Abstract
The Shapiro-Wilk test for non-normality is highly sensitive to the presence of ties due to grouping or rounding of the raw data, and should not be used if the grouping interval exceeds 0.1 standard deviation units. A simple method of modifying Wfor known grouping intervals is presented, and the case of unequal intervals is covered. The modified Wtest may be applied over a wide range of grouping intervals (0-1.5 S.D. units) and sample sizes (7-2000), and is shown to have high power whencompared with a variant of the Pearson χ2test.

This publication has 16 references indexed in Scilit: