When does blocking help? Should we always do it if we can? In this post we talk about how there are actually different kinds of blocking, and different ways of thinking about where blocks come from. The question of whether blocking can ever cause harm turns out to depend on these perspectives.

An annoyance (for me) is when I want to compare different groups portrayed in a nice plot that is kind enough to provide error bars. Consider a hypothetical data collection effort where we have 5 groups and are looking at the mean outcome of these 5 groups. Our estimated means and standard deviations are as follows: …

Once an analysis is in, we need to think carefully about how to talk about our results. Statistics is about inference, about capturing the uncertainty in our findings given limitations of our data and estimation error. We want the reader to quickly assess how seriously to take an estimated difference, and we want to …