# Fractional factorial designs: lower $p$ vs. more replications

If you were to choose between a 2(X-1) with only one replication or a 2(X-2) with two replications, what would be preferred and why? The goal is to both predict the values of the effect parameters as well as possible, as well as the significant effects.

• That cannot be answered without knowing more about the specifics, and the goal of this experiment. Higher $p$ if you dont know which variables are relevant, so you need a screening experiment. – kjetil b halvorsen Dec 15 '16 at 9:04