**Type-I error**: The
mistake of rejecting a true null hypothesis. A maximum acceptable probability
of Type-I error should be set during the design stage, before statistical
analysis. Across much of the biological sciences, it is conventionally taken as
*α* = 0.05, in which case the
analysis will show significant effects if outputs yield *P* < 0.05.

The associated Type-II error is the mistake of accepting a
false null hypothesis. A maximum acceptable probability of Type-II error should
also be set during the design stage, before data collection, because the power of the design equals 1 - *β*.
A probability *β* = 0.20 is often
regarded as acceptable, meaning that the analysis will have a 0.8 chance of
rejecting a false null hypothesis, and therefore a high probability of
identifying pattern in the data if it exists.

Doncaster, C. P. & Davey, A. J. H. (2007) *Analysis of Variance and Covariance: How to
Choose and Construct Models for the Life Sciences*. Cambridge: Cambridge
University Press.

http://www.southampton.ac.uk/~cpd/anovas/datasets/