(Please note that this new version 06-28-2018 supports By groups. Otherwise, it is the same as before.)

JMP provides the Shapiro-Wilk test in the Distribution platform for a departure from a normal distribution. This test has been shown to be more powerful than alternative tests, including the Anderson-Darling test (1). (Note that all such tests have low power with small samples.) Still, you might have a precedent for using the A-D test for normality so this add-in implements it. The test estimates the mean and the standard deviation of the normal distribution by default. Either one of these parameters may instead be specified if known for a test against a particular normal distribution. The test results are appended to a Distribution platform.

Simply open the add-in file to register it with JMP before use. Then, open the data table with the data column to be tested and select **Add-Ins** > **Anderson-Darling Normality Test**. Select the data column and click **Y, Response**. Optionally, select the column with the group identifiers and click **By**. If the mean is known, then select the **Known** radio button for the **Mean** and enter the value in the box provided. The same is true for the standard deviation.

Click **OK**. An alert window will briefly appear for each sample. The *p*-value for the A square test statistic is computed by a Monte Carlo simulation of 100,000 samples under the null hypothesis.

The test report is seen at the bottom of the window. The A square and adjusted A square test statistics are reported. The approximate *p*-value for the second test statistic is computed from a set of four interpolating functions (2).

**Note** that a previous version (before June 5, 2014) of this test as a script computed the *p*-value for A square **incorrectly** and should not be used. Please replace that version with this add-in if you downloaded the version before the correction became available.

**References**

(1) Razali, Normadiah Modh, and Yap Bee Wah (2011) *Power Comparisons of Shapiro-Wilk, Kolmogorov-Smirnov, Lilliefors and Anderson-Darling Tests*, Journal of Statistical Modeling and Analytics, **2**(1)21-33.

(2 ) R.B. D'Augostino and M.A. Stephens, Eds., 1986, *Goodness-of-Fit Techniques*, Marcel Dekker.