Dear SPM community,

 

My colleagues and I have reached an impasse with our local ethics committee regarding power and sample size calculations. The committee “requests” that we include a power analysis before they will give us final approval.

 

While in principle I agree that power analyses are important it is, as we know, not straight forward in functional imaging.

 

I have read Jeanette Mumford’s recent paper in SCAN and looked at both fmripower and PowerMap. As is common in all power/sample size analysis, pilot data is required to estimate effect size and additionally an idea of location. While we have pilot data for the normal control arm of our study, we now wish to look at a clinical population and a new control group which is why we require ethical approval from this particular committee. This, of course, means we don’t have pilot data for the clinical arm.

 

Usually we would use previous data, which we don’t have, or data from the literature to estimate response. However, our task is novel, and therefore it is difficult to find valid estimates of effect size.

 

We are meeting with our institutional statistics department in the new year to get an independent opinion from a medical statistician but we wonder if there is a form of words that others have used to explain the issues involved, or whether any of the statisticians on the list with neuroimaging expertise can help.

 

Thanks in advance and a Merry Christmas and a Happy New Year,

 

Gordon

 

 

Gordon D. Waiter PhD CSci MIPEM CPhys MInstP
Aberdeen Biomedical Imaging Centre
Division of Applied Medicine
University of Aberdeen
Research MRI Centre
Lilian Sutton Building
Foresterhill
Aberdeen
AB25 2ZD

Please note new telephone number

Tel: +44 (0)1224 438356 (Internal 8356)
Fax: +44 (0)1224 438364

[log in to unmask]
www.abdn.ac.uk/ims/imaging