A way to obtain a resolution IV design based on two designs of resolution III (R_III). Used when confirmation runs from a R_III design differ substantially from their prediction, and the experimenter wants to de-alias the two-way interactions from the main effects.