Testing whether confidence intervals overlap is a poor test for statistical significance -- confidence intervals can overlap and still represent a statistically

Testing whether confidence intervals overlap is a poor test for statistical significance -- confidence intervals can overlap and still represent a statistically significant difference between groups. http://www.refsmmat.com/statistics/significant-differences.html#when-significant-differences-are-missed http://www.refsmmat.com/statistics/significant-differences.h... Skimming through the paper, it appears they did not correct for making multiple comparisons. That will raise their false positive rate. Oh well...