To conduct the Chow Test for regression stability, we need to define our sample subsets.
For this tutorial, we selected the first nine (9) observations as our first data set, and the last ten (10) observations as our second data set.
Now we are ready to conduct the test. First, locate the statistical test icon in the NumXL tab and click on the “Regression Stability” item in the drop‐down list.
- Select an empty cell to store the Regression stability test output table
- locate the statistical test icon in the NumXL tab and click on the “Regression Stability” item in the drop‐down list.
- The regression stability wizard (form) pops up.
- Select the cell range for the dependent variable and the explanatory variables for data set 1, and do the same for data set 2.
Note: In this case, the regression intercept or constant will be estimated from our data. In other cases, where the intercept is fixed to a given number (e.g. zero), then we should put this number there.
- Selecting the input data cell range will enable the “Options” and “Missing Values” tabs. Click the “Options” Tab.
- In the options tab.
- If your data include one or more intermediate observations with missing values, click the “Missing Values” tab.
- Initially, missing values are dropped from the sample data. If you wish to change this behavior/treatment, select an alternative method.
- Now, Click “OK” to generate the test output table.
The Chow test (regression stability test) wizard generates the output table as shown below:
This regression stability test function and wizard are available starting with version 1.60.