a type of experimental design in which some conditions are omitted, such that not all levels of one or more of the independent variables are combined with all other levels of the other variables. A Latin square is an example. Fractional factorial designs might be used because of a small overall sample available for study or because of difficulty obtaining participants or assigning them to some conditions. Additionally, there may simply be too many combinations to study (e.g., in a 3 × 4 × 2 × 5 design, there are 120 combinations of levels). Also called incomplete factorial design. Compare complete factorial design.