Quantitative/ Research design
Factorial designs
Cross manipulations to expose interactions
Factorial designs combine levels of two or more factors so their main effects and interactions can be estimated. Factors can be between participants, within participants or mixed across both. An interaction asks whether one factor’s effect changes across another’s levels. Feasibility and interpretability can deteriorate rapidly as factors and levels multiply.
WHEN IT FITS
Use when the question concerns combinations of interventions or whether an effect depends on context. Ensure the relevant factor combinations are feasible and that interaction estimates can be precise enough to interpret.
Strengths
- Tests combinations and effect heterogeneity directly
- Shares experimental resources across several questions
Limitations
- Many cells increase recruitment and interpretation demands
- Main effects can conceal crossed or context-specific patterns
Know the boundary
A factorial design does not mean every factor is manipulated or randomised; causal claims depend on each factor’s assignment.
USED ACROSS
PsychologyEducationBusiness & MBAComputer science