The Stack Overflow podcast is back! Listen to an interview with our new CEO.

Experimental designs that use only a subset of possible factor combinations. Typically, some factors are intentionally confounded w/ some interactions, making the interactions impossible to estimate, but requiring a smaller N for the study.

Fractional factorial designs are experimental designs that use only a subset of possible factor combinations. Typically, some factors (i.e., main effects) are intentionally confounded with some interactions. This makes the confounded interactions impossible to estimate, but allows equivalent statistical power to detect the effects the experimenter does care about with a smaller $N$ for the study.

history | excerpt history