Footnote 1

The level of significance measures the so-called Type I error, or the likelihood of accepting an observed difference as real when it is, in fact, due to chance alone. A complementary concept is the Type II error, which measures the likelihood of rejecting an observed difference as unreal when it is, in fact, real. When the level of significance used to test an observed difference is made more stringent, the power of the test to detect a real difference is decreased. The power of a statistical test is very important when negative results, i.e. no effect is seen, are observed. The design of the study and the sample size predetermine the probability of a false negative result.