Don't trust small sample sizes
BlogJuly 15, 2019Tags:Why small sample sizes might be even worse than you think

Why small sample sizes might be even worse than you think
A short course on R for non-programmers
Why you should never say 'a trend towards significance'