Hello! I've been reading the example posted on your website regarding design evaluation of D-optimal and I-optimal designs (https://www.jmp.com/support/help/en/16.2/index.shtml#page/jmp/comparison-of-a-doptimal-and-an-ioptimal-response-surface-design.shtml)
Somewhere in that post, it compares the relative prediction variance of both designs and states: "...This means that the relative standard error is 0.732 for the D-optimal design and 0.456 for the I-optimal design..." furthermore, it finishes saying something like "...confidence intervals for the expected response based on the D-optimal design are about 60% wider..."
May I understand how those values (expected response CI and relative standard error (RSE) are computed??)
In the case of the RSE, would you be so kind to share what's the appropriate interpretation of that value?
Thanks and regards!
