Facebook Pixel
Brookbush Institute Logo
  1. Glossary
  2. Crossover Study
Crossover Study - Glossary Term Illustration

Crossover Study

A crossover study is an experimental research design in which the same participants complete each intervention in sequence. The order of interventions is randomized; generally, a washout period separates phases to limit carryover, and participants then switch ("cross-over") to the alternate intervention.

Crossover Study

A crossover study is an experimental research design in which the same participants complete each intervention in sequence. The order of interventions is randomized; generally, a washout period separates phases to limit carryover, and participants then switch ("cross-over") to the alternate intervention. Because every participant receives all conditions, each person serves as their own control.

Semantic Clarification

  • “Crossover”: Each participant switches from one intervention to another within the same study.
  • “Randomized order”: The starting intervention is assigned randomly to reduce sequence bias.
  • “Washout period”: A break between interventions that minimizes lingering effects from the first condition.
  • “Within-subject control”: Each participant acts as their own control, strengthening comparisons and reducing variability.

Historical Perspective

The crossover design emerged from 19th- and early-20th-century agricultural and statistical work on “change-over” experiments (e.g., Lawes at Rothamsted, 1853) and later formal treatments of carryover and period effects (e.g., Cochran, 1939/1941), establishing the logic of applying sequential treatments to the same experimental unit. A clear, peer-reviewed clinical exemplar arrived with Hogben & Sim (1953) in the British Journal of Preventive and Social Medicine, which used repeated treatment periods within the same patients and explicitly discussed wash-in/wash-out and carryover—core features that define contemporary crossover trials.

Applied Example

Included in the course: Acute Variables: Periodization Training

Oliveira et al. (2018) conducted a randomized cross-over study in 23 elite male wrestlers (24.5 ± 3.9 years; injury-free ≥6 months; no ergogenic supplements/drugs). Athletes completed two 3-month training protocols in sequence, separated by a 2-month transition:

  • Non-periodized (NP) protocol: 5 sessions/week; 4 sets/exercise; 8 reps/set; ~90% 1RM; 1 hour/session of wrestling drills to improve strength.
  • Block linear periodization (BLP) protocol: 5 sessions/week; aerobic + resistance training;
    • Month 1: aerobic only (running) ~1 hour/session; no resistance training
    • Month 2: 2 sets/exercise; 20 reps/set; 60% 1RM; running ~45 min/session (before resistance)
    • Month 3: 3 sets/exercise; 10 reps/set; 80% 1RM; running ~30 min/session (before resistance)
      All participants continued wrestling-specific training and used a full-body resistance program (leg press, bench press, lat pulldown, seated row, shoulder press, crunches, hamstring curl).

Findings

  • Isometric handgrip strength increased only during BLP.
  • Peak strength (unspecified exercise) increased significantly and similarly after both protocols.
  • Bench press, lat pulldown, and squat 1RM increased more after BLP.
  • Serum creatine kinase (CK) rose significantly during both protocols; intra-intervention CK at 1 and 2 months was higher in NP.

Reference:
Oliveira, A. L., Sposito-Araujo, C. A., Senna, G. W., Lopes, T. C., Godoy, E. S., Scudese, E., ... & Dantas, E. H. (2018). Comparison of the Matveev periodization model and the Verkhoshansky periodization model. Journal of Exercise Physiology Online, 21, 60–67.

Strengths and Weaknesses

Strengths

  • Each person is their own control. This reduces noise from individual differences and makes true treatment differences easier to detect.
  • Fewer participants needed. Compared with a parallel-group RCT asking the same question, a crossover typically needs a smaller sample.
  • Direct within-person comparisons. Effects are estimated within the same individual, which may simplify the interpretation of results.
  • Order can be randomized. Randomizing the sequence helps reveal or balance period/order effects rather than confounding results.
  • A Good Tool for Certain Scenarios: Works well when the condition is stable and outcomes can be measured repeatedly over short intervals.

Limitations

  • Carryover/learning effects. Residual effects from the first treatment (or skill learning, fatigue) can contaminate the second period if washout is inadequate.
  • Time trends (period effects). If the condition naturally improves/worsens over time, it can bias the comparison across periods.
  • Greater participant burden. Longer total participation per person increases the risk of dropout.
  • Not suitable for lasting effects. Inappropriate when treatments or outcomes are irreversible or long-lasting (e.g., surgery, long-term adaptations).
  • More complex design/analysis. Requires careful planning of washout length, sequence balance, and statistical modeling to avoid biased estimates.
  • Practical/ethical constraints. Scheduling washouts, ensuring adherence, and exposing participants to multiple conditions can pose additional risks.

Frequently Asked Questions (FAQ)

Is a crossover study a randomized controlled trial (RCT)?

  • Yes. A crossover study is typically a type of RCT in which participants receive multiple interventions in a random sequence, with each participant serving as their own control. A “washout” period is usually included between interventions to reduce carryover effects.

What is the difference between a parallel and a crossover study?

  • In a parallel study, participants are randomly assigned to one of the interventions and remain in that group for the duration of the study. In a crossover study, each participant receives multiple interventions in sequence, allowing direct within-subject comparisons. Parallel designs are simpler and avoid carryover, while crossover designs increase efficiency and statistical power by reducing inter-subject variability.

What is a case crossover study?

  • A case crossover study is an observational design used to evaluate the effect of transient exposures on acute events. Each participant serves as their own control, with exposure immediately before the event (case period) being compared with exposure at other times (control periods). For example, researchers might compare whether intense physical exertion occurred in the hours leading up to a cardiac event versus other non-event times.

Why do you do a crossover trial?

  • Crossover trials are valuable because they allow each participant to act as their own control, minimizing inter-individual variability and often requiring fewer participants to achieve statistical power. They are especially useful when studying short-term, reversible effects (e.g., acute physiological responses to exercise or nutrition interventions). However, they are less appropriate for long-lasting or irreversible outcomes.

Brookbush Institute Perspective

Crossover designs are especially valuable in exercise and rehabilitation research, where between-subject variability is high, and the effects of an intervention are not long-lasting or reversible. Using each participant as their own control can sharpen comparisons between protocols. However, interpretation of findings should consider the potential for adaptations that are durable or learning-dependent, which risks contamination across phases.

Discussion

Comments

Guest