I think there may be a relationship I’m missing and wondering if someone can help me get it.
With respect to the Durbin-Watson test, it appears that sometimes we relate the test statistic to the number 2 (higher than = negative correlation, lower than = positive correlation) while at either times we assess the test statistic against the range in the tables (which are all less than 2). What’s different and why would we use one method in some circumstances?