Contents - Index

Include any quadratic terms in the model


This functions becomes available after multiple linear regression. The following example is the multiple regression results using yield (YLD) as dependent variable. It indicates that kernel weight (KW), lodging, days to heading, and days to maturity were variables significantly associated with yield, and they altogether explained nearly 36% of the yield variation.



Under the main menu Association then Multivariate, click Include Quadratic Terms. The biplot becomes:



A single quadratic term KW*KW is added to the model, which explained an additional 4% yield variation. This indicates that although there is strong linear relationship between KW and yield, extremely large KW may lead to lower yield. 


The computing process of adding quadratic terms to the model is similar to that of adding interaction terms to the model.