Hello everyone,

I would like to investigate three factors using a central composite design.

Each factor has 5 levels (+1 and -1, 2 axial points and 1 center point).

I chose my high and low levels (+ and -1) based on a screening DoE I did previously using the same factors.

I chose an alpha of 1.681 for the axial points because I would like my model to be rotatable. However, for one of the three factors, one of the axial points is outside the feasable range (negative CaCl concentration....). I thought of increasing my low level for this factor to avoid this. Lets say, increasing the value from 0.05 to 0.1 to avoid reaching the negative range with the axial point, but I was wondering, if this would effect the reliablity of my model?

Another option would be to change the design to one that has no axial points outside the design points. However, this is actually my area of interest.

Can anyone help?