6.3 ARRIVE 2.0 for In Vivo Animal Research
Key Takeaways
ARRIVE 2.0 is a reporting guideline for in vivo animal research, organized as the Essential 10 and Recommended Set; it does not itself authorize animal or first-in-human research.
The sample-size item requires the exact number of experimental units, total animals, and an explanation of how size was decided, with an a priori calculation reported if one was performed—not a universal power calculation for every study.
Randomization, blinding, inclusion/exclusion criteria, outcome measures, and statistical methods should be reported as actually used, including steps taken to minimize bias and the experimental unit for each analysis.
Pseudoreplication occurs when correlated technical observations are analyzed as independent biological units; the analysis must respect clustering and the true experimental unit.
Ethical review, animal welfare law, the Guide for the Care and Use of Laboratory Animals, institutional policy, and regulatory nonclinical standards are separate from ARRIVE reporting.
ARRIVE 2.0 for In Vivo Animal Research
The ARRIVE 2.0 guidelines help authors report in vivo animal research so readers can assess its reliability and reproduce or build on it. They apply broadly to research involving living animals. They are reporting guidance, not a substitute for ethical approval, animal-welfare law, institutional authorization, nonclinical laboratory standards, or regulatory evidence requirements.
Essential 10 and Recommended Set
ARRIVE 2.0 prioritizes items in two groups.
The Essential 10 cover:
- study design;
- sample size;
- inclusion and exclusion criteria;
- randomization;
- blinding;
- outcome measures;
- statistical methods;
- experimental animals;
- experimental procedures;
- results.
The Recommended Set adds context including abstract, background, objectives, ethical statement, housing and husbandry, animal care and monitoring, interpretation and scientific implications, generalizability or translation, protocol registration, data access, and declarations of interests.
“Essential” means these items are prioritized for assessing reliability; it does not mean every experiment must use the same design choice. The report should say what was done and why. If randomization or blinding was not used, state that and explain rather than inventing a procedure after the fact.
Study design and experimental unit
Identify the groups, comparator, intervention, and unit assigned to treatment. The experimental unit is the smallest independent entity allocated to an intervention. It may be an individual animal, litter, cage, tank, or another unit depending on the design.
Confusing observations with experimental units creates pseudoreplication. Forty tissue sections from four mice are not automatically 40 independent biological replicates. A model can account for measurements nested within animals, or investigators may summarize technical replicates at the animal level where justified. The manuscript should state the unit used in each analysis and explain repeated or clustered data.
Sample size
Report the exact number of experimental units in each group and total number of animals used. Explain how the number was decided. If an a priori sample-size calculation was performed, report it, including the effect size, variability, power, significance level, and method as applicable.
ARRIVE does not say that every exploratory or feasibility study must claim a formal power calculation. A transparent resource- or precision-based rationale is better than a fabricated calculation. Small size affects precision and reliability and should be acknowledged.
Inclusion, exclusion, and attrition
Define criteria for including and excluding animals or data, preferably before data collection. Report every exclusion and its reason. State whether criteria were established prospectively and whether any data were excluded after the analysis began. Reconcile planned, used, analyzed, and reported numbers.
Post hoc removal of “outliers” because they weaken an effect can bias results. If an objective rule was applied, identify it and consider a sensitivity analysis when exclusion affects conclusions.
Randomization and blinding
Report how animals were allocated to groups, including the randomization method if used and steps taken to control confounders such as cage location or order of treatment. “Animals were randomized” without a method is incomplete.
Report who knew group allocation during allocation, procedures, outcome assessment, and analysis. Some interventions make blinding impractical; describe what was possible and the likely risk of bias. Randomization and blinding improve design, but ARRIVE's role is to make the actual process visible.
Outcomes and statistical methods
Define primary and secondary outcome measures and, for hypothesis-testing studies, identify the primary outcome used for the sample-size rationale where applicable. Report statistical methods with enough detail to understand the analysis, including assumptions, transformations, multiple comparisons, repeated measures, missing data, and the unit of analysis.
A p-value alone is not a complete result. Report descriptive statistics, effect estimates, and precision where appropriate. Match the method to the design; a simple two-sample t test is usually unsuitable when many correlated sections from a few animals are treated as independent.
Animals, procedures, and welfare context
Describe species, strain, sex, age or developmental stage, weight where relevant, genotype, health or immune status, source, and previous procedures. For each procedure, report what was done, how, when, where, why, and with what anesthesia, analgesia, dose, route, and monitoring as applicable.
The Recommended Set includes the ethical statement and details of housing, husbandry, animal care, monitoring, humane endpoints, and adverse events. These are scientifically important because environment and welfare can affect outcomes. Their placement in the Recommended Set does not mean ethical approval or welfare obligations are optional. Those obligations arise from applicable law, ethics bodies, institutions, funders, and other standards.
Translation and first-in-human development
Animal evidence may contribute to a first-in-human package, but ARRIVE compliance does not “validate” safety or authorize a clinical trial. Translation also depends on model relevance, pharmacology, toxicology, manufacturing, exposure, regulatory requirements, ethics review, and the total evidence.
A publication professional should avoid claiming that journal reporting alone protects human participants. The useful contribution of ARRIVE is transparency: regulators, reviewers, and researchers can see the design, limitations, and results and judge how much weight they deserve.
Publication workflow
Use the checklist early, not only at submission. Map each item to the protocol, laboratory records, analysis output, and draft. Ask investigators about missing methods rather than guessing. Check journal requirements and submit the checklist if requested. Preserve source links and author decisions.
Important
In exam scenarios, distinguish an unreported method from a method not performed, count experimental units correctly, and never treat checklist completion as proof that an animal study is ethical, reproducible, or sufficient for human testing.
Which pair belongs to the ARRIVE 2.0 Essential 10?
Ethical statement and housing conditions
Protocol registration and declaration of interests
Sample size and randomization
Abstract and data access
Four mice per group each produce ten tissue sections, and a t test treats all 40 sections per group as independent. What is the central problem?
Recall bias
Publication bias
Lack of a journal impact factor
Pseudoreplication, because correlated sections from the same mouse are not independent biological units.
What is the correct relationship between ARRIVE reporting and authorization of a first-in-human trial?
ARRIVE compliance automatically authorizes the trial.
ARRIVE improves transparent reporting of animal evidence, while clinical authorization depends on the total nonclinical package and applicable regulatory and ethics requirements.
A journal editor replaces the regulator when ARRIVE is complete.
ARRIVE eliminates the need for toxicology and exposure assessment.
Sections you finish are checked off in the contents.