— A Curated Q&A on Experimental Design, Analysis, and Reproducibility
Free webinar (via Google Meet)
Up to 1,000 registered participants
First 500 to log in will have interactive access (full participation)
Additional participants will join in view-only mode
Recording available exclusively to registered participants
Certificates of attendance provided to all live attendees (interactive and view-only)
Dr. Carlos O. S. Sorzano is a senior researcher and statistician at the National Center of Biotechnology (CSIC, Spain) with more than 15 years of experience teaching and applying statistical experimental design in biomedical and animal research.
His core expertise spans rigorous hypothesis definition, randomization, blocking strategies, sample-size and power calculations, variance estimation, identifying pseudoreplication, and ensuring the correct matching of statistical analysis to experimental design.
He is the author of Statistical Experiment Design for Animal Research (2025), has served as an active member of an animal research ethics committee, and participated in the FELASA Working Group on Experimental Design in Education and Training.
Throughout his career, Dr. Sorzano has placed particular emphasis on the 3Rs—especially Reduction—by developing statistical frameworks that yield robust, reproducible scientific conclusions with the minimum number of animals compatible with research objectives.
The most critical statistical questions in animal research arise long before data collection begins: What is the true experimental unit? How many animals are genuinely required? How do we handle blocking, repeated measures, or missing values? And what can still be rescued when an experimental design was not ideal?
This interactive, attendee-driven session is structured around your real-world statistical and design challenges. Participants are invited to submit their specific questions in advance, shaping a tailored discussion that bridges statistical theory with practical laboratory application. Key focal areas include sample-size and power calculations, true randomization, blocking, identifying pseudoreplication, selecting statistical models, interpreting p-values and confidence intervals, and distinguishing exploratory from confirmatory research.
Rather than delivering a dry catalogue of statistical tests, Dr. Carlos O. S. Sorzano demonstrates how sound statistical reasoning enhances experimental design, maximizes scientific reproducibility, and upholds the 3Rs (Reduction).
Selected references:
Sorzano, C. O. S. (2025). Statistical Experiment Design for Animal Research. Springer/CSIC
Fry D, Berdoy M, Forni M, Sorzano COS, Steckler T, Franco NH (2025). Teaching experimental design: outputs from the FELASA Working Group. Laboratory Animals. 59(5):614–629
How do I identify the experimental unit: the animal, the cage, the litter, the batch, or something else?
How many animals do I really need, and what information is required to estimate the sample size?
I’ve already collected the data. How can I check whether my planned statistical analysis matches the study design?
Distinguish true N (animal, cage, litter, batch) to eliminate pseudoreplication
Estimate sample sizes using effect size, variance estimates, and statistical power
Apply blocking factors (sex, litter, batch) to control variance and improve precision
Match statistical models (ANOVA, mixed-effects, non-parametric) correctly to experimental designs
Apply appropriate statistical inference and P-value interpretations based on study objectives
Statistics starts before data collection; faulty design cannot be rescued by statistical testing
Misidentifying replicates as true N inflates false-positive rates and ruins reproducibility
Blocking and standardizing procedures reduce required animal numbers more effectively than simple sample-size formulas
Maintain a strict distinction between exploratory data mining and confirmatory hypothesis testing
Repeated measures, batch effects, and missing data require dedicated modeling rather than standard t-tests.