Читать книгу Health Psychology - Michael Murray - Страница 109

It is common in research to make multiple statistical comparisons. At the .05 or 5% significance (α) level with 20 tests there is a 64% chance of observing at least one significant result, even if none is actually significant (a Type I error). In a research project, numerous simultaneous tests may be required and the probability of getting a significant result simply by chance may be very high. The α level must be adjusted in some way, so that the probability of observing at least one significant result due to chance remains below the desired significance level. The Bonferroni correction places the significance cut-off at α/n. For example, with 20 tests and α at .05, you would only reject a null hypothesis if the p-value were less than .05/20 = 0.0025.