As Senn, Stevens, & Chaturverdi (Statist Med, 2000,
19, 861-877) discuss, MANOVA doesn't provide useful
estimates of the treatment effect. 

Better alternative include the summary measures
approach (eg, regression through the origin), random
coefficients, or generalized estimating equations.

SR Millis


--- Rich Ulrich <[EMAIL PROTECTED]> wrote:
> On 3 Dec 2003 04:41:09 -0800, [EMAIL PROTECTED]
> (Karl) wrote:
> 
> > Hallo!
> > 
> > I have data of the following design: 
> > N subjects were measured at Baseline (visit 1) and
> at 3 following time
> > points (visit 2-4). There is or is not a
> treatment. Say, the results
> > of visits 1 to 4 are Y1, Y2, Y3, Y4.
> > 
> > T: Variable Treatment (yes/no)
> > V: Variable Visit (2, 3, 4 - without baseline)
> > id: Variable subject identification number
> > N: number of subjects
> > 
> > Most interesting is the question if there is a
> difference in treatment
> > between the results of visit 4 and baseline. (The
> other time points
> > are also of interest.)
> > 
> 
> The MANOVA suggested by Jim Clark sounds like good
> advice  to me, given the absence of warning signs.
> 
> Is change expected in both groups?  - Are equal
> changes 
> expected in each time period, or would more change
> be 
> expected early? - early change could justify using
> the 
> average of three followups as the main criterion.
> 
> How serious are you about the single end-point at
> visit 4?
>  - that suggests a simpler design, where you compare
> the
> treatments at 4  while using 1  as a covariate.  
> 
> For any design, it will be useful to confirm that
> there 
> were no initial differences, despite randomization
> of the
> treatment condition.  Of course, the equal-start is
> more 
> critical if subjects  were not randomized.
> 
> -- 
> Rich Ulrich, [EMAIL PROTECTED]
> http://www.pitt.edu/~wpilib/index.html
> "Taxes are the price we pay for civilization." 
> .
> .
>
=================================================================
> Instructions for joining and leaving this list,
> remarks about the
> problem of INAPPROPRIATE MESSAGES, and archives are
> available at:
> .                  http://jse.stat.ncsu.edu/        
>            .
>
=================================================================
> 
> 



=====
Scott R Millis, PhD, MEd, ABPP (CN & RP)
Director, Office of Clinical Trials
Kessler Medical Rehabilitation Research & Education Corp
1199 Pleasant Valley Way
West Orange, NJ  07052

*********************************************************
This electronic message may contain information that is confidential
and/or legally privileged. It is intended only for the use of the
individual(s) and entity named as recipients in the message. If you are
not an intended recipient of this message, please notify the sender
immediately and delete the material from any computer. Do not deliver,
distribute or copy this message, and do not disclose its contents or
take any action in reliance on the information it contains. Thank you.
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to