A peptide experiment can generate precise measurements without reliably answering its biological question. Repeated readings may reduce measurement noise, but they cannot repair an allocation scheme that confounds treatment with experimental day or counts dependent observations as independent evidence.
The starting point is the relationship between question, intervention, experimental unit and intended inference—not simply the number of measurements. ARRIVE 2.0 makes these relationships explicit for living-animal research. Its reporting framework can inform planning; it is not a certification system for every laboratory experiment. ARRIVE explanation and elaboration
Identify independent intervention assignments
An experimental unit is the entity assigned independently to an intervention. It may differ from the entity measured. In animal research it might be an animal, cage or litter, depending on assignment and dependence. ARRIVE study design
Consider a hypothetical experiment assigning treatment to whole cages and measuring several animals in each cage. The animals provide information within those cages; they are not all independent treatment assignments. Conversely, housing animals in cages does not automatically make cage the relevant unit. The design determines the answer.
A useful report identifies the allocation unit and observation hierarchy. “Six replicates” is insufficient if readers cannot tell whether they are independent units, subsamples or repeated readings.
Technical repeats do not create independent experimental units
Pseudoreplication occurs when dependent observations are analyzed as independent. Lazic describes repeated measurements, nested samples and correlations across time or space as routes to this problem. Lazic, full methodological paper
For example, multiple images of the same independently treated specimen can improve its characterization. They do not automatically create that many independent treatment assignments. Collecting more images is not the problem; losing their relationship to the source is.
Nor is there a universal instruction to discard lower-level observations. Summaries at the independent-unit level or models representing the hierarchy may be appropriate. The analysis must account for dependence rather than inflate apparent sample size.
Separate exploratory and confirmatory questions
A confirmatory study evaluates a specified hypothesis with a defined analysis plan. Exploratory work investigates patterns and generates hypotheses. Both can be useful, but their interpretation differs. ARRIVE distinguishes these purposes and discusses defining outcomes and analytical decisions appropriately. ARRIVE guidance
A result chosen after inspecting many outcomes should not be presented as the sole prospectively specified test. In peptide research, identify what the primary comparison can establish. One altered cellular signal does not resolve all questions about mechanism, selectivity or wider relevance.
Make allocation and assessment visible
Randomization concerns assignment. Blinding concerns what investigators know at relevant stages. Exclusion criteria determine which observations enter analysis. These address different vulnerabilities; one cannot silently stand in for another.
ARRIVE asks researchers to report these features and potential confounding variables instead of leaving readers to assume they were addressed. Essential reporting items
An illustrative problem is assigning every control sample to one experimental day and every treated sample to another. Day and treatment become inseparable. A numerical difference cannot identify the explanation responsible; recording day afterward does not undo the allocation problem.
Explain sample size and uncertainty
Sample size should refer to the relevant independent units, with subsamples described separately. Its justification should match purpose and assumptions rather than rely on a customary number.
ARRIVE includes sample-size reasoning, statistical methods, inclusion/exclusion criteria and results among its Essential 10 items. Lazic emphasizes reporting sample size, degrees of freedom, test statistics and precise probability values so readers can assess the analysis. ARRIVE Lazic
The final report should make its evidence structure recoverable: what was independently assigned, what was repeatedly measured, what was excluded and how uncertainty was calculated. Transparent reporting cannot rescue every flawed design, but it lets readers distinguish a supported conclusion from an attractive result whose independence remains unresolved.
Connect material evidence without confusing it with design
Lot traceability is useful for reconstructing which material entered an experiment. It does not establish independent replication, unbiased assignment or adequate statistical analysis. Conversely, sound allocation cannot authenticate an unknown sample.
The COA guide addresses material documentation. The biological-activity explainer addresses measured responses. A defensible study needs the relevant evidence in both areas, without treating either as a substitute for experimental design.
References
- Percie du Sert N et al. Reporting animal research: Explanation and elaboration for ARRIVE guidelines 2.0. PLOS Biology. 2020;18:e3000411.
- Lazic SE. The problem of pseudoreplication in neuroscientific studies. BMC Neuroscience. 2010;11:5.
Laboratory research use only. Not for human or veterinary use.
