Crossover Solomon Four-Group Design
Crossover Solomon Four-Group Experimental Design · Also known as: crossover S4G design, within-subjects Solomon design, repeated-measures Solomon four-group design
The Crossover Solomon Four-Group Design merges two powerful experimental strategies: the Solomon four-group design's control for pretest sensitization and the crossover design's within-subjects efficiency. Participants are randomly assigned to one of four groups that vary in whether they receive a pretest and in the sequence of treatment and control conditions, allowing the researcher to simultaneously estimate treatment effects, pretest effects, and their interaction while controlling for individual differences through repeated measurement.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use this design when you need to detect whether the act of pretesting sensitizes participants to the treatment — and you also want the statistical power benefits of a within-subjects crossover. It is best suited to stable traits or conditions where a meaningful washout period can plausibly return participants to baseline before the second condition. Good candidates include educational intervention studies, psychological training programs, and clinical trials of reversible symptom outcomes. Do NOT use it when carryover effects are irreversible (e.g., learning or surgical interventions), when the study condition produces lasting change that prevents a true return to baseline, when attrition across two phases is likely to be high, or when the logistical burden of two full treatment phases is prohibitive. For straightforward pretest-sensitization control without within-subjects repetition, use the standard Solomon four-group design instead.
Strengths & limitations
- Controls for pretest sensitization (testing effect) while gaining within-subjects efficiency — a combination unavailable in simpler designs.
- Each participant serves as their own control, reducing variance due to individual differences and lowering required sample size relative to a pure between-subjects Solomon design.
- Generates rich information: treatment effect, pretest effect, pretest-by-treatment interaction, sequence effects, and period effects are all estimable.
- Supports strong causal inference through randomization of both group assignment and condition sequence.
- Useful when sample size is limited but pretest sensitization is a genuine methodological concern.
- Assumes no or negligible carryover effects between conditions — a washout period helps but cannot guarantee this, and violation invalidates within-subjects comparisons.
- More complex to implement and analyze than either the standard Solomon design or a simple crossover design taken alone; requires specialized mixed-model analysis.
- Participant burden is high: individuals must complete two full treatment phases plus measurements, increasing dropout risk.
- Statistical power benefits materialize only if between-person variability is large relative to within-person variability; if individual differences are small, the crossover advantage shrinks.
- Rarely used in practice — limited published guidance and examples make implementation decisions harder to benchmark against existing studies.
Frequently asked
What is the key difference between a standard Solomon four-group design and this crossover version?
In the standard Solomon four-group design, each participant is assigned to exactly one group and receives either the treatment or the control — it is a purely between-subjects design. The crossover version has participants rotate through both the treatment and control conditions across two sequential phases, making it a within-subjects (repeated-measures) variant. This reduces required sample size but introduces the risk of carryover effects and increases participant burden.
How do I determine whether a washout period is long enough?
The required washout duration depends on the mechanism of the intervention. For pharmacological treatments, it is typically several half-lives of the drug. For behavioral or psychological interventions, the washout must be long enough that any practice effects, mood shifts, or attitude changes from the first period have returned to baseline. Pilot data and substantive theory about the intervention's persistence are the best guides. If a genuine washout is implausible, do not use a crossover design.
How large a sample do I need?
Because each participant contributes observations in both conditions, the required total sample is generally smaller than for a between-subjects design of equivalent power. However, the four-group structure means the sample must be divided four ways, so each cell should have sufficient observations for stable estimation — typically at least 10–15 per cell as a rough minimum, yielding 40–60 total participants. A formal power analysis using a mixed-model framework with estimates of within-person and between-person variance is strongly recommended.
What statistical model should I use?
A linear mixed-effects model (LME) is the most appropriate and flexible approach. Fixed effects should include: treatment condition, period (first vs. second), sequence group (AB vs. BA), pretest exposure (yes vs. no), and the pretest-by-treatment interaction. A random intercept per participant accounts for the repeated-measures structure. If carryover is suspected, add a carryover term (first-period treatment condition as a predictor of second-period outcome). Specialized ANOVA for crossover designs can also be used but is less flexible for unbalanced data.
Is this design used in practice?
The full crossover Solomon four-group design is used infrequently, primarily because of its logistical complexity and the demanding conditions required (reversible outcomes, feasible washout, low attrition risk). Researchers more commonly use either the standard Solomon four-group design or a standard crossover RCT, but not both simultaneously. When both pretest sensitization and individual-differences control are genuine priorities, the combined design is theoretically well-motivated.
Sources
- Solomon, R. L. (1949). An extension of control group design. Psychological Bulletin, 46(2), 137–150. DOI: 10.1037/h0062958 ↗
- Braver, M. C. W., & Braver, S. L. (1988). Statistical treatment of the Solomon four-group design: A meta-analytic approach. Psychological Bulletin, 104(1), 150–154. DOI: 10.1037/0033-2909.104.1.150 ↗
How to cite this page
ScholarGate. (2026, June 3). Crossover Solomon Four-Group Experimental Design. ScholarGate. https://scholargate.app/en/experimental-design/crossover-solomon-four-group-design
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Control Group Experimental DesignExperimental design↔ compare
- Crossover Pretest-Posttest Experimental DesignExperimental design↔ compare
- Crossover Randomized Controlled TrialExperimental design↔ compare
- Pretest-Posttest Experimental DesignExperimental design↔ compare
- Solomon Four-Group DesignExperimental design↔ compare