Then I came across a recent fire ecology paper in a highly respected journal. The researchers simply compared variables across a number of groups; if the 95% confidence intervals didn't overlap then they considered there to be a ...