Believe nothing and be on your guard against everything.
The full-factorial approach to experimentation covers all combinations of factors, providing valuable information on interactions. However, the number of experimental runs increases rapidly, even if you test the factors at only two levels each. Fortunately, by cutting back to a “fractional factorial,” you may find it possible to study more factors within a given experimental budget. Table 5.1 shows how few runs we suggest you pare back to for four to seven factors (see Appendix 2 for details on layout.). These particular designs are especially good for “screening” the main effects of many factors in search of the vital few or more in-depth ...