For reset tests the null hypothesis is no effect of higher order (>1) terms in the RHS of your model. Have you plotted your data.? This should show some such effects., If they exist. Now two different tests of the same hypothesis will likely depend on somewhat different assumptions otherwise there should be no difference. If you now consult google or other references you should find out the differences between the two tests. This and/or the plots should help you choose the appropriate test.to use. Now I must ask are you looking at these tests just because they appeared on your printout or were you concerned about possible assumption violations.? If the first what do the p-values tell you about anything I would suggest nothing. If you should now assumption worries choose the. more appropriate test for that assumption check. Remember multiple tests can change decision probabilities. This means do not use a test just because it pops up on the output. Thinking rather than robot like aaproaches is valuable in data analyses.

Good luck, D. Booth

F-Test, because the test statistic includes an estimate of the (residual) variance that is independent of the expectation. The notations is somewhat confusin, however: the F-test is also a LR-test, based on the profie LR (profiling out the variance). I suspect that the LR test mentioned is simply using the LR obtained for the variance fixed at its MLE.

I agree mostly with Jochen but the I suspect means he has not closely checked. YOU CHECK YOUR OWN RESEARCH. YOU ARE RESPONSIBLE FOR YOUR OWN RESEARCH not someone else. Best wishes for a safe journey, David Booth

David Eugene Booth Thank you a lot for your kind reply and explanation. I really appreciate that. I will follow your advice for sure :)

Jochen Wilhelm Many thanks dear for your kind reply and valuable interpretation which I found very useful. It is highly appreciated to get answers from influential researchers like you.