Dear everyone,
Sorry to intervene but to move on to a more philosophical (cynical level), have power analyses been applied successfully at all for planning fMRI studies (leaving aside that they are a very good idea of course)?
Based on Web of Knowledge the Mumford & Nichols (2008) paper was cited 39 times, the guide by Mumford (2012) in SCAN was cited 5 times. Looking at the titles several of those are of methodological nature. Of the remaining ones I checked a few, one just stated that fMRI analyses were often underpowered, another seems to rely on a posthoc power analysis, another states that based on power calculations their no. of trials was "sufficient" without any further information on how they came to that conclusion. Guo et al. (2014, https://dx.doi.org/10.1016/j.neuroimage.2013.08.012 ) report that from 100 chosen fMRI papers a single study included a power calculation.
Citations doesn't mean that much, but based on these numbers it seems as if hardly anyone relies on power calculations, which is somewhat disturbing. Maybe it's not just ignorance of the researchers but also limitations of the available methods (their assumptions)?
Best
Helmut
|