jmp-l
[Top] [All Lists]

Re: Comparing Regression Slopes

To: <jmp-l@lists.biostat.wustl.edu>
Subject: Re: Comparing Regression Slopes
From: <Gunter.Hartel@csl.com.au>
Date: Mon, 20 Jun 2005 10:57:00 +1000
Thread-index: AcVy/BTPsfj7veMVTTaAyoKM3zRDMACNq+zw
Thread-topic: [jmp-l] Comparing Regression Slopes
Sean,
I wrote a short article on that topic for JMP'er Cable 
(http://www.jmp.com/news/jmpercable/15_summer_2004.pdf ).  
You might find it useful.
Cheers
Gunter

                -----Original Message-----
                From:   jmp-l-owner@lists.biostat.wustl.edu 
[mailto:jmp-l-owner@lists.biostat.wustl.edu] On Behalf Of Davern, Sean
                Sent:   Friday, 17 June 2005 2:41 PM
                To:     'jmp-l@lists.biostat.wustl.edu'
                Subject:        [jmp-l] Comparing Regression Slopes

                 << File: Growth Rate Analysis.JMP >> I have a relatively 
simple analysis that I can't determine how to
                appropriately do with JMP.

                I'm measuring log-phase cell growth in two different vessels.  
So, from
                first principles I know:
                VCD = VCD0 * exp( a * time)
                Where VCD is viable cell density, VCD0 is the initial cell 
density and a is
                the growth rate constant.
                  I want to compare growth rate (a) in two vessels.  Lets say I 
run 2 (1
                replicate) of each vessels.  So, I inoculate 4 vessels and 
measure VCD0.
                Daily I take samples, record time and measure VCD (a 
measurement with large
                variability).  I take measurements for 3 days giving 4 points 
for each
                curve.  Lets not assume that VCD0 is the same for all tanks 
even if I
                inoculate them with a common inoculum.

                  I now want to know if a is a function of vessel type. I 
believe that I
                might also be able to determine if differences of growth 
between two vessels
                of the same type was measured.

                  This site references J. Zar and describes a statistical 
analysis for
                comparing slopes and intercepts of linear regressions:
                http://www.curvefit.com/1_model__2_datasets.htm
                I haven't gotten the reference yet, so I haven't read the 
background.

                  I'm wondering if I can use the Fit Model platform if I use a 
log
                transformation the VCD data as Y.  Then I want to determine if 
the slopes of
                a linear regression are distinguishable.  The question is how 
do I structure
                the model effects?  I believe that if I include "time [vessel]" 
(the nested
                combination) and "vessel" as my model effects I'll get the 
ANCOVA analysis.
                p-values for the "vessel(type):time" parameter estimates will 
tell me if
                slopes (a) are statistically distinguishable.  Furthermore, if 
I also nest
                "Run ID" appropriately I'll also get the within-vessel 
comparison.

                  How do I specify the nesting correctly?  Will this give me 
the correct
                statistical analysis?  It seems to me that the p-values will 
tell me if the
                model terms are significant with respect to the model but it 
isn't obvious
                to me that they tell me the instances are statistically 
distinguishable.

                  The attached JMP ver 5.1.2 file contains example data.
                 <<Growth Rate Analysis.JMP>> 
                Thanks, in advance, for any guidance.
                Sean

                Sean Davern
                Amgen
                1201 Amgen Court West
                Mail Stop AW2/D2152
                Seattle, WA 98119-3105
                Phone: (206) 265-7074
                


****************************************************************************************************************
This email and any attached files are intended solely for the named addressee, 
are confidential and may contain legally privileged information. The copying or 
distribution of them or any information they contain, by anyone other than the 
addressee, is prohibited.  If you have received this email in error, please let 
us know by telephone or return the email to the sender
and destroy all copies. Thank you.

CSL Limited A.C.N. 051 588 348
45 Poplar Road Parkville Victoria 3052 Australia
Phone: +61 3 9389 1911  Fax: +61 3 9389 1434
***************************************************************************************************************


<Prev in Thread] Current Thread [Next in Thread>