Assume a software is providing you with 1st order autocorrelation post-regression test output such that DurbinWatson statistic = 2.19 and p-value = 0.02. This result asserts with sufficiently high confidence that there is:

positive autocorrelation

or no autocorrelation

or negative autocorrelation

or none of the previous answers is valid