Im working on panel regression for school, using 45 cross-sectional (countries) over a 3 year period ('05,'10,'15), 118 observations and 5 independent variables (including const).
When running the regression on gretl, I found the durbin-watson equal to 1.817, meaning there is positive autocorrelation right?
But when I looked up the critical values (dl=1.63029;du=1.78873), I found the durbin-watson is greater than 'du'. Does this mean there is no autocorrelation? If there is autocorrelation, how can I fix my model to prevent autocorrelation?