paula posted on Thursday, April 20, 2006 - 2:43 pm
Hi, I m searching for the best model using Akaike s information criterion and R2 (analysis OLS using the program SAM). The response variable is the species richness which is spatial strutured. I found a model (R2=0.916)using 4 predictors varibles and 2 espatial variables (latitude and longitude in a second order polinomial). The problem is that the R2 of Predictors variables is 0.234 and the R2 of Space is 0.29 and THE SHARE EXPLAINED VARIANCE is -0.392. I can t understand this negative variance!!!! please help me... Thank you very much!!!!!