# How to Justify Sample Size in Grant Proposals and Manuscripts ## Quick Answer - Justify sample size through formal power analysis for confirmatory studies, reporting the effect size, alpha, power, and calculated n per group. - For pilot or feasibility studies, justify sample size based on the study goal, such as estimating feasibility parameters or reliability coefficients, not statistical power. - Reviewers reject weak justifications that omit effect-size rationale, use arbitrary rules of thumb, or fail to account for attrition and design complexity. ## Why Sample Size Justification Matters in Grant Proposals and Manuscripts Peer reviewers and grant panelists treat sample size as a direct indicator of scientific rigor. A study with too few biological or technical replicates cannot detect meaningful effects, while an excessively large sample wastes resources and may expose more animals or human subjects to research procedures than necessary. The justification you write must convince reviewers that the chosen n follows from the study objectives, the expected variability in your system, and the statistical tests you plan to run. Funding agencies and journals increasingly expect transparent reporting of how sample sizes were determined. The National Institutes of Health provides official guidance on grant applications and review criteria, and reviewers are instructed to assess whether the proposed sample size is adequate to answer the scientific question [1]. Similarly, the EQUATOR Network catalogs reporting guidelines that many journals require or recommend, and these guidelines typically ask authors to state how the sample size was arrived at and to describe any methods used to estimate it [2]. When you submit a manuscript, the methods section must allow a reader to reproduce your sample size decision and verify that the n was sufficient. The consequences of weak justification extend beyond a single rejection. Editors and reviewers may question the validity of your entire dataset if they cannot see a logical path from the research question to the number of observations collected. In animal research, institutional animal care and use committees also review sample size justifications to ensure that the number of animals requested is the minimum necessary to achieve the scientific objective. A rigorous justification therefore serves multiple audiences: funders, journal reviewers, ethics committees, and your own laboratory team. ## Core Principles of Sample Size Estimation ### The Statistical Framework Sample size estimation rests on four interconnected quantities. The first is the effect size, which is the magnitude of the difference or association you expect to detect. The second is the significance level, conventionally set at 0.05, which is the probability of declaring a true null effect significant. The third is statistical power, conventionally set at 0.80 or higher, which is the probability of detecting a true effect when it exists. The fourth is the variability of the outcome measure, often expressed as the standard deviation. These four quantities determine the required sample size through a mathematical relationship specific to the planned statistical test. For a two-sample t-test comparing means, the required n per group increases as the effect size decreases, as the variability increases, as the significance level becomes more stringent, and as the target power increases. You cannot specify a defensible sample size without committing to values for all four inputs. ### Effect Size as the Central Decision The effect size is the most consequential and the most contested input in any sample size calculation. Reviewers will ask where your effect size estimate came from. Acceptable sources include prior published studies in the same or a closely related system, preliminary data from your own laboratory, or a clinically or biologically meaningful threshold that you define in advance. A common error is to use the effect size that produces a convenient n instead of the effect size that the science demands. When you define a meaningful effect size, you are making a scientific claim about what difference matters. For example, if you are comparing two treatment conditions in a cell culture assay, you must decide whether a 10 percent change in the outcome is biologically important or whether only a 50 percent change warrants detection. That decision belongs in the methods section with a citation or a reasoned argument. Reviewers accept a range of defensible choices, but they reject calculations that appear to reverse-engineer the n from a desired sample size. ### Variability and Its Sources The standard deviation used in a power calculation must reflect the variability you expect in your actual experiment, not the variability reported in an idealized published dataset. In life science research, variability arises from multiple sources: biological variation among individuals or cell lines, technical variation from instruments and reagents, and variation introduced by different operators or batches. Your justification should state which sources of variability are included in the estimate and how you obtained it. Pilot data are a legitimate source of variability estimates, but pilot estimates are often imprecise. A pilot trial that estimates the standard deviation of a continuous outcome can inform the main trial sample size, yet the pilot itself may be too small to pin down the standard deviation reliably [3]. When you use pilot data, acknowledge the uncertainty and consider inflating the main trial sample size to account for it. ## At a Glance: Sample Size Justification Components | Component | What Reviewers Expect | Common Deficiency | | --- | --- | --- | | Primary outcome | A single clearly defined outcome variable linked to the study hypothesis | Listing multiple outcomes without identifying the one used for the calculation | | Effect size | A numeric value with a cited source or explicit biological rationale | An unstated or arbitrary effect size that conveniently yields a small n | | Variability estimate | A standard deviation or variance with its source and measurement units | Borrowing a standard deviation from a dissimilar system without justification | | Statistical test | The exact test used in the calculation, matching the planned analysis | A generic calculation that does not match the actual analysis plan | | Attrition allowance | A stated expected dropout or technical failure rate and how it was estimated | No allowance for lost samples, failed runs, or animal losses | | Software and version | The program and version used for the calculation | No software citation, making the calculation impossible to verify | ## Practical Workflow for Writing a Sample Size Justification ### Step 1: Define the Primary Outcome and Hypothesis Write a single sentence that states the primary outcome variable and the direction of the expected effect. If your study has multiple aims, identify which aim drives the sample size. Reviewers need to see that the sample size is tied to the most important comparison, not to a secondary exploratory analysis. State whether the hypothesis is one-sided or two-sided. Most confirmatory biological experiments use two-sided tests because the direction of an effect is not guaranteed. A one-sided test requires a strong a priori rationale and reduces the required sample size, but reviewers may view it skeptically if the justification is weak. ### Step 2: Select the Effect Size and Variability Choose the effect size before you run the calculation. The most defensible approach is to use a published effect from a closely related study, your own preliminary data, or a threshold that your field regards as biologically meaningful. Write one paragraph that explains the origin of the effect size and why it is appropriate for your system. For the variability estimate, use a standard deviation from your own pilot data if available, or from published work that used the same assay and similar biological material. If you must borrow a standard deviation from a different system, explain why the variability is likely to transfer and consider using a larger value to be conservative. ### Step 3: Run the Calculation and Report All Inputs Use a statistical software package or an online calculator that allows you to specify the test, the effect size, the significance level, and the power. Record the software name and version. Report the calculated n per group and the total n after accounting for the number of groups. The calculation must match the planned analysis. If you plan to use analysis of variance with three groups, do not present a two-sample t-test calculation. If you plan to use a mixed model with repeated measures, the calculation must account for the correlation among repeated observations. A mismatch between the power calculation and the analysis plan is a common reason for reviewer criticism. ### Step 4: Add an Attrition Allowance Decide how many samples or animals you expect to lose to technical failure, contamination, dropout, or death. Multiply the calculated n by a factor that accounts for this loss. State the expected attrition rate and its basis. For example, if you expect 20 percent attrition in a cell culture experiment due to contamination and failed passages, you would increase the n by 25 percent to end with the required number of analyzable samples. Attrition allowances are study-specific. A behavioral study with human participants might expect 15 to 20 percent dropout. A long-term animal study might expect losses from illness or unexpected death. A high-throughput sequencing experiment might lose a fraction of libraries to quality control failures. Your justification should name the specific risk and the rate you assigned to it. ### Step 5: Write the Justification Paragraph Assemble the inputs into a single paragraph that a reviewer can follow without referring to supplementary files. A complete justification states the primary outcome, the expected effect size and its source, the standard deviation and its source, the significance level, the target power, the statistical test, the calculated n, the attrition allowance, and the final n per group and in total. The paragraph should read as a logical argument, not a list of numbers. Explain why the effect size is meaningful, why the variability estimate is appropriate, and why the attrition rate is realistic. A reviewer who disagrees with your assumptions can then challenge specific inputs instead of dismissing the entire calculation. ## Options and Tradeoffs in Sample Size Approaches ### Formal Power Analysis for Confirmatory Studies Formal power analysis is the standard expectation for studies that test a specific hypothesis with a planned statistical comparison. The calculation produces a sample size that gives a stated probability of detecting the assumed effect at the chosen significance level. This approach is appropriate for most grant-funded research and for manuscripts reporting confirmatory results. The main tradeoff is that the calculation is only as good as its inputs. If the effect size is wrong, the power calculation may produce an n that is too small to detect the true effect or larger than necessary. Reviewers understand this limitation and will accept a calculation with clearly stated assumptions, but they will reject a calculation that hides its assumptions. ### Feasibility-Based Justification for Pilot Studies Pilot studies serve a different purpose from confirmatory studies. A pilot study assesses whether a larger trial can be conducted, providing information about recruitment, adherence, measurement feasibility, and variability. Formal statistical power is generally not required for a pilot study, but the sample size must be justified on the basis of the pilot study's goal [4]. If the pilot goal is to assess the reliability of a questionnaire, sample size tables based on kappa agreement, intra-class correlation, or Cronbach's alpha provide minimum requirements. For these reliability tests, with alpha set at 0.05 and power at 80 percent, the recommended minimum sample sizes are at least 15 subjects for kappa agreement, 22 for intra-class correlation, and 24 for Cronbach's alpha. Allowing for a 20 percent non-response rate, a minimum of 30 respondents is sufficient to assess questionnaire reliability [5]. For a pilot trial that will estimate the standard deviation of a continuous outcome for a future main trial, the pilot sample size can be chosen to minimize the total sample size of the pilot and main trial combined. For a main trial designed with 90 percent power and two-sided 5 percent significance, recommended pilot sample sizes per treatment arm are 75 for an extra small standardized effect size of 0.1 or less, 25 for a small effect of 0.2, 15 for a medium effect of 0.5, and 10 for a large effect of 0.8 [3]. These recommendations recognize that pilot trials estimate standard deviations imprecisely and that the pilot size should be chosen with the overall trial program in mind. ### Resource-Limited Justification When funding or animal availability constrains the sample size, you must still provide a justification, but the logic changes. State the maximum feasible sample size given your resources, then report the power that this sample size provides for the effect size you consider meaningful. If the power is below 80 percent, acknowledge this limitation explicitly and discuss the implications for interpreting null results. A resource-limited justification is weaker than a full power analysis, but it is more honest than a calculation that pretends the constraints do not exist. Reviewers may accept a lower-powered study if the limitation is acknowledged and the results are interpreted accordingly. They will not accept a study that claims adequate power without showing the calculation. ## Pilot Study Sample Size Recommendations by Goal | Pilot Study Goal | Recommended Minimum Sample Size | Basis for Recommendation | | --- | --- | --- | | Questionnaire reliability with kappa agreement test | 15 subjects | Alpha at 0.05, power at 80 percent, ideal effect size [5] | | Questionnaire reliability with intra-class correlation test | 22 subjects | Alpha at 0.05, power at 80 percent, ideal effect size [5] | | Questionnaire reliability with Cronbach's alpha test | 24 subjects | Alpha at 0.05, power at 80 percent, ideal effect size [5] | | Questionnaire reliability with 20 percent non-response allowance | 30 respondents | Minimum of 30 respondents sufficient after non-response allowance [5] | | External pilot for main trial with extra small effect (0.1 or less) | 75 per arm | Main trial at 90 percent power, two-sided 5 percent significance [3] | | External pilot for main trial with small effect (0.2) | 25 per arm | Main trial at 90 percent power, two-sided 5 percent significance [3] | | External pilot for main trial with medium effect (0.5) | 15 per arm | Main trial at 90 percent power, two-sided 5 percent significance [3] | | External pilot for main trial with large effect (0.8) | 10 per arm | Main trial at 90 percent power, two-sided 5 percent significance [3] | ## Records and Measurements for Sample Size Documentation ### What to Record in Your Laboratory Notebook Document the sample size decision process in your laboratory notebook or electronic research record. Record the date of the calculation, the software and version used, the primary outcome, the effect size and its source, the standard deviation and its source, the significance level, the target power, the calculated n, and the attrition allowance. Include any correspondence with a statistician or collaborator about the assumptions. This record serves multiple purposes. It provides the material for the methods section of your manuscript. It demonstrates to reviewers and ethics committees that the sample size was determined before data collection began. It also protects you if a reviewer asks for the basis of a specific assumption months after the experiment was completed. ### Post Hoc Power Calculations Post hoc power calculations, which use the observed effect size to compute the power of a completed study, are generally discouraged. The observed effect size is an estimate with its own uncertainty, and a post hoc calculation based on it does not answer the question of whether the study was adequately powered to detect a meaningful effect. If a reviewer asks about power after seeing your results, the appropriate response is to report the confidence interval around the observed effect and discuss whether the study was large enough to detect the effect you set out to find. ### Reporting in the Manuscript The methods section should contain the full sample size justification, not a summary that omits key inputs. State the software and version used for the calculation. Report the effect size, the standard deviation, the alpha level, the power, the calculated n, and the final n after attrition. If you used a pilot study to inform the main trial sample size, describe the pilot and how its results were incorporated. The results section should report the actual number of samples or animals analyzed, which may differ from the planned number due to attrition. If the final n is lower than planned, state the reasons and discuss the impact on the study's ability to detect the planned effect. ## Common Failure Patterns in Sample Size Justifications ### The Missing Effect Size The most frequent failure is a justification that reports alpha, power, and n but never states the effect size. Without an effect size, the calculation is meaningless because any n can be justified by choosing a sufficiently small effect. Reviewers will reject this pattern immediately. ### The Arbitrary Rule of Thumb Rules of thumb such as a fixed n of three or five per group, or a minimum of 30 total samples, have no general statistical basis. These rules ignore the variability of the outcome, the size of the effect, and the planned analysis. A rule of thumb may occasionally produce an adequate sample size, but it cannot be justified as a general principle. The only exception is when the rule is tied to a specific statistical goal, such as the minimum sample sizes for reliability testing described above [5]. ### The Mismatched Calculation A calculation that does not match the planned analysis is a serious flaw. If you plan to compare four groups with analysis of variance but present a two-sample t-test calculation, the n will be wrong. If you plan a mixed model with repeated measures but present a calculation for independent groups, the n will be wrong. Reviewers who notice the mismatch will question whether any part of the analysis plan is reliable. ### The Ignored Attrition A justification that does not account for expected losses will leave the study underpowered. If you calculate that you need 10 samples per group but expect 20 percent technical failure, you must start with 13 per group to end with 10 analyzable samples. Failing to include this allowance means the final analysis will have less power than the calculation promised. ### The Post Hoc Justification Writing the sample size justification after the experiment is complete, using the observed results to select the inputs, is circular reasoning. The justification must be written before data collection begins. If you find yourself constructing a justification after the fact, the appropriate response is to report the study as exploratory and discuss the limitations of the sample size. ## Welfare and Safety Context for Sample Size Decisions ### Animal Research and the Three Rs In animal research, sample size justification is a welfare issue as well as a statistical issue. The principle of reduction, one of the Three Rs, requires that the number of animals used be the minimum necessary to achieve the scientific objective. An underpowered study wastes animals because it cannot produce a reliable answer. An overpowered study uses more animals than necessary. The sample size justification demonstrates that you have balanced these considerations. Institutional animal care and use committees review sample size justifications as part of protocol approval. A protocol that requests animals without a statistical justification may be returned for revision. The justification should state the expected effect size, the variability estimate, the power, and the number of animals per group, with the total number of animals clearly calculated. ### Biosafety and Sample Size Sample size decisions interact with biosafety in several ways. A larger sample size means more samples to process, more reagents to dispose of, and more potential exposure events for laboratory personnel. The World Health Organization Laboratory Biosafety Manual provides guidance on safe handling of biological materials, and your sample size should not exceed what your laboratory can process safely and competently [6]. If a proposed sample size would require processing more samples than the laboratory can handle within the safety protocols, the justification should be revised. ### Human Subjects Research For studies involving human participants, the sample size justification must account for the burden on participants and the risk of harm. An unnecessarily large sample exposes more people to research procedures without scientific benefit. The justification should demonstrate that the sample size is the minimum needed to answer the question, and it should account for expected dropout without inflating the sample beyond what is needed. ## Professional Escalation Criteria for Sample Size Questions ### When to Consult a Statistician You should consult a statistician before finalizing a sample size justification in several situations. If you are using a complex design such as a mixed model, a cluster randomized trial, or a longitudinal study with repeated measures, the calculation requires specialized methods. If you are uncertain about the appropriate effect size or variability estimate, a statistician can help you identify defensible sources. If a reviewer has questioned your sample size and you need to respond, a statistician can help you evaluate whether the criticism is valid. ### When to Revise a Proposal or Manuscript If a grant reviewer or journal reviewer requests additional justification, treat the request as an opportunity to strengthen the work. Provide the full calculation with all inputs stated. If the reviewer questions the effect size, provide additional citations or preliminary data. If the reviewer questions the attrition allowance, explain the basis for the rate you used. If the review reveals that your study is underpowered, you have three options. You can increase the sample size if resources permit. You can narrow the research question to focus on a larger effect that the current sample size can detect. Or you can acknowledge the limitation and frame the study as exploratory. The choice depends on the stage of the research and the resources available. ### When to Seek Institutional Guidance Funding agencies and journals have specific requirements for sample size reporting. The National Institutes of Health provides official guidance on grant applications and review criteria, and your institutional grants office can help you interpret these requirements [1]. The NIH Data Management and Sharing Policy also affects how you plan to store and share the data your sample size is designed to generate, and this planning should occur before you finalize the proposal [7]. ## Interpreting Sample Size Justifications in the Literature ### What to Look for When Reading Published Studies When you read a published study, examine the sample size justification critically. A complete justification states the primary outcome, the effect size and its source, the variability estimate and its source, the significance level, the power, the statistical test, and the attrition allowance. If any of these components is missing, the justification is incomplete. Pay attention to whether the reported sample size matches the calculation. If the authors calculated 10 per group but analyzed 8, the study may be underpowered. If they calculated 10 per group but analyzed 15, the calculation may not have accounted for the actual design. The results section should report the number of samples or animals analyzed and explain any differences from the planned number. ### Using Published Justifications as Models Published sample size justifications in your field provide useful models for your own writing. Identify papers that used similar designs and outcome measures, and examine how the authors structured their justification. Note the level of detail they provided and the sources they cited for effect size and variability estimates. Use these as templates, but do not copy the numbers without checking whether they apply to your system. ## The Role of Reporting Guidelines Reporting guidelines provide checklists for the information that should appear in a manuscript. The EQUATOR Network maintains a comprehensive database of reporting guidelines for different study types, and many journals require authors to follow the relevant guideline [2]. These guidelines typically include items about sample size, and following them ensures that your manuscript contains the information reviewers expect. For a randomized trial, the CONSORT statement requires a description of how the sample size was determined. For observational studies, the STROBE statement asks for an explanation of how the study size was arrived at. For animal research, the ARRIVE guidelines require a description of the sample size and how it was determined. Consulting the relevant guideline before you write the methods section will help you include all required elements. ## Data Management and Sample Size Planning ### The Connection Between Sample Size and Data Sharing The NIH Data Management and Sharing Policy requires researchers to plan for the management and sharing of scientific data generated from NIH-funded research [7]. This planning includes describing the types of data that will be produced, which is directly related to the sample size. A larger sample size produces more data, which may require more storage, more documentation, and more effort to prepare for sharing. Your data management plan should describe the data that the sample size will generate, including the number of samples, the types of measurements, and the file formats. The plan should also describe how the data will be preserved and shared, and it should address any restrictions on sharing that arise from the nature of the data. The sample size justification and the data management plan should be consistent with each other. ### Documentation for Reproducibility Reproducible research requires that another investigator can understand how the sample size was determined. The documentation should include the software and version used for the calculation, the inputs, and the output. If you used a custom script or a simulation, the script should be archived and referenced in the manuscript. The National Center for Biotechnology Information provides literature and data resources that can support the documentation and sharing of research outputs [8]. ## Pilot Studies and Their Sample Size Justifications ### The Purpose of a Pilot Study A pilot study is a small-scale investigation conducted to assess whether a larger trial should be initiated. Pilot studies provide critical information about feasibility, including whether participants can be recruited, whether the intervention can be delivered, whether measurements can be collected, and whether the outcome varies as expected [4]. The sample size for a pilot study is justified by the goal of the pilot, not by a formal power calculation for the main trial. ### Justifying Pilot Sample Size by Goal If the pilot goal is to assess the reliability of a questionnaire, the sample size should be based on the reliability test you plan to use. For kappa agreement, intra-class correlation, and Cronbach's alpha, with alpha at 0.05 and power at 80 percent, the recommended minimum sample sizes are 15, 22, and 24 subjects respectively. With a 20 percent non-response allowance, a minimum of 30 respondents is sufficient [5]. If the pilot goal is to estimate the standard deviation of a continuous outcome for a future main trial, the pilot sample size should be chosen to minimize the total sample size of the pilot and main trial combined. For a main trial with 90 percent power and two-sided 5 percent significance, recommended pilot sample sizes per arm are 75 for an extra small effect, 25 for a small effect, 15 for a medium effect, and 10 for a large effect [3]. ### The Limitation of Pilot Estimates Pilot trials estimate standard deviations imprecisely because the pilot sample is small. This imprecision means that the main trial sample size calculated from pilot data may be too small or too large. When you use pilot data to plan a main trial, acknowledge this uncertainty and consider using a conservative estimate of the standard deviation or inflating the main trial sample size. ## The Role of Researcher Identity and Collaboration ### Documenting Contributions to Sample Size Decisions The sample size justification is a scientific contribution that should be documented in the research record. If a statistician or collaborator contributed to the sample size calculation, their contribution should be acknowledged according to the authorship standards of the journal or funding agency. The Committee on Publication Ethics Core Practices address authorship and contributor standards, and these practices apply to all aspects of the research, including the statistical design [9]. ### Maintaining Your Research Record Your ORCID record provides a persistent identifier that links you to your research outputs, including publications and datasets [10]. Keeping this record current ensures that your contributions to sample size methodology and statistical design are discoverable. When you publish a paper with a rigorous sample size justification, the record of that work is part of your scholarly identity. ## Frequently Asked Questions ### What is the minimum acceptable power for a sample size calculation? The conventional minimum is 80 percent, meaning the study has an 80 percent chance of detecting the assumed effect at the chosen significance level. Some funders and reviewers expect higher power, particularly for studies with important clinical or policy implications. The power should be stated explicitly in the justification, and the choice should be defended if it deviates from the conventional 80 percent. ### Can I justify sample size without a formal power calculation? Yes, if the study is a pilot or feasibility study. For pilot studies, the sample size must be justified on the basis of the study goal, such as estimating feasibility parameters or reliability coefficients [4]. For confirmatory studies, a formal power calculation is expected. A resource-limited study can state the maximum feasible sample size and report the power it provides, but this is a weaker justification. ### How do I choose the effect size for my power calculation? Use a published effect from a closely related study, your own preliminary data, or a threshold that your field regards as biologically or clinically meaningful. State the source of the effect size in the justification. Avoid choosing an effect size that conveniently produces a small n without a scientific rationale. ### What should I do if my pilot data give an imprecise standard deviation estimate? Acknowledge the imprecision and consider using a conservative estimate of the standard deviation or inflating the main trial sample size. Pilot trials estimate standard deviations imprecisely, and the pilot sample size should be chosen with the overall trial program in mind [3]. Consulting a statistician can help you decide how to handle the uncertainty. ### How do I account for attrition in my sample size calculation? Estimate the expected rate of loss from technical failure, dropout, or death, and increase the calculated n accordingly. State the expected attrition rate and its basis in the justification. For example, if you expect 20 percent attrition, multiply the calculated n by 1.25 to end with the required number of analyzable samples. ### Is a post hoc power calculation acceptable? Post hoc power calculations based on the observed effect size are generally discouraged because the observed effect is an estimate with uncertainty. If a reviewer asks about power after seeing your results, report the confidence interval around the observed effect and discuss whether the study was large enough to detect the effect you set out to find. ### What reporting guidelines apply to sample size? The relevant reporting guideline depends on the study type. The EQUATOR Network maintains a database of reporting guidelines, and many journals require authors to follow the applicable guideline [2]. For randomized trials, CONSORT applies. For observational studies, STROBE applies. For animal research, ARRIVE applies. These guidelines specify the sample size information that should appear in the manuscript. ### When should I consult a statistician about sample size? Consult a statistician for complex designs such as mixed models, cluster randomized trials, or longitudinal studies. Also consult if you are uncertain about the appropriate effect size or variability estimate, or if a reviewer has questioned your sample size and you need to respond. A statistician can help you evaluate whether the criticism is valid and how to address it. ## Related Diagnostic Guides - [Immunoassay Test Types and Clinical Applications](/knowledge/diagnostics/serology-immunology/immunoassay-test-types-and-clinical-applications) - [RNA Sequencing Protocol: From Sample to Data](/knowledge/diagnostics/molecular/rna-sequencing-protocol-from-sample-to-data) - [Animal Biology](/blog/guides/animal-biology) ## References and Further Reading [1] [NIH Grants and Funding](https://grants.nih.gov/). National Institutes of Health. [2] [EQUATOR Network](https://www.equator-network.org/). EQUATOR Network. [3] [Estimating the sample size for a pilot randomised trial to minimise the overall trial sample size for the external pilot and main trial for a continuous outcome variable.](https://pubmed.ncbi.nlm.nih.gov/26092476). Statistical methods in medical research, 2016. [4] [A brief overview of pilot studies and their sample size justification.](https://pubmed.ncbi.nlm.nih.gov/38331310). Fertility and sterility, 2024. [5] [Sample size determination for conducting a pilot study to assess reliability of a questionnaire.](https://pubmed.ncbi.nlm.nih.gov/38449496). Restorative dentistry & endodontics, 2024. [6] [Laboratory Biosafety Manual](https://www.who.int/publications/i/item/9789240011311). World Health Organization. [7] [Data Management and Sharing Policy](https://sharing.nih.gov/data-management-and-sharing-policy). National Institutes of Health. [8] [NCBI Literature Resources](https://www.ncbi.nlm.nih.gov/). National Center for Biotechnology Information. [9] [Core Practices](https://publicationethics.org/core-practices). Committee on Publication Ethics. [10] [ORCID for Researchers](https://info.orcid.org/researchers). ORCID. > This article is educational and does not replace validated laboratory procedures, institutional biosafety review, manufacturer instructions, or professional interpretation.