Matching the ruggedness test to an efficient design

It is unlikely that any ruggedness test could justify the outlay of more than 30 experiments, in fact it is rare that more than 20 experiments would be carried out. 

If a method of analysis is fast or can be fully automated and requires the testing of few factors (three or less) then the larger designs can be considered. Good choices are central composite designs, or if a linear factor response is expected a full factorial design at two levels. 

As can be seen from the discussion above the most critical limitation is the number of factors that need to be tested. A complex HPLC method may require 10 factors to be tested at three levels and thus a reflected Plackett-Burman design would provide an efficient solution with 23 experiments. 

However, a simple UV spectroscopic method may require the testing of five factors and thus an eleven experiment star design would be sufficient. 

THE NUMBER OF EXPERIMENTS REQUIRED FOR EXPERIMENTAL DESIGNS WITH VARIOUS k VALUES 

---