the model and then find the combination of X_1, X_2, ..., X_k that maximizes the response Y. For this specific application (involving this particular product) usually the maximum is attained at a point where several of the X_1, X_2, ..., X_k variables are 0 which means it is on the boundary of the design space.

We know that the confidence interval of prediction is very wide at the boundary (no matter what optimality criteria we use), so the confidence interval on the maximum reponse is likely to be very wide.

So here is my question: Is there a way to design an experiment so that "it has lower prediction variance at the boundary"? My intuition says that while designing the experiment put more points at the corners and sides of the design space and less inside the design space. Is that a good strategy or can you suggest another better one? Won't that destroy some desired properties of DOE? What will I lose? Is there a standard way to handle such design of experiments?

Remember that the X variables are ingredients, so the minimum value is 0(cannot be negative).

Thanks,

Pradipta.

Unlimited freedom, unlimited storage. Get it now

#2900

From: KAUSHIK BHATTACHARJEE <kabonline07@...>
Date:: Wed Sep 24, 2008 10:55 am
Subject:: filtering problem in Time Series

kabonline07

Offline

Hi All,

I posted this problem earlier...did not get any response. Hope this time I will get some .

The description of the problem is as follows:

X_t=f(Y_t ,Z_t) +random error term

Xt is observable but Y_t and Z_t are not observable-however both are rvs ;i.e. I have a time series of X_ts only. Is there any(filtering) technique you know/think/used to disentangle the effect of Y_t and Z_t. Here t=1 day. Observations are supposed to be taken at equal intervals like 1 day etc.In other words if on a given day �a*Xt � is due to Yand �(1-a)*Xt� is due to Z-is there any procedure to estimate 'a'?

Please write anything you think might prove to be useful. Even if it is an one liner e.g. �try using Kalman Filter/spectral analysis�---pls do.

Just for your info-When it boils down to modeling, we may expect that Y_t possesses a time trend but Z_t does not.

Kaushik Bhattacharjee

#2899

From: hilal a�fffffe�fffffdk <asikhilal@...>
Date:: Wed Sep 24, 2008 8:35 am
Subject:: power of chi-square test

asikhilal

Offline

MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

Hi,=0Ais there a=A0computer program or a practical way of calculating power=
  of chi-quare test or any other test. is the program spss compute power of =
tests like chi-square,shapiro wilk or....=0A=0A=0A

#2898

From: hilal a�fffffe�fffffdk <asikhilal@...>
Date:: Wed Sep 24, 2008 6:57 am
Subject:: Re: time series

asikhilal

Offline

Hi ,

My real prolem is that I dont have background about time series. from my university education I remember that time series is used in datas about economy but I forget the important details and how I can use it in biostatistics.

----- Original Message ----
From: farah dhiab <farahdhiab@...>
To: Statisticians_group@...
Sent: Wednesday, September 24, 2008 1:00:04 AM
Subject: Re: [Statisticians_group] time series

Hi,

In spss tool bar, you chose analyze then time series box.

If you have background in time series analysis then you can continue by your own, else i would be glad to help you.

Farah

--- On Tue, 9/23/08, hilal a�fffffe�fffffdk <asikhilal@...> wrote:

From: hilal a�fffffe�fffffdk <asikhilal@...>
Subject: [Statisticians_group] time series
To: statisticians_group@...
Date: Tuesday, September 23, 2008, 10:31 AM

Hi all,
I need help about application of time series in biostatistics. I have data about canser between the years 1980-2008. and I hope to find a significant change in some years. and for this data � want to use spss packet program. bu I do not know how.
thanks

#2897

From: kishore reddy <kishore_pingili@...>
Date:: Wed Sep 24, 2008 5:15 am
Subject:: Re: A DOE Question

kishore_pingili

Offline

Hi
better to send SPSS File

Rgds,
kshr

--- On Tue, 23/9/08, Zidan M. Houmani <houmani67@...> wrote:

From: Zidan M. Houmani <houmani67@...>
Subject: Re: [Statisticians_group] A DOE Question
To: Statisticians_group@...
Date: Tuesday, 23 September, 2008, 1:41 PM

i have all of my results data in SPSS FILES and word files which one you want me to send to you

You’re truly

Mr.Zidan Mohamed Houmani

Nuclear Science Building
School of Applied Physics
Faculty of Science and Technology
University of Kebangsaan Malaysia
43600 UKM Bangi Selangor D.E,
Malaysia.

Tel (H): 00603-87344616
H/P : 006016-3103526
E-mail: houmani67@yahoo. com & houmani67@gmail. com

--- On Thu, 9/18/08, farah dhiab <farahdhiab@yahoo. com> wrote:
From: farah dhiab <farahdhiab@yahoo. com>
Subject: Re: [Statisticians_ group] A DOE Question
To: Statisticians_ group@yahoogroup s.co.in
Date: Thursday, September 18, 2008, 3:17 PM

I have some questions considering your experiment, how many levels did u set for each variable x? And can u send your design plan so I can have better idea of your experiment?
Also, have you thought of using robust parameter design?? If you think that idea might help u to get better optimum results, email me back and we will discuss it in focus.
Regards,
Farah

--- On Thu, 9/18/08, Pradipta Sarkar <sarkarp2002@ yahoo.com> wrote:

From: Pradipta Sarkar <sarkarp2002@ yahoo.com>
Subject: [Statisticians_ group] A DOE Question
To: Statisticians_ group@yahoogroup s.co.in
Date: Thursday, September 18, 2008, 6:49 AM

I have a real life problem .... would appreciate any help you can provide.

Let us say we have k variables (which are ingredients of a product, k is about 20). We want to study some technical response Y as a function of the k variables X_1, X_2, ..., X_k. The variable X_i can take value between 0 and a_i

So we can create a response surface design using some optimality criteria such as D-Optimality or I-Optimality. Ater running the experiment and collecting data we can fit the model and then find the combination of X_1, X_2, ..., X_k that maximizes the response Y. For this specific application (involving this particular product) usually the maximum is attained at a point where several of the X_1, X_2, ..., X_k variables are 0 which means it is on the boundary of the design space.

We know that the confidence interval of prediction is very wide at the boundary (no matter what optimality criteria we use), so the confidence interval on the maximum reponse is likely to be very wide.

So here is my question: Is there a way to design an experiment so that "it has lower prediction variance at the boundary"? My intuition says that while designing the experiment put more points at the corners and sides of the design space and less inside the design space. Is that a good strategy or can you suggest another better one? Won't that destroy some desired properties of DOE? What will I lose? Is there a standard way to handle such design of experiments?

Remember that the X variables are ingredients, so the minimum value is 0(cannot be negative).

Thanks,

Pradipta.

Unlimited freedom, unlimited storage. Get it now

#2896

From: farah dhiab <farahdhiab@...>
Date:: Tue Sep 23, 2008 10:00 pm
Subject:: Re: time series

farahdhiab

Offline

Hi,

In spss tool bar, you chose analyze then time series box.

If you have background in time series analysis then you can continue by your own, else i would be glad to help you.

Farah

--- On Tue, 9/23/08, hilal a�fffffe�fffffdk <asikhilal@...> wrote:

From: hilal a�fffffe�fffffdk <asikhilal@...>
Subject: [Statisticians_group] time series
To: statisticians_group@...
Date: Tuesday, September 23, 2008, 10:31 AM

Hi all,
I need help about application of time series in biostatistics. I have data about canser between the years 1980-2008. and I hope to find a significant change in some years. and for this data � want to use spss packet program. bu I do not know how.
thanks

#2895

From: "Zidan M. Houmani" <houmani67@...>
Date:: Tue Sep 23, 2008 8:11 am
Subject:: Re: A DOE Question

houmani67

Offline

i have all of my results data in SPSS FILES and word files which one you want me to send to you

You�re truly

Mr.Zidan Mohamed Houmani

Nuclear Science Building
School of Applied Physics
Faculty of Science and Technology
University of Kebangsaan Malaysia
43600 UKM Bangi Selangor D.E,
Malaysia.

Tel (H): 00603-87344616
H/P : 006016-3103526
E-mail: houmani67@... & houmani67@...

--- On Thu, 9/18/08, farah dhiab <farahdhiab@...> wrote:

From: farah dhiab <farahdhiab@...>
Subject: Re: [Statisticians_group] A DOE Question
To: Statisticians_group@...
Date: Thursday, September 18, 2008, 3:17 PM

I have some questions considering your experiment, how many levels did u set for each variable x? And can u send your design plan so I can have better idea of your experiment?
Also, have you thought of using robust parameter design?? If you think that idea might help u to get better optimum results, email me back and we will discuss it in focus.
Regards,
Farah

--- On Thu, 9/18/08, Pradipta Sarkar <sarkarp2002@ yahoo.com> wrote:

From: Pradipta Sarkar <sarkarp2002@ yahoo.com>
Subject: [Statisticians_ group] A DOE Question
To: Statisticians_ group@yahoogroup s.co.in
Date: Thursday, September 18, 2008, 6:49 AM

I have a real life problem .... would appreciate any help you can provide.

Let us say we have k variables (which are ingredients of a product, k is about 20). We want to study some technical response Y as a function of the k variables X_1, X_2, ..., X_k. The variable X_i can take value between 0 and a_i

So we can create a response surface design using some optimality criteria such as D-Optimality or I-Optimality. Ater running the experiment and collecting data we can fit the model and then find the combination of X_1, X_2, ..., X_k that maximizes the response Y. For this specific application (involving this particular product) usually the maximum is attained at a point where several of the X_1, X_2, ..., X_k variables are 0 which means it is on the boundary of the design space.

We know that the confidence interval of prediction is very wide at the boundary (no matter what optimality criteria we use), so the confidence interval on the maximum reponse is likely to be very wide.

So here is my question: Is there a way to design an experiment so that "it has lower prediction variance at the boundary"? My intuition says that while designing the experiment put more points at the corners and sides of the design space and less inside the design space. Is that a good strategy or can you suggest another better one? Won't that destroy some desired properties of DOE? What will I lose? Is there a standard way to handle such design of experiments?

Remember that the X variables are ingredients, so the minimum value is 0(cannot be negative).

Thanks,

Pradipta.

#2894

From: hilal aÿfffffeÿfffffdk <asikhilal@...>
Date:: Tue Sep 23, 2008 7:31 am
Subject:: time series

asikhilal

Offline

Hi all,
 I need help about application of time series in biostatistics. I have data
about canser between the years 1980-2008. and I hope to find a significant
change in some years. and for this data ı want to use spss packet program. bu I
do not know how.
thanks

#2893

From: Hemant Trivedi <trivedihemant69@...>
Date:: Mon Sep 22, 2008 2:58 pm
Subject:: (No subject)

trivedihemant69

Offline

Dear All,

I need help on Probit Analysis from its basics/fundamentals to advance level. If any one has any stuff related to this kindly let me know.

Hemant Trivedi

#2892

From: "mchulet" <mchulet@...>
Date:: Mon Sep 22, 2008 2:06 pm
Subject:: T Distribution calculator

mchulet

Offline

Hi,
   Need you help, I need to implement regression analysis with a
hypothesis test of non negative slope. For this I am using t-test but
the problem is that to find the Critical p-value in Java code I need to
implement T-Distribution calculator similar to the one that can be
found at http://stattrek.com/Tables/T.aspx . Can someone please tell me
how mathematically this is derived or point me to a resource that
details this out

Greatly appreciate your help

-Thanks

#2891

From: Scott <bigmagumba@...>
Date:: Mon Sep 22, 2008 2:01 am
Subject:: Re: Urgent! How to collect the white noise from a panel data regression in STATA

bigmagumba

Offline

Hi Alexis,

This is a new one for me. Not sure I've ever heard it put this way.

So maybe this will help (?) ...

In a perfect regression exercise, theoretically, one explains variance and the residual that occurs between observed and estimated values is random error (white noise?)

Of course, there is no such thing as a perfect regression.

Scott

--- On Sun, 9/21/08, Siwaig <asym001@...> wrote:

From: Siwaig <asym001@...>
Subject: [Statisticians_group] Urgent! How to collect the white noise from a panel data regression in STATA
To: Statisticians_group@...
Date: Sunday, September 21, 2008, 12:08 PM

Hi everybody!
Does anybody knows how to collect the white noise from a panel data
regression in STATA??
Pleaseee.. need some help!

alexis

#2890

From: "Siwaig" <asym001@...>
Date:: Sun Sep 21, 2008 4:08 pm
Subject:: Urgent! How to collect the white noise from a panel data regression in STATA

siwaigyong

Offline

Hi everybody!
Does anybody knows how to collect the white noise from a panel data
regression in STATA??
Pleaseee.. need some help!

alexis

#2889

From: "Madan Gopal Kundu" <madan4331@...>
Date:: Sun Sep 21, 2008 3:55 pm
Subject:: Re: Online Surveys to earn more

madan4331

Offline

I discourage you to post this kind of mails in future.

--- In Statisticians_group@..., pradeep jegannathan
<pradeep_j1982@...> wrote:
>
>
>
>
>
> EARN $10 To $50
> PER HOUR FROM HOME EASILY.
>
>
>
> Minimum payout $75 Via Alert pay, cheque OR pay
> pal.
>
>
>
> Get $1.25 that Registers for Free.
>
>
>
> For more details please log on to
>
>
>
> http://www.AWSurveys.com/HomeMain.cfm?RefID=ambinu
>
>
>
> Instead of ambinu you can use your USERNAME.
>
>
>
> CLICK Create A Free Account Link In right side
> top.
>
>
>
> STEP 1.FILL THE FORM WITH USER NAME &
> PASSWORD.
>
>
>
> STEP 2FILL THE FORM WITH FIRST NAME & LAST
> NAME.
>
>
>
> STEP 3.CHOOSE YOUR GENDER & AGE WITH VALID
> EMAIL ADDRESS.
>
>
>
> STEP 4.CONFIRM YOUR REGISTRATION BY TYPING THE
> IMAGE & CLICK CREATE FREE ACCOUNT.
>
>
>
> ALL THE FOUR steps HARDLY TAKE A MAXIMUM OF
> 5MINUTES, BUT MAKE YOU GROW 10 TIMES OF YOUR CURRENT ECONOMIC
STATUS.
>
>
>
> How to View Account Details?
>
>
>
> http://www.awsurveys.com/login.cfm
>
>
>
> Enter your user name and Password & click
> Login.
>
>
>
> How to Make Changes In Account Information?
>
>
>
> Click MY ACCOUNT SETTINGS DISPLAYED NEAR YOUR
> id make the changes bring the cursor to the bottom of the page
Click UPDATE
> Account Info
>
>
>
>
>
> HOW TO EARN MONEY FROM AW SURVEYS?
>
>
>
> A $6.00 Website Evaluation is Available.
>
> A $4.00 Website Evaluation is Available
>
> A $4.00 Website Evaluation is Available
>
> A $4.00 Website Evaluation is Available
>
> A $4.00 Website Evaluation is Available
>
> A $4.00 Website Evaluation is Available
>
> A Monthly Bonus Website Evaluation is
> Available.
>
>
>
> Take the 7 Surveys in Homepage & get $27
> immediately.
>
>
>
> Refer Friends & Increase your earnings
> easily.
>
>
>
> Payment request will be processed within 7
> days.
>
>
>
> Wish You All the Best for Future earnings.
> AND
>
>
>
>
>
> I am giving some wonderful online free opportunities �" work
> from home.
>
> Most legitimate online home based jobs, we can work from
> home, part
>
> time jobs, work at any time, monthly payouts once your payment
amount
>
> reaches Rs.1000
>
> Go to the following web page by clicking or copy paste the link in
to
>
> your browser
>
>
>
> http://www.365jobs4u.com/idevaffiliate/pages/700.php
>
>
> Free sign up and bonus Rs.400. Sign up immediately
>
>
>
>
>
> http://www.onlinefreejobs2all.blogspot.com
>
> http://www.freenetjob.co.cc
>
> http://www.easyonlineincome.co.cc
>
>
>
>
>
>
>
>       Get your new Email address!
> Grab the Email name you&#39;ve always wanted before someone else
does!
> http://mail.promotions.yahoo.com/newdomains/aa/
>

#2888

From: Madan Kundu <madan4331@...>
Date:: Sat Sep 20, 2008 4:17 pm
Subject:: Fw: Urgent requirement for SAS Professional with MNC

madan4331

Offline

Those who are interested please contact with naveen@.... Please don't send me any mail directly to me regarding this.

--------------

Madan Gopal Kundu
Biostatistician, Ranbaxy Labs. Ltd.

Gurgaon, Haryana

India

Web: http://www.freewebs.com/madanstata

mobile: 91-9868788406

e-mail: Madan.Kundu@...

--- On Sat, 20/9/08, naveen@... <naveen@...> wrote:

From: naveen@... <naveen@...>
Subject: Urgent requirement for SAS Professional with MNC
To: madan4331@...
Date: Saturday, 20 September, 2008, 3:04 PM

Hi !

Hope you are doing well !!

We are looking for many candidates for SAS for Bangalore/Chennai/Delhi work locations for SE/SSE levels (2 to 5 yrs. exp..)

(There would be Telephonic round of interview on Short Notice).

Kindly send your latest updated resume along with below filled in inputs asap so that we can process it further for the interview.

Full Name:

Present Location :

Choice Work Location:(Bangalore/Chennai,Delhi)

Present Company Name :

Total IT Exp :

SAS Exp :

Present Role :

Qualifications :

Current CTC :

Expected CTC :

Total Team Size :

Total Onsite Exp (if any) :

Notice Period :

D.O.B:

Contact No. to attend Telephonic Interview :

Kindly send your updated cv asap so that we process it further.

Also refer us few of your colleagues and friends who would be interested for the same.

Thanks and Regards
_______________________________________
Naveen Kumar
Email : naveen@...
INTEREC IT GLOBAL EXECUTIVE SEARCH
NOIDA. CONTACT-91-120- 4623619

Disclaimer:

The sender of this email is registered with naukri.com as Interec IT Global Executive Search (, Sector 27, Shop No 2, Nr Vinayak Hospital, Noida, Uttar Pradesh - 201303) using Naukri.com services. The responsibility of checking the authenticity of offers/correspondence lies with you.

If you consider the content of this email inappropriate or spam, you may:

Report abuse by forwarding this email to: abuse@...

You can also Block this Company from searching your resume in the database.

Please note that Naukri.com does NOT endorse any requests for money payments, or sharing of bank account details.

Ads by Google

Employment Jobs Career Opportunity

Partner Sites

Find your dream match on Jeevansathi.com Get career advice on asknaukri.com

Buy/Sell/Rent Property on 99acres.com Brijj your network to win a trip to London

Sponsored Links

Click and get the best Airfares @ MMT!

Add more friends to your messenger and enjoy! Invite them now.

#2887

From: pradeep jegannathan <pradeep_j1982@...>
Date:: Sat Sep 20, 2008 3:12 am
Subject:: Online Surveys to earn more

pradeep_j1982

Offline

http://www.365jobs4u.com/idevaffiliate/pages/700.php

EARN $10 To $50 PER HOUR FROM HOME EASILY.

Minimum payout $75 Via Alert pay, cheque OR pay pal.

Get $1.25 that Registers for Free.

For more details please log on to

http://www.AWSurveys.com/HomeMain.cfm?RefID=ambinu

Instead of ambinu you can use your USERNAME.

CLICK Create A Free Account Link In right side top.

STEP 1.FILL THE FORM WITH USER NAME & PASSWORD.

STEP 2FILL THE FORM WITH FIRST NAME & LAST NAME.

STEP 3.CHOOSE YOUR GENDER & AGE WITH VALID EMAIL ADDRESS.

STEP 4.CONFIRM YOUR REGISTRATION BY TYPING THE IMAGE & CLICK CREATE FREE ACCOUNT.

ALL THE FOUR steps HARDLY TAKE A MAXIMUM OF 5MINUTES, BUT MAKE YOU GROW 10 TIMES OF YOUR CURRENT ECONOMIC STATUS.

How to View Account Details?

http://www.awsurveys.com/login.cfm

Enter your user name and Password & click Login.

How to Make Changes In Account Information?

Click MY ACCOUNT SETTINGS DISPLAYED NEAR YOUR id make the changes bring the cursor to the bottom of the page Click UPDATE Account Info

HOW TO EARN MONEY FROM AW SURVEYS?

A $6.00 Website Evaluation is Available.
A $4.00 Website Evaluation is Available
A $4.00 Website Evaluation is Available
A $4.00 Website Evaluation is Available
A $4.00 Website Evaluation is Available
A $4.00 Website Evaluation is Available
A Monthly Bonus Website Evaluation is Available.

Take the 7 Surveys in Homepage & get $27 immediately.

Refer Friends & Increase your earnings easily.

Payment request will be processed within 7 days.

Wish You All the Best for Future earnings.

AND

I am giving some wonderful online free opportunities – work from home.

Most legitimate online home based jobs, we can work from home, part
time jobs, work at any time, monthly payouts once your payment amount
reaches Rs.1000

Go to the following web page by clicking or copy paste the link in to
your browser

Free sign up and bonus Rs.400. Sign up immediately

http://www.onlinefreejobs2all.blogspot.com

http://www.freenetjob.co.cc

http://www.easyonlineincome.co.cc

New Email names for you!
Get the Email name you've always wanted on the new @ymail and @rocketmail.
Hurry before someone else does!

#2886

From: Jitendra Salunkhe <jitu_stat@...>
Date:: Fri Sep 19, 2008 3:13 pm
Subject:: Resume for Statistician(Analyst)

jitu_stat

Offline

Dear Friends,

I am recently completed M. Sc.(Industrial Statistics), now I am Looking for job, I am Stay at Mumbai, if any where requirement please consider as an application, please go through attachment for resume

Thanks,

Regard,

Jitendra_

09421014533,09819601421

Get an email ID as yourname@... or yourname@.... Click here.

#2885

From: farah dhiab <farahdhiab@...>
Date:: Thu Sep 18, 2008 10:17 pm
Subject:: Re: A DOE Question

farahdhiab

Offline

I have some questions considering your experiment, how many levels did u set for each variable x? And can u send your design plan so I can have better idea of your experiment?

Also, have you thought of using robust parameter design?? If you think that idea might help u to get better optimum results, email me back and we will discuss it in focus.

Regards,

Farah

--- On Thu, 9/18/08, Pradipta Sarkar <sarkarp2002@...> wrote:

From: Pradipta Sarkar <sarkarp2002@...>
Subject: [Statisticians_group] A DOE Question
To: Statisticians_group@...
Date: Thursday, September 18, 2008, 6:49 AM

I have a real life problem .... would appreciate any help you can provide.

Let us say we have k variables (which are ingredients of a product, k is about 20). We want to study some technical response Y as a function of the k variables X_1, X_2, ..., X_k. The variable X_i can take value between 0 and a_i

So we can create a response surface design using some optimality criteria such as D-Optimality or I-Optimality. Ater running the experiment and collecting data we can fit the model and then find the combination of X_1, X_2, ..., X_k that maximizes the response Y. For this specific application (involving this particular product) usually the maximum is attained at a point where several of the X_1, X_2, ..., X_k variables are 0 which means it is on the boundary of the design space.

We know that the confidence interval of prediction is very wide at the boundary (no matter what optimality criteria we use), so the confidence interval on the maximum reponse is likely to be very wide.

So here is my question: Is there a way to design an experiment so that "it has lower prediction variance at the boundary"? My intuition says that while designing the experiment put more points at the corners and sides of the design space and less inside the design space. Is that a good strategy or can you suggest another better one? Won't that destroy some desired properties of DOE? What will I lose? Is there a standard way to handle such design of experiments?

Remember that the X variables are ingredients, so the minimum value is 0(cannot be negative).

Thanks,

Pradipta.

#2884

From: Pradipta Sarkar <sarkarp2002@...>
Date:: Thu Sep 18, 2008 3:49 am
Subject:: A DOE Question

sarkarp2002

Offline

I have a real life problem .... would appreciate any help you can provide.

Let us say we have k variables (which are ingredients of a product, k is about
20).  We want to study some technical response Y as a function of the k
variables X_1, X_2, ..., X_k.  The variable X_i can take value between 0 and a_i

So we can create a response surface design using some optimality criteria such
as D-Optimality or I-Optimality. Ater running the experiment and collecting data
we can fit the model and then find the combination of X_1, X_2, ..., X_k that
maximizes the response Y.  For this specific application (involving this
particular product) usually the maximum is attained at a point where several of
the X_1, X_2, ..., X_k variables are 0 which means it is on the boundary of the
design space.

We know that the confidence interval of prediction is very wide at the boundary
(no matter what optimality criteria we use), so the confidence interval on the
maximum reponse is likely to be very wide.

So here is my question: Is there a way to design an experiment so that "it has
lower prediction variance at the boundary"?  My intuition says that while
designing the experiment put more points at the corners and sides of the design
space and less inside the design space.  Is that a good strategy or can you
suggest another better one?  Won't that destroy some desired properties of DOE? 
What will I lose?   Is there a standard way to handle such design of
experiments?

Remember that the X variables are ingredients, so the minimum value is 0(cannot
be negative).

Thanks,

Pradipta.

#2883

From: "poornima285" <poornima285@...>
Date:: Wed Sep 17, 2008 4:22 pm
Subject:: slpoping control charts

poornima285

Offline

hi everbody,

please expalin me the concept of sloping control charts.

#2882

From: poornima padmanabhan <poornima285@...>
Date:: Wed Sep 17, 2008 4:09 pm
Subject:: sloping control charts

poornima285

Offline

hi everybody,

iam poornima frm bangalore. i have completed my MSc in statistics. iam currently searching for jobs. kindly guide me.

could anybody explain wht r sloping control charts?

thanks

poornima

Unlimited freedom, unlimited storage. Get it now

#2881

From: "Zidan M. Houmani" <houmani67@...>
Date:: Tue Sep 16, 2008 6:55 am
Subject:: NEED HELP

houmani67

Offline

Dear group members

I am using SPSS software and i analysis my data by it, Could you please one of you who has knowledge about it help me to write a report on this data results.

You�re truly

Mr.Zidan Mohamed Houmani

#2880

From: "Robert Newcombe" <newcombe@...>
Date:: Wed Sep 10, 2008 11:40 am
Subject:: Re: (Likert scale)

newcomberobert

Offline

A couple of points, further to Jay's response. Many would argue that Likert
scales are inherently
only suitable for analysis by non-parametric methods - which would of course
rule out calculating
the mean (and indeed the SD). The counter-argument is that the alternative
measures of central
tendency - of which the simplest are the median and the mode - are simply
insufficiently
informative, particularly when we come to compare between two samples. For a
Likert scale with k
points, running from 1 to k, the median and the mode can only take 2k+1 possible
values, i.e. 1,
1.5, 2, 2.5, ..., k. A k-point Likert scale fom 0 to k-1 is exactly equivalent,
the median and the
mode can only take 2k+1 possible values, i.e. 0, 0.5, 1, 1.5, ..., k-1.
Particularly when k is
small, this is a blunt instrument to characterise differences between groups.

The meaningfulness of calculating the mean depends crucially on one assumption,
namely that in some
meaningful sense the scale is equally spaced. I guess this is what is often
meant by an interval
scale, but it is much more specific to define it this way.

Some time ago I was prompted to consider Likert scales with just 3 points. A
holiday operator
distributed a questionnaire on the return flight, in which the respondent had to
choose one of three
Chernoff faces (or smileys) to say what s/he thought of various aspects of the
holiday. One, with
the edges of the mouth upturned, was meant to represent happy, the opposite
extreme was a mouth with
edges downturned, indicating unhappy, the middle point in the scale was a
straight mouth indicating
neutral. It was obviously intended to be equally spaced. Here, summarising by
the median or mode
gives just 5 possible values, which is unsatisfactory. The mean is obviously
easy to calculate - but
what about calculating a confidence interval for it?

Calculating a CI for the mean of a variable on a 0-1-2 scale also arises in
other contexts, such as
surgical procedures on paired organs such as breasts, or estimating gene
frequency without assuming
Hardy-Weinberg equilibrium. Getting a CI for such a mean turns out to be
equivalent to getting a CI
for a paired differences of proportions. I wrote a paper on this, which was
published as

Newcombe RG.  Confidence intervals for the mean of a variable taking the values
0, 1 and 2.
Statistics in Medicine, Sept 2003, 22(17), 2737-2750.

I'll be happy to email a pdf to anyone who would like a copy.

Robert G. Newcombe PhD CStat FFPH
Professor of Medical Statistics
Department of Primary Care and Public Health
Centre for Health Sciences Research
Cardiff University
4th floor, Neuadd Meirionnydd
Heath Park, Cardiff CF14 4YS

Tel: 029 2068 7260
Fax: 029 2068 7236

Home page
http://www.cf.ac.uk/medic/contactsandpeople/n/newcombe-robert-gordon-prof-overvi\
ew_new.html
For location see
http://www.cardiff.ac.uk/locations/resources/Faxable%20Map-07.pdf

>>> Jay Warner <quality@...> 20/08/08 19:23 >>>

Short answer:  If you use 0 to 4, you still have five marked points
on the scale.  You can still calculate averages and standard
deviations (if you believe it is close enough to an interval scale to
do this).

Just don't confuse it with a ratio scale, where 0 times another
point  = 0.  Unless 0 really means that you have none of the measured
item.

Long answer:  As I understand it, a Likert Scale is as you make it.
Likert used 1 to 7 as well as 1 to 5 himself.

Having one point in the absolute center is generally a GOOD IDEA.
Fence sitters respond that way.

If you consider that partial - midpoint answers such as halfway
between 3 and 4 - 3.5 -  are unacceptable, than you have forced the
scale to be less sensitive to nuanced responses, not more.

The difficulty is that different respondents may see a given point on
the scale as being a different level - degree of response - than others.

Likert used it for (basically) degree-of-emotion type responses.  It
can work for lots of other weakly defined responses as well.

If you can define a few points on the scale by example, so that the
respondents more or less agree, then you can really make headway.
For physical parts, a sample of a "bad" one and a "good" one at
points 1 and 5 on the 5 point scale make all the difference in the
world.  And let people put down 3.5 if they like.

Jay Warner
Principal Scientist
Warner Consulting, Inc.
4444 North Green Bay Road
Racine, WI 53404-1216
USA

Ph:       262.634.9100
Fax:     262.681.1133
email:  quality@...
web:    www.a2q.com

On Aug 20, 2008, at 9:58:32 AM, Hemant Trivedi wrote:

> Is it necessary that Likert Scale should be from 1 to 5 or it can
> be from 0 to 4 also.

Reply | Forward | Messages in this Topic (3)

#2879

From: "sch0pp" <sch0pp@...>
Date:: Fri Sep 5, 2008 11:59 am
Subject:: Statistical Management Openings with Capital One!

sch0pp

Offline

Thank you for taking a look at this message, I'll be as brief as
possible I know you're busy.

Capital One (Auto Finance and U.S. Card) is looking to expand.  I've
been tasked to find the best in their fields regarding the highly
touted positions.

If you're interested in hearing more, let me know (via e-mail or
phone).  If not, please forward this position to others as I've
several openings to fill.

----------------------------------------------------------------------
------------------------------------------
Senior Statistical Analysis Manager in Plano, TX

Basic Qualifications:
Masters or higher degree in statistics, economics, or biostatistics
with 2+ years of applied experience
1+ years of hands on risk modeling experience
Manager of Statistical Analysis - Richmond, VA OR Plano, TX

Basic Qualifications:

Master's degree in Statistics or related quantitative field
four years of experience working with classic Statistical tools such
as linear models, logistic regression, hypothesis testing or
sampling.
Proficient in SAS programming
----------------------------------------------------------------------
------------------------------------------

Please give me a call.

Thank you and have a great day.

Greg Schopp
      Recruiting Consultant for CapitalOne (Auto Finance)


E-mail: Greg.Schopp@...

#2878

From: rakeshkumar khowal <rakeshshail2010@...>
Date:: Thu Sep 4, 2008 6:01 pm
Subject:: Business Presentation (Need Your Comments)

rakeshshail2010

Offline

Hi Friends,

I was provided this presentation. I have been assigned the work to update the presentation with the result data (that you can follow lookin at the attachment) and do commentory. Though my commentory has been approved but, I am not satisfied with the kind of commentory I did. Please give me more insights..

Regards & Thanks,

Rakesh Khowal .

Unlimited freedom, unlimited storage. Get it now