I am using Box-Behnken Design in RSM for optimization. In the ANOVA for response surface Quadratic model I am getting pure error 0 ie sum of squares and mean square is zero due to which there is no p value and F value for LOF test.Is it correct?
Dear Shilpi,
I think the reason of this result is that the value Sum of Squares (SS) of regression in your model is larger so much than SS of the residual. In the residual, there are two parts: lack of fit error (LOF) and pure error. Therefore, in your problem, SS and mean squares (MS) of pure error are zero, I think you should not worry because these values are very small and negligible. It depends on your design by using RSM as well as results of your experiments. I read some publications that both of SS and MS are zero, that means there is no P-value and F-value for LOF test.
For more details, you can read in following publication
Good luck
Dang
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA