Log in
RegisterSample Size and CV for Replicate Design [Power / Sample Size]
Hi J.,
OK,
in a way it sounds like you are talking about confidence limits for the variability now and not for the point estimate. If that is the case then you are entering a very complicated territory where I have little to offer.
Generally, most people seek to calculate the sample size that assures a power of X, given a CV of Y and a point estimate (theta, test:reference ratio) of Z.
When I re-read your original post, it almost sounds like it is a goal in itself to keep the CV above 30%. For ref-scaling purposes? I must admit it sounds wrong to me to aim for an observed CV above a certain limit, so I probably misunderstood something?!
Could you perhaps reformulate (I hate that word!) your assignment or put more words to it?
Many thanks.
❝ I downloaded the package and indeed I see the 2x2x4 design is supported. However, given the information I have available and my ultimate goal, I'm not sure how to specify what I need. I see targetpower=0.80 and alpha=0.05, as well as a few other arguments such as CV and theta. Do I have enough information in what I posted in the initial post to solve my problem using PowerTOST? I should mention I've never used the CV metric before, and I don't see an argument to specify both the 0.35 and 0.30 numbers. Are these values some how related to the 'theta' values that I'm also asked to specify?
OK,
in a way it sounds like you are talking about confidence limits for the variability now and not for the point estimate. If that is the case then you are entering a very complicated territory where I have little to offer.
Generally, most people seek to calculate the sample size that assures a power of X, given a CV of Y and a point estimate (theta, test:reference ratio) of Z.
When I re-read your original post, it almost sounds like it is a goal in itself to keep the CV above 30%. For ref-scaling purposes? I must admit it sounds wrong to me to aim for an observed CV above a certain limit, so I probably misunderstood something?!
Could you perhaps reformulate (I hate that word!) your assignment or put more words to it?
Many thanks.
—
Pass or fail!
ElMaestro
Pass or fail!
ElMaestro
Complete thread:
- Sample Size and CV for Replicate Design Jamesmartinn 2011-11-26 20:22
![[Mix]](https://static.bebac.at/img/mix.png "open in Mix view")
- Sample Size and CV for Replicate Design ElMaestro 2011-11-26 22:38
- Sample Size and CV for Replicate Design Jamesmartinn 2011-11-26 23:52
- Sample Size and CV for Replicate DesignElMaestro 2011-11-27 00:43
- Sample Size and CV for Replicate Design Jamesmartinn 2011-11-27 00:50
- Sample Size and CV for Replicate Design ElMaestro 2011-11-27 01:24
- Sample Size and CV for Replicate Design Jamesmartinn 2011-11-27 01:50
- Strange questions… Helmut 2011-11-27 04:47
- Strange questions… Jamesmartinn 2011-11-27 14:43
- Misusing PowerTOST for superiority? Helmut 2011-11-27 17:46
- Strange questions… ElMaestro 2011-11-27 15:14
- Strange questions… Helmut 2011-11-27 18:07
- Strange questions… Jamesmartinn 2011-11-27 14:43
- Strange questions… Helmut 2011-11-27 04:47
- Sample Size and CV for Replicate Design Jamesmartinn 2011-11-27 01:50
- Sample Size and CV for Replicate Design ElMaestro 2011-11-27 01:24
- Sample Size and CV for Replicate Design Jamesmartinn 2011-11-27 00:50
- Sample Size and CV for Replicate DesignElMaestro 2011-11-27 00:43
- Sample Size and CV for Replicate Design Jamesmartinn 2011-11-26 23:52
- Sample Size and CV for Replicate Design ElMaestro 2011-11-26 22:38
Ing. Helmut Schütz