The video shows how to do a Chow test for regression stability with NumXL 1.60 in Microsoft Excel
Hello and welcome to NumXL 1.6 tutorial. In this video I'll demonstrate how to do a CHOW test for a regression stability. I'll use a sample data gathered from 20 different salespeople.
This regression model will attempt to explain and predict a person's weekly sales which is a dependent variable. This will be done using two explanatory variables, intelligence, IQ and extraversion.
First select a cell in your worksheet where you want the analysis output to be located. Next, find the statistical test icon in the NumXL tab and from the drop down menu click on regression stability test.
A regression stability test wizard pops up. For each data set select a cells range for the dependent variable Y and explanatory variable X. Let's start with data set one. Then select data for data set two.
For the intercept or the constant value, if your model forces the intercept to be a fixed value then enter it here, otherwise leave it blank to indicate that it is a parameter calculated from the data set.
Once the data sets are selected the options and missing values tabs become available. Click on the options tab.
The options tab shows which methods are supported, currently we implement CHOW test.
In the missing values tab leave the default selection which removes observations with any missing value in X or Y. Now click OK, to display the test results.
And there you go, the table shows the test stat score p-value and compares it against a significance level. In this example the regression model seems to be stable over the sample data.
That is it for now, thank you for watching!