Double-Blind Pretest-Posttest Experimental Design
Also known as: DB-pretest-posttest design, double-blind pre-post design, masked pretest-posttest RCT, double-masked pre-post experiment
The double-blind pretest-posttest experimental design is a true experiment in which participants are randomly assigned to treatment and control conditions, outcome data are collected both before and after the intervention, and neither participants nor outcome assessors know which condition each participant received. Combining baseline measurement with strong blinding, the design controls for both pre-existing group differences and expectancy-driven bias, making it a gold-standard approach in clinical and behavioral research.
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 establish causal efficacy of an intervention, participants can plausibly be blinded to their condition, and baseline measurement is feasible. It is ideal for clinical drug trials, behavioral interventions, educational programs, and any setting where expectancy effects or pre-existing group differences threaten validity. Do not use it when blinding is impossible (e.g., surgical technique comparisons where patients inevitably know their procedure), when there is no practicable control condition, when the pretest measurement itself might sensitize participants and alter their response to the intervention (testing effects — consider the Solomon four-group design instead), or when the outcome cannot be meaningfully measured before the intervention begins.
Strengths & limitations
- Controls pre-existing group differences by using each participant as their own baseline, increasing statistical power.
- Double-blinding eliminates expectancy and performance bias from both participants and assessors, protecting internal validity.
- Randomization plus blinding together represent the strongest available design for establishing causal inference.
- Change-score or ANCOVA analysis is more sensitive than post-only comparison, reducing required sample size for equivalent power.
- Pretest data allow verification of randomization success and enable subgroup analyses by baseline severity.
- Blinding is not always achievable: active treatments with distinctive side-effects or obvious physical changes may inadvertently unblind participants.
- The pretest may sensitize participants to the outcome measure, producing testing effects that interact with the treatment (threatening external validity).
- Attrition between pretest and posttest can introduce bias if dropout is differential across groups.
- Implementing and verifying double-blind conditions requires careful logistical planning and quality-checking throughout the study.
Frequently asked
Is ANCOVA always better than a change-score analysis for this design?
ANCOVA with the pretest as covariate is generally preferred: it accounts for regression to the mean, is more powerful than a raw change-score t-test when the pretest-posttest correlation is below 0.5, and handles residual baseline imbalance. However, when the pretest-posttest correlation is very high (above 0.8) and groups are well-balanced at baseline, change-score analysis and ANCOVA yield similar results. Pre-register the chosen analysis to avoid outcome-reporting bias.
What should I do if blinding is broken for some participants?
Report the unblinding events transparently in the manuscript. Conduct a sensitivity analysis excluding unblinded participants or using a per-protocol analysis alongside the intention-to-treat analysis to assess whether the results are robust. If unblinding is widespread (e.g., more than 20% of participants correctly identified their assignment), treat the blinding claim with caution in interpretation.
How does testing effect threaten this design, and how can I address it?
A testing effect occurs when familiarity with the pretest instrument causes improvement on the posttest regardless of the intervention, inflating apparent change in both groups. Because both arms experience the same pretest, the effect is usually symmetric and cancels in the group comparison — but it can interact with treatment (treatment by prior-testing interaction). To detect this, add a no-pretest control group (creating a Solomon four-group design) or use alternate forms of the instrument at pre and post.
Can I use this design with a within-subjects (crossover) structure?
Yes, but the two structures solve different problems. A crossover design controls for between-person variance by having the same person receive both conditions across periods; the pretest-posttest design controls for within-person change over time. Combining them (crossover with pre- and post-period measures) is possible but requires careful handling of carryover effects and period effects in the analysis.
What sample size considerations are specific to this design?
Sample size calculations should be based on the expected change score (mean difference in posttest-minus-pretest between groups) and the standard deviation of change scores rather than of post scores alone. Because within-person change scores are typically less variable than raw scores, smaller samples can achieve the same power — the advantage grows as the pretest-posttest correlation increases. Use the ANCOVA-adjusted formula for the most accurate estimate.
Sources
- Campbell, D. T., & Stanley, J. C. (1963). Experimental and quasi-experimental designs for research. In N. L. Gage (Ed.), Handbook of Research on Teaching (pp. 171-246). Rand McNally. link ↗
- Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and Quasi-Experimental Designs for Generalized Causal Inference. Houghton Mifflin. ISBN: 978-0395615560
How to cite this page
ScholarGate. (2026, June 3). Double-Blind Pretest-Posttest Experimental Design. ScholarGate. https://scholargate.app/en/experimental-design/double-blind-pretest-posttest-experimental-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
- Pretest-Posttest Experimental DesignExperimental design↔ compare
- Randomized Controlled TrialExperimental design↔ compare
- Single-blind pretest-posttest experimental designExperimental design↔ compare
- Solomon Four-Group DesignExperimental design↔ compare