Since the experimental values are calculated by numerical simulation, the values of each repeated test are the same. Therefore, when performing response surface analysis in Design-Expet, the 'Lack of Fit' item does not report p and F values, nor does it report significant or insignificant. But the p-value of the model shows that the model is significant. In this case, can we assume that the model is valid?
I don't think I got really what you meant! Is it that you had nothing as "the lack of fit" ??
How is that possible?
You will either get an insignificant or a significant one.
Please, explain more your problem, and feel free to text me in private if you want.
Anyways, the lack of fit means the difference between the predicted values by the model and the measured values, so it has to be "non-significant" so that we can rely on the model to give "the most real" values.
This is a screenshot of 'Analysis of Variance' in Design-Expert. The column 'Lack of fit' does not have F and p values.
Aicha Akouz Thank you so much. I have upload a screenshot about 'Analysis of Variance' in Design-Expert software. The 'Lack of fit' does not have any values, and not report 'significent' or 'insignificent'.
All objecive values are obtained by simulation. So the values of the repeated experiments are the same. Is the absence of 'Lack of fit' caused by this?
Dear Jiajie Yang,
Sorry for the late reply.
I agree with your explanation, I also think it might be due to the simulation.
I have sent you a private text to discuss this case.
Kind regards
- Badges
- Science topic