Crossover Full Factorial Experiment — Within-Subject Full Factorial Design
Also known as: within-subject full factorial design, repeated-measures full factorial experiment, crossover factorial trial, full factorial crossover design
A crossover full factorial experiment combines the efficiency of a crossover (within-subject) design with the comprehensiveness of a full factorial design. Every participant receives all combinations of the factor levels across successive treatment periods, separated by washout intervals, allowing complete estimation of all main effects and interactions while using each participant as their own control.
Key highlights
- Participants serve as their own controls, eliminating between-subject variability from the error term and substantially increasing statistical power.
- Full factorial structure estimates all main effects and all interaction effects without confounding — no information is left unexamined.
- Requires fewer participants than a parallel-group full factorial design to achieve equivalent power.
- Reveals how the effect of one factor depends on the levels of other factors, providing richer mechanistic insight.
- Efficient use of participant time: each person contributes data for every treatment combination.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use a crossover full factorial experiment when: (1) the research question requires estimating all main effects and interactions among multiple factors; (2) the outcome is reversible — the treatment has no permanent effect that would carry over after washout; (3) the participant pool is limited and within-subject comparison greatly increases statistical power; and (4) participants can realistically complete all treatment periods. Ideal in pharmacology, nutrition science, ergonomics, and cognitive psychology. Do NOT use when: treatments have permanent or long-lasting effects (e.g., surgery, vaccines); washout periods cannot practically separate treatment phases; the number of factor combinations is so large that trial duration becomes burdensome or attrition prohibitive; or when carryover effects are suspected to be treatment-period interaction-dependent (non-uniform carryover).
Strengths & limitations
- Participants serve as their own controls, eliminating between-subject variability from the error term and substantially increasing statistical power.
- Full factorial structure estimates all main effects and all interaction effects without confounding — no information is left unexamined.
- Requires fewer participants than a parallel-group full factorial design to achieve equivalent power.
- Reveals how the effect of one factor depends on the levels of other factors, providing richer mechanistic insight.
- Efficient use of participant time: each person contributes data for every treatment combination.
- Carryover effects are a fundamental threat; if a treatment's effect persists into the next period, within-subject comparisons are biased.
- The number of treatment combinations grows exponentially with factors and levels (2^k for two-level factors), making the design impractical with many factors.
- Long total trial duration — all periods plus washouts — increases dropout risk and fatigue effects.
- Statistical analysis is more complex than a simple between-subjects design, requiring mixed models that correctly handle within-subject correlation.
- Cannot be used when treatments produce irreversible outcomes.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
How is this different from a standard crossover design?
A standard crossover design typically compares two (or a few) treatments in sequence. A crossover full factorial experiment extends this by requiring every participant to receive all combinations of multiple factors — a full factorial structure applied within a crossover framework. This allows estimation of interaction effects between factors, not just main effects.
What is carryover and why does it matter so much here?
Carryover occurs when the effect of a treatment administered in one period persists into the next period, distorting the outcome measured under the new treatment. In a crossover design the within-subject comparison is only valid if each period's measurement reflects only the current treatment. With many treatment periods (as in a full factorial crossover), the risk of carryover accumulating across periods is higher, making washout planning critical.
How do I choose the washout length?
Washout should be at least five biological half-lives of the active treatment for pharmacological studies. For behavioural or psychological interventions, washout length is guided by prior evidence about how long the manipulation's effects persist. When in doubt, err on the longer side and include a check for residual effects at the start of each period.
When should I use a fractional factorial crossover instead?
When the number of full factorial combinations makes the total trial duration impractical or dropout risk too high, a fractional factorial crossover design omits some higher-order combinations using a principled aliasing structure. This sacrifices the ability to estimate high-order interactions but retains main effects and lower-order interactions, greatly reducing participant burden.
What statistical model should I use for the analysis?
A linear mixed-effects model is the standard approach. Fixed effects include treatment combination, period, and sequence; participant is included as a random effect. This model correctly accounts for the repeated-measures structure and allows unbiased estimation of all factorial main effects and interactions. If carryover effects are suspected, include a carryover term and interpret with caution.
Sources
- 1.Jones, B., & Kenward, M. G. (2003). Design and Analysis of Cross-Over Trials (2nd ed.). Chapman and Hall/CRC.ISBN 978-1584883429
- 2.Montgomery, D. C. (2017). Design and Analysis of Experiments (9th ed.). Wiley.ISBN 978-1119320937
You have read it. What now?
Cite this page
ScholarGate. (2026, June 3). Crossover Full Factorial Experiment. ScholarGate. https://scholargate.app/experimental-design/crossover-full-factorial-experiment