Not regarding this study, but in general those who are saying they don't need any sig test because they only look at the effect size know *squat* about statistics.++With small sample sizes a large difference means *nothing*.++With large sample sizes you can detect very small differences and these differences and if there *is* a difference you will detect it, however small or meaningful if not there won't be.++At any rate if you look at the trend the difference is not meaningful either since the graphs are very clearly intersecting unless there is a MAJOR unforseen change in trend in the next little while completely different from the past half year.
