nag_shapiro_wilk_test (g01ddc) (PDF version)
g01 Chapter Contents
g01 Chapter Introduction
NAG Library Manual

NAG Library Function Document

nag_shapiro_wilk_test (g01ddc)

+ Contents

    1  Purpose
    7  Accuracy

1  Purpose

nag_shapiro_wilk_test (g01ddc) calculates Shapiro and Wilk's W statistic and its significance level for testing Normality.

2  Specification

#include <nag.h>
#include <nagg01.h>
void  nag_shapiro_wilk_test (Integer n, const double x[], Nag_Boolean calc_wts, double a[], double *w, double *pw, NagError *fail)

3  Description

nag_shapiro_wilk_test (g01ddc) calculates Shapiro and Wilk's W statistic and its significance level for any sample size between 3 and 5000. It is an adaptation of the Applied Statistics Algorithm AS R94, see Royston (1995). The full description of the theory behind this algorithm is given in Royston (1992).
Given a set of observations x1,x2,,xn sorted into either ascending or descending order (nag_double_sort (m01cac) may be used to sort the data) this function calculates the value of Shapiro and Wilk's W statistic defined as:
W= i=1naixi 2 i=1n xi-x- 2 ,
where x- = 1n 1n xi  is the sample mean and ai , for i=1,2,,n , are a set of ‘weights’ whose values depend only on the sample size n.
On exit, the values of ai, for i=1,2,,n, are only of interest should you wish to call the function again to calculate w and its significance level for a different sample of the same size.
It is recommended that the function is used in conjunction with a Normal Q-Q plot of the data. Function nag_normal_scores_exact (g01dac) can be used to obtain the required Normal scores.

4  References

Royston J P (1982) Algorithm AS 181: the W test for normality Appl. Statist. 31 176–180
Royston J P (1986) A remark on AS 181: the W test for normality Appl. Statist. 35 232–234
Royston J P (1992) Approximating the Shapiro–Wilk's W test for non-normality Statistics & Computing 2 117–119
Royston J P (1995) A remark on AS R94: A remark on Algorithm AS 181: the W test for normality Appl. Statist. 44(4) 547–551

5  Arguments

1:     nIntegerInput
On entry: n, the sample size.
Constraint: 3n5000.
2:     x[n]const doubleInput
On entry: the ordered sample values, xi, for i=1,2,,n.
3:     calc_wtsNag_BooleanInput
On entry: must be set to Nag_TRUE if you wish nag_shapiro_wilk_test (g01ddc) to calculate the elements of a.
calc_wts should be set to Nag_FALSE if you have saved the values in a from a previous call to nag_shapiro_wilk_test (g01ddc).
If in doubt, set calc_wts equal to Nag_TRUE.
4:     a[n]doubleInput/Output
On entry: if calc_wts has been set to Nag_FALSE then before entry a must contain the n weights as calculated in a previous call to nag_shapiro_wilk_test (g01ddc), otherwise a need not be set.
On exit: the n weights required to calculate w.
5:     wdouble *Output
On exit: the value of the statistic, w.
6:     pwdouble *Output
On exit: the significance level of w.
7:     failNagError *Input/Output
The NAG error argument (see Section 3.6 in the Essential Introduction).

6  Error Indicators and Warnings

NE_ALL_ELEMENTS_EQUAL
On entry, all elements of x are equal.
NE_BAD_PARAM
On entry, argument value had an illegal value.
NE_INT_ARG_GT
On entry, n=value.
Constraint: nvalue.
NE_INT_ARG_LT
On entry, n=value.
Constraint: n3.
NE_INTERNAL_ERROR
An internal error has occurred in this function. Check the function call and any array sizes. If the call is correct then please contact NAG for assistance.
NE_NON_MONOTONIC
On entry, elements of x not in order. x[value]=value, x[value]=value, x[value]=value.

7  Accuracy

There may be a loss of significant figures for large n.

8  Parallelism and Performance

Not applicable.

9  Further Comments

The time taken by nag_shapiro_wilk_test (g01ddc) depends roughly linearly on the value of n.
For very small samples the power of the test may not be very high.
The contents of the array a should not be modified between calls to nag_shapiro_wilk_test (g01ddc) for a given sample size, unless calc_wts is reset to Nag_TRUE before each call of nag_shapiro_wilk_test (g01ddc).
The Shapiro and Wilk's W test is very sensitive to ties. If the data has been rounded the test can be improved by using Sheppard's correction to adjust the sum of squares about the mean. This produces an adjusted value of w,
WA=W xi - x- 2 i=1n xi=x- 2 - n-1 12 ω2 ,
where ω is the rounding width. WA can be compared with a standard Normal distribution, but a further approximation is given by Royston (1986).
If n>5000, a value for w and pw is returned, but its accuracy may not be acceptable. See Section 4 for more details.

10  Example

This example tests the following two samples (each of size 20) for Normality.
Sample
Number
Data
1 0.11, 7.87, 4.61, 10.14, 7.95, 3.14, 0.46, 4.43, 0.21, 4.75, 0.71, 1.52, 3.24, 0.93, 0.42, 4.97, 9.53, 4.55, 0.47, 6.66
2 1.36, 1.14, 2.92, 2.55, 1.46, 1.06, 5.27, -1.11, 3.48, 1.10, 0.88, -0.51, 1.46, 0.52, 6.20, 1.69, 0.08, 3.67, 2.81, 3.49
The elements of a are calculated only in the first call of nag_shapiro_wilk_test (g01ddc), and are re-used in the second call.

10.1  Program Text

Program Text (g01ddce.c)

10.2  Program Data

Program Data (g01ddce.d)

10.3  Program Results

Program Results (g01ddce.r)


nag_shapiro_wilk_test (g01ddc) (PDF version)
g01 Chapter Contents
g01 Chapter Introduction
NAG Library Manual

© The Numerical Algorithms Group Ltd, Oxford, UK. 2014