Now Playing: Normality Tests (2:16)
The Normality Tests procedure in PASS is used to determine sample size or evaluate power for eight different Normality Tests.
In this video, we will examine the power of the Shapiro-Wilk Normality test to indicate non-Normality for an Exponential distribution, for various sample sizes.
We begin by opening the Normality Tests procedure in PASS.
We wish to solve for power, based on the Shapiro-Wilk Normality test.
We will use one thousand simulations for the sake of time, although ten thousand or more should probably be used in practice.
We set Alpha to 0.05.
We will examine samples sizes ranging from 5 to 50.
The data distribution is set to Exponential, with a mean of 10.
The Calculate button is pressed to generate the output.
The numeric report shows the powers that are generated for each sample size.
An additional report shows the power values for each of eight Normality tests.
The first power curve shows that the power quickly increases until the sample size is 20, after which the improvement in power is much less dramatic.
The second power curve chart shows a comparison of the power of the Shapiro-Wilk test compared to other common Normality tests. The Shapiro-Wilk test tends to be the most powerful.
Thank you for watching this brief introduction to the Normality Tests procedure in PASS. For more details, please download and install the free trial of PASS.