GraphPad Statistics Guide

Q&A: Normality tests

Q&A: Normality tests

Previous topic Next topic Expand/collapse all hidden text  

Q&A: Normality tests

Previous topic Next topic JavaScript is required for expanding text JavaScript is required for the print function Mail us feedback on this topic!  


Why the term "normality"?
Which normality test is best?
Why do the different normality tests give different results?
How many values are needed to compute a normality test?
What question does the normality test answer?
What cut-off does Prism use when deciding whether or not a data set passed a normality test?
So it tells me whether a data set is Gaussian?
Are any data sets truly sampled from ideal Gaussian distributions?
But don't t tests, ANOVA, and regression assume Gaussian distributions?
So do the normality tests figure out whether the data are close enough to Gaussian to use one of those tests?
Isn't the whole point of a normality test to decide when to use nonparametric tests?
How should I interpet the K2, KS or W values reported by the normality test?


How useful are normality tests?


1 RB D'Agostino, "Tests for Normal Distribution" in Goodness-Of-Fit Techniques edited by RB D'Agostino and MA Stepenes, Macel Decker, 1986.

Parts of this page are excerpted from Chapter 24 of Motulsky, H.J. (2010). Intuitive Biostatistics, 2nd edition. Oxford University Press. ISBN=978-0-19-973006-3.