Hi All-- Hope you are having a great December! (Quite cold here in KY)
Just wanted to check my thinking on this. Customer wants to run an experiment. He wants one factor to be "Polymer" (Categorical, 10 levels) and the other being Tackifier (Categorical, 6 levels). He will blend one Polymer with one Tackifier for any given test, always using 90% polymer and 10% tackifier.
The question is: Can he get away with running fewer than 10*6=60 tests? (He can afford 20.)
In my thinking, if he assumes that a given tackifier behaves the same across all polymers, he might be able to reduce the test size. But if there is any chance of an interaction between a given polymer and a given tackifier that would strongly influence responses, he runs the risk of missing it if he runs a subset of all combinations.
Thoughts? and Thanks!
