restrict attribute levels for particular alternatives

[bblachly]

Hi,
Thank you for the useful package. Your 2020 paper in J. Statistical Software suggests it is possible to restrict the attribute levels that appear in a particular (labelled) alternative. However, I cannot find an example. Can you help?
Many thanks,
Ben

[traets]

Hi Ben,
For now you cannot restrict labeled alternatives (I am working on it).
You can restrict unlabeled alternatives. The easiest way to do this would be to remove undesirable profiles (= alternatives) from the candidate set (generated with the "Profiles" function). Then you can use the Modfed algorithm to optimize the design based on the restricted candidate set.

Kind regards,
Frits

[tallysfeldens]

Dear Dr Traets,

Thank you so much for the work put into this package development.
I have an issue here.
I could successfully constrain one of the attributes from my labelled design experiment by using the option constraints. However, it is a case where my attributes have a fixed level for one alternative and varying levels for the other. Although it worked for the design, I later found it unsuitable for estimation due to collinearities with my label design. Do you recommend a workaround to this type of problem? Thank you so much,