Performing the Durbin-Watson autocorrelation test
The Durbin-Watson autocorrelation test determines whether past data influences the present so that we can make predictions with it. This test returns a value interpretation, depending on its range. The function returns a value between 0
and 4
. The interpretation is outlined here:
- The value is 2 – there is no autocorrelation.
- The value is between 0 and 2 – it has positive autocorrelation, meaning that the relationship is growing over time. This is common in time-series data.
- The value is between 2 and 4 – it has negative autocorrelation. The values are decreasing over time.
The test returns the following results:
- Accept the
null
hypothesis and that the data has no autocorrelation. - Reject the
null
hypothesis and accept the alternative hypothesis that the data has autocorrelation.
We are going to perform the Durbin-Watson test step by step and analyze its value. The formula...