Regression fitted values

#1
I'm working with a dataset that I'm using to model total biomass (which was a variable created by adding three components of biomass). I also modeled each of these components as a function of the same variable.

I am comparing the predicted value of the total biomass model with the sum of the predicted values of the components. I wasn't expecting the values to be equal, but I am not sure why the values aren't the same.

Does it have to do with the error associated to the models?

Thanks for any help!
 

noetsi

No cake for spunky
#2
That is a possiblity. It would think its possible that outliers in the individual component data sets with high leverage might distort their regression lines, but might not influence calculation of the overall total biomass the same way. But that is wild guess....

Of course that assumes you used regression. I am not sure what method you utilized.
 

noetsi

No cake for spunky
#4
I am not sure how non-linear models influences the effect of outliers or leverage (my assumption is they don't because discussions of these in the context of regression don't suggest that non-linear relationships change the basis reality here) but you might check and see if indeed there are high leverage points in the modeling of the the subsets of biomass that don't have signficant effect on your estimates of overall biomass.