Autocorrelation is a correlation derived from a time series variable and its values over a given period. It is useful for identifying patterns and validating assumptions in your hypothesis testing.
The Durbin Watson statistic is a number that tests for autocorrelation in the residuals from a statistical regression analysis.