In this week’s issue of Tips & Tricks, we use the NumXL package to demystify the normality test, a common diagnostic tool in econometric and time series analysis.
Using the Analysis Pack Add-in in Excel, we generate 5 series of random numbers, each from a different distribution. Then we apply three major normality tests – the Jarque-Bera, the Shapiro-Wilk, and the Anderson-Darling – to examine the sensitivity of each test in detecting deviation from normality for different sample sizes. The exercise highlights the pros and cons of each test and provides a useful introduction to using the Normality Test function in NumXL.
You can find more details and a step-by-step tutorial, along with a downloadable spreadsheet and PDF, at the following link: Facts and Myths about Normality Test
