In the pooling contribution technique that are used to analyze the results in ANOVA technique, what does it mean if we get a negative contribution? Is it possible to get such a negative contribution? If yes, then what does it indicate?
In most cases you are violating ANOVA's assumptions, your estimates are biased, your model is wrong, your model is incomplete, or all of the above. Check the publications attached as well as the link. Without knowing your problem and model is difficult to say something about it. May be try with ML or REML estimation instead.
The results of Test F is always positive . You can obtain negative results using multiple comparison or using contrasts to detect diferences between groups, meaning that the 2nd group or the 2nd mean is superior to the first one, becasue they are t tests.
Helena, I agree that Jose needs to be more specific as to what is meant by a negative contribution. The assumption for ANOVA is that the variables are normally distributed. Please let us know what parameters are being contrasted and how they are being contrasted.