Stine Fangel, Statistician, Bavarian Nordic
Often we are setting variables to hard-to-change and very-hard-to-change when complete randomisation of experiments is not possible. When defining variables as hard-to-change and very-hard-to-change, we randomise where possible while ensuring valid testing of conditions. I will provide examples where the design from the JMP DOE platform is used as-is and examples on how the design can be altered to meet requirements such as balanced design (equal number of runs for each treatment) and blocking of a hard-to-change variable.
