What is the range of the Durbin-Watson statistic?

The Durbin-Watson statistic is designed to test for autocorrelation in the residuals of a regression analysis. Its value typically ranges from 0 to 4. A value around 2 suggests that there is no autocorrelation; values below 2 indicate positive autocorrelation, while values above 2 suggest negative autocorrelation.

The range of 0 to 4 is derived from the fact that the statistic is calculated based on the differences between consecutive residuals, and extremes in positive or negative autocorrelation can lead to values closer to 0 or 4, respectively. This comprehensive range allows analysts to assess the degree of autocorrelation and diagnose issues that may affect model performance.