Cohen's conventional benchmarks: 0.2 = small, 0.5 = medium, 0.8 = large. Small effects are not necessarily unimportant.