Critical values for the Durbin-Watson test in large samples

Abstract

This paper estimates functional relationships between the Durbin-Watson critical bounds and the sample size. The functions estimated provide a quick and easy way to calculate critical values for sample sizes not reported as standard in books of statistical tables. They also allow estimation of the sample sizes necessary for the convergence of the Durbin-Watson lower and upper bounds to those derived from the normal distribution and each other. Our results indicate that what constitutes a ‘large’ sample is highly sensitive to the number of independent variables in the regression equation.