I was interested in exploring if spatial scale and the presence of invasion contributed to differences in response variables; however, due to small sample sizes (replicates were 'taken up' by calculating effect sizes), this analysis on the effect sizes themselves was descriptive.
For a gene showing average within-group variability, our sample size (four replicates with 30 mM benzoate and four replicates without benzoate, 4,351 ORFs) provided statistical power of 98% to detect a 2-fold difference in gene expression between experimental conditions.
Others leave more to the evaluator by asking if the sample size and replicates is sufficient (see [11]).
The reported aspects connected to "statistical design" are: statistical method and results including significance levels and estimates of variability, (sufficient) sample size and replicates, randomized treatments and independent observations.
While this reanalysis study has the strengths of a larger sample size, independent replicates, and well characterized patient samples from specific areas of cortex, the finding should be interpreted cautiously as this study has some limitations.
This analysis and post-mortem sample size with replicates is noteworthy for an adequately powered sample to detect 1.3 fold changes, improving sensitivity, reliability, and signal to noise issues.
A recent study estimated the required sample size to replicate an association finding with different amounts of between-study heterogeneity [28].
This may be explained by different environmental risk profiles between Europeans and Asians, body composition and genetic backgrounds, or that we have insufficient power with current sample size to replicate some of these previously reported risk variants.
The categorical nature of five of the seven variables meant that reduction of the sample size to replicate the latent class structure and cross-validate the results was not possible.
The trial was of a randomised triplicate design, the sample size per replicate and treatment was 30 bunches.
