Hostname: page-component-8448b6f56d-sxzjt Total loading time: 0 Render date: 2024-04-18T05:52:50.666Z Has data issue: false hasContentIssue false

Peer review of clinical and translational research manuscripts: Perspectives from statistical collaborators

Published online by Cambridge University Press:  04 January 2024

Phillip J. Schulte
Division of Clinical Trials and Biostatistics, Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN, USA
Judith D. Goldberg
Division of Biostatistics, Department of Population Health, New York University Grossman School of Medicine, New York, NY, USA
Robert A. Oster
Department of Medicine, Division of Preventive Medicine, University of Alabama at Birmingham, Birmingham, AL, USA
Walter T. Ambrosius
Department of Biostatistics and Data Science, Wake Forest University School of Medicine, Winston-Salem, NC, USA
Lauren Balmert Bonner
Department of Preventive Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL, USA
Howard Cabral
Department of Biostatistics, Boston University School of Public Health, Boston, MA, USA
Rickey E. Carter
Department of Quantitative Health Sciences, Mayo Clinic, Jacksonville, FL, USA
Ye Chen
Biostatistics, Epidemiology and Research Design (BERD), Tufts Clinical and Translational Science Institute (CTSI), Boston, MA, USA
Manisha Desai
Quantitative Sciences Unit, Departments of Medicine, Biomedical Data Science, and Epidemiology and Population Health, Stanford University, Stanford, CA, USA
Dongmei Li
Department of Clinical and Translational Research, Obstetrics and Gynecology and Public Health Sciences, University of Rochester Medical Center, Rochester, NY, USA
Christopher J. Lindsell
Department of Biostatistics and Bioinformatics, Duke University, Durham, NC, USA
Gina-Maria Pomann
Department of Biostatistics and Bioinformatics, Duke University, Durham, NC, USA
Emily Slade
Department of Biostatistics, University of Kentucky, Lexington, KY, USA
Tor D. Tosteson
Department of Biomedical Data Science, Geisel School of Medicine, Dartmouth College, Hanover, NH, USA
Fang Yu
Department of Biostatistics, University of Nebraska Medical Center, Omaha, NE, USA
Heidi Spratt*
Department of Biostatistics and Data Science, School of Public and Population Health, University of Texas Medical Branch, Galveston, TX, USA
Corresponding author: H. Spratt, PhD; Email:
Rights & Permissions [Opens in a new window]


Research articles in the clinical and translational science literature commonly use quantitative data to inform evaluation of interventions, learn about the etiology of disease, or develop methods for diagnostic testing or risk prediction of future events. The peer review process must evaluate the methodology used therein, including use of quantitative statistical methods. In this manuscript, we provide guidance for peer reviewers tasked with assessing quantitative methodology, intended to complement guidelines and recommendations that exist for manuscript authors. We describe components of clinical and translational science research manuscripts that require assessment including study design and hypothesis evaluation, sampling and data acquisition, interventions (for studies that include an intervention), measurement of data, statistical analysis methods, presentation of the study results, and interpretation of the study results. For each component, we describe what reviewers should look for and assess; how reviewers should provide helpful comments for fixable errors or omissions; and how reviewers should communicate uncorrectable and irreparable errors. We then discuss the critical concepts of transparency and acceptance/revision guidelines when communicating with responsible journal editors.

Review Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (, which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
© The Author(s), 2024. Published by Cambridge University Press on behalf of Association for Clinical and Translational Science


The types of articles and clinical research studies evaluated and considered for publication by medical journals are broad and diverse, as are the associated statistical methods appropriate for analysis of those studies. Review of statistical methodology is a core component of the peer review process, but the statistical review process for submissions varies greatly among journals, as does the quality and depth of statistical review among individual reviewers, even within the same journal. Some journals retain biostatistics experts for peer review, although this practice is not widespread. Identification of experts in quantitative methodology and experience or expertise in the clinical or translational research domain with interest and availability to serve as a reviewer or editorial board member is a common challenge for editors [Reference Dance1Reference Fox4]. Even when appropriate reviewers are identified, there is often little guidance provided to help the reviewer and to enable appropriate, equitable review practices. Given the diversity of reviewers and practices, the scientific community would benefit from guidance that could streamline the review process and provide a unified, structured, and systematic approach. Such guidance could provide more transparent, constructive, and consistent reviews to authors and promote robust scientific practice [2].

Many publications provide statistical guidelines for authors and research collaborators, including those published in general medical journals [Reference Altman, Gore, Gardner and Pocock5,Reference Harrington, D’Agostino and Gatsonis6] or as books [Reference Lang and Secic7,Reference Peat and Barton8] and those in subject-specific medical journals [Reference Armstrong, Davies, Dunne and Gilmartin9Reference Johnson, Stone, Bunn, Lyons and Navalta19]. Resources targeting statistics and quantitative methodology professionals also provide relevant guidelines[Reference Wasserstein, Schirm and Lazar20Reference Smith22]. Alternatively, many clinical and translational science journals provide guidelines on websites in the instructions for authors [2326]. While these publications provide guidance to the authors when writing a paper and performing analyses that are reported, these guidelines are generally tailored to basic or commonly used statistical methods. Moreover, ongoing developments in quantitative methods mean that best practices are ever-changing, and analysis guidance documents may struggle to stay current. Reporting guidelines were developed to focus on transparent reporting of results, but not necessarily to provide guidance for research conduct, for randomized trials [Reference Schulz, Altman and Moher27], observational studies [Reference von Elm, Altman, Egger, Pocock, Gøtzsche and Vandenbroucke28], meta-analyses and systematic reviews [Reference Page, McKenzie and Bossuyt29], diagnostic/prognostic prediction studies [Reference Collins, Reitsma, Altman and Moons30Reference Stevens, Mortazavi, Deo, Curtis and Kao32], and preclinical animal studies [Reference Percie du Sert, Hurst and Ahluwalia33], among many[34]. Ethical guidelines also exist for peer review but without focus on the content of a conducted review [35]. Recommendations for statistical analysis and methods reporting for authors also exist [Reference Lang and Altman36,Reference Bailar and Mosteller37]. Taken together, these publications and reports provide comprehensive recommendations for authors but lack guidance for the peer reviewer tasked with reviewing the scientific content of a submitted manuscript and the application of quantitative methodology and biostatistics in that manuscript. While prior papers have provided recommendations for the peer reviewer of submissions to medical journals, recommendations are often specific and lack generalizability to the broader medical research community or are tailored to a narrow audience of biostatisticians with formal training in statistical methods [Reference Altman38]. Recommendations for the peer reviewer should consider the diversity of scientific work while maintaining consistency with ongoing and future developments in the statistical sciences. These recommendations should also be accessible to a broader community of reviewers who evaluate quantitative methods in a manuscript [Reference Greenwood and Freeman39]. Further, the need exists for guidance and recommendations in communication to editors and authors to articulate statistical concerns and necessary revisions clearly.

The purpose of this manuscript is to provide guidance for reviewers who are responsible for assessing the application of quantitative methodology and is intended to complement the guidelines and recommendations for authors previously described. We describe key components of clinical and translational science research manuscripts that require attention to the statistical methodology or that have considerable implications for statistical methods. These include study design and hypothesis, sampling (selection of the study subjects), interventions (for studies that include an intervention), measurement (assessment of study variables including outcome measures), statistical analysis methods, presentation of the study results, and interpretation of the study results. For each component, we describe what reviewers should look for and assess; we offer recommendations for how reviewers should provide helpful comments for fixable errors or omissions; and we describe how reviewers should communicate uncorrectable and irreparable errors. We next provide a checklist summarizing what to look for and assess when reviewing a manuscript (Table 1). We then discuss the critical concepts of transparency and acceptance/revision guidelines when communicating with responsible journal editors.

Table 1. Checklist guide of items to consider in biostatistical review of clinical and translational manuscripts

Objective, hypothesis, and design


The study objective or aim, and study design set the framework for a scientific manuscript. Reviewers must first assess whether the research objective or aim is coherent, sensible, and statistically evaluable; subsequently, the reviewer evaluates whether a proposed study design supports the research objective. The research objectives are defined inclusively to encompass comparisons of interventions or therapies (randomized or observational), epidemiological and etiological studies, prediction and diagnosis objectives, mechanistic studies, and more.

Suggestions for Revision

Fixable flaws from a study design perspective generally relate to a lack of clarity in the description of the design or a poor description of the objectives. Comments to the authors’ revision may focus on requests for clarification or additional detail, or revisions to wording to clarify aims. For example, studies that specify a testable hypothesis are typically described as the statistical alternative hypothesis. That is, authors might hypothesize that a treatment causes a difference in outcome, on average, between those treated and untreated. This corresponds to the statistical alternative hypothesis whereas the null hypothesis is “no difference” between groups. On occasion, authors will incorrectly describe the hypothesis using language corresponding to the null hypothesis of the statistical methods used to analyze the data, providing a hypothesis that “groups will be similar.” This may necessitate a comment suggesting a change in analysis methods or a change in how hypotheses are described. This is because statistical tests in the frequentist framework either reject or do not reject the null hypothesis and the colloquial phrase: absence of evidence to reject the null hypothesis is not evidence of the null hypothesis.

Irreparable Issues

A study design that is not aligned with the research objective is generally not fixable since this would require a new study to be conducted using a more appropriate design. For example, the aim of an epidemiological study might be to evaluate the association between an exposure and outcome. If the exposure is measured and ascertained after the outcome, conclusions may be muddied by reverse causation and other logical flaws. When the study design does not provide a statistically valid approach to address the objectives or aims, a recommendation of rejection can be provided to the editor with a well-reasoned description of the concern to the authors.

Unhelpful Comments

Reviewers should be cautious in distinguishing between inappropriate study designs and designs that are adequate but perhaps not the reviewer’s preferred approach. For example, a well-reasoned, -analyzed, and -reported observational comparative effectiveness study may be a sound design advancing the science of the research objective despite not being a gold-standard randomized trial. In general, there may be multiple statistically principled study designs that can address a given research question; thus, comments to authors should not suggest a reviewer-preferred approach or recommend rejection of the manuscript if the existing approach is still valid and interpreted correctly.



Researchers refer to the set of study subjects used in a research objective as a “sample” whereas “population” reflects a broader group to which the conclusions or results of the study are applicable. Reviewers should typically expect a description of how researchers planned to collect the sample of data, including (i) description of the population of interest, (ii) framework for acquiring a sample, and (iii) justification of sample size. The population of interest is typically described by inclusion and exclusion criteria, and the reviewer should ensure that these criteria align with the study aims. In clinical and translational science, samples are most often driven by convenience such as electronic medical records from a particular clinic, subjects enrolled and providing informed consent at a participating site or sites, mice or animals bred by a single supplier, or tissue and serum from patients seen at a hospital or clinic [Reference Sedgwick40]. Stratified sampling or clustered sampling may be used in some studies. A reviewer must first assess whether this sampling framework is described well but also evaluate whether subsequent methods align with data collection and whether limitations of a sampling framework are appropriately acknowledged [Reference Cochran41]. In observational data analyses, the framework for sampling may require a description of whether data are cross-sectional, cohort, or case-control, for example, with appropriate justification for whether that data can address the aims of the study. Missing data are not a sample characteristic but rather a Measurement concern, while complete case analysis is Analysis Methods decision or choice rather than a characteristic of the sample; critique of missing data is subsequently addressed in those sections. Finally, sample size should often be justified before data are collected (a priori), such as evaluating statistical power [Reference Adcock42Reference Lenth45], anticipated Bayesian posterior distributions [Reference Wang, Chow and Chen46,Reference Kunzmann, Grayling, Lee, Robertson, Rufibach and Wason47], or relevant metrics for prediction model development [Reference Riley, Ensor and Snell48Reference Riley, Debray and Collins51]. Whether a numeric calculation such as this is described or a qualitative justification is given, reviewers should assess the robustness of this justification.

After the study is conducted, the characteristics of the selected sample should be summarized to facilitate an assessment on how the sample might differ from the target population, thus allowing for an evaluation of the generalizability and transportability of the results. This evaluation may be assisted by a table showing summary statistics for subject characteristics. Generalizability and transportability reflect whether results obtained in the sample are reasonably applied to other subjects in the population [Reference Degtiar and Rose52]. Biased representation of the sample to the target population or poor generalizability – such as a convenience sample lacking racial diversity present in the population, a tertiary referral hospital with more complex cases than a typical hospital, or a preclinical study using only male mice to study a condition that affects both sexes – may unknowingly affect the interpretation of the results. The lack of generalizability or transportability may exacerbate inequalities and lead to poor medical decision-making – even harm [Reference Degtiar and Rose52,Reference Chen, Pierson, Rose, Joshi, Ferryman and Ghassemi53]. Thus, the reviewer should critically evaluate whether the sample of data reasonably reflects the population in which the intervention, statistical model, or characteristic will be applied.

Suggestions for Revision

Reviewers should first ensure that sufficient information is provided to assess potential biases. The impact of the sampling process, even with major limitations that make the sample a “convenience” sample, can be mitigated in some contexts with appropriate experimental designs or statistical analysis methods. Otherwise, reviewers should look for a thorough discussion of the implications of the sampling process. If any of the items mentioned above (target population, sample size determination, sample selection, and representativeness of the sample discussion) are missing, the reviewer can suggest the authors provide such information to ensure the study was rigorously conducted.

If a sample size justification was not prepared prior to the analysis, the reviewer should request a revised method section with detail how the existing sample size was determined and a quantitative calculation provided but clearly denoted as being done after the analysis was conducted. Post hoc or observed power is never an appropriate request from the reviewer, but the reviewer may request power to detect a “small effect” or calculate the difference between groups that the sample size provides 80% or 90% power to detect [Reference Lakens54]. When sample size is fixed due to external constraints such as a retrospective data source or limited population such as a rare disease setting, a similar quantitative sample size justification may be provided along with the qualitative description of the constraints.

Irreparable Issues

Reviewers should first give authors an opportunity to explain why the particular sampling approach was used and address limitations. We note, however, that studies that employ a sample with poor generalization to the population or exclude a subset of the population are not generally fixable through peer review. Such instances may include studies that were performed on a single sex, gender, race/ethnic group, or age group when the disease/virus/predicament under study is applicable to many.

Unhelpful Comments

It is generally not helpful to suggest that the authors obtain new data under an alternative sampling approach. Unless the manuscript is describing future planned research, at the time a manuscript is under review, data have already been collected and the reviewer should respect contributions of animals or human subjects or samples to the research when possible. Instead, reviewers should require a robust discussion of limitations or future research needs.



Not all clinical and translational research involves the study of interventions, therapies, or treatments applied to samples. Rather, some studies assess etiology of disease, ability to diagnose conditions, or predict future outcomes, for example. However, those who evaluate interventions, treatments, or therapies will be expected to explain when and how the intervention was applied, with a level of fidelity that allows readers or future researchers to replicate the results [Reference Hoffmann, Glasziou and Boutron55]. In trials that prospectively prescribe intervention (broadly encompassing single arm, crossover, cluster-randomized, parallel arm randomized, and others), this intervention should be prescribed a priori in the study protocol and subsequently described in the submitted manuscript. In analysis of observational studies with retrospective or prospective data collection, authors should include a comprehensive description of how the intervention is defined and data to assess the effects of the intervention were obtained.

The timing of intervention relative to collection of outcomes should reflect the potential biological mechanisms of action so that outcomes are not ascertained too early nor too late relative to the application of the intervention. In prospective studies, outcome assessments can be controlled at observation timepoints with biological relevance. In retrospective studies, the reviewer must carefully assess whether authors’ collection of these data could bias the study, such as a survivorship bias or immortal time bias [Reference Klein, Van Houwelingen, Ibrahim and Scheike56].

Suggestions for Revision

Even if the intervention is well-described intervention, there may be a need to recommend a robust discussion of study limitations. Additional suggestions may depend on the study design. A reviewer, for example, might suggest authors develop a Directed Acyclic Graph [Reference Shrier and Platt57] with subject expertise to assess potential confounding in an observational comparative effectiveness study. While proper randomization and intention-to-treat principles prevent bias due to confounding, adjustment for prognostic variables in randomized interventional studies should be similarly pre-specified [Reference Moher, Hopewell and Schulz58]. Crossover trials [Reference Mills, Chan, Wu, Vail, Guyatt and Altman59Reference Senn61], before-after implementation studies [Reference Bernal, Cummins and Gasparrini62,Reference Wagner, Soumerai, Zhang and Ross-Degnan63], cluster-randomized trials, and others [Reference Kim, Bretz, Cheung and Hampson64] each have implications that reviewers should consider to ensure authors have provided an adequate description of the intervention and limitations of the intervention applicable to the specified design.

Irreparable Issues

In retrospective studies, authors may be limited by available data and granularity of that data. For example, an administrative database may document interventions applied during an intensive care unit admission without recording specific timing of the intervention. A study to evaluate outcomes also measured during the intensive care unit admission could be unable to determine whether the timing of intervention relative to outcome reflects biological plausibility. Reviewers would need to rely on subject-matter knowledge to understand whether the data can plausibly evaluate the objective or hypothesis, and whether concerns can be mitigated through identified limitations or whether the data are not adequate to evaluate for the study objective.

Unhelpful Comments

Suggested changes regarding how the interventions were implemented are not helpful in studies that are underway or complete. Rather, the reviewer should give constructive comments or suggestions to improve the description of the intervention and suggest corresponding limitations.



Measurement refers to both the collection or recording of data, data quality, and data fidelity. Measurement may reflect the timing of data collection such as a visit schedule in a prospective trial. This domain also includes whether data are missing, and the extent and reasons for missing data. Measurement also reflects how data are documented in source or raw data for the study, recorded for later analysis as analysis data, and whether such data are either continuous, interval, count, binary, nominal categorical, or ordinal categorical. Time-to-event data include a binary or categorical event or state status and continuous or interval time. Sufficient descriptions of measurements are necessary for the reader to evaluate data presented and allow for reproducibility in future research.

The reviewer should assess how the missingness arose in the study and whether treatment of missing data is appropriately handled and discussed in the manuscript. Longitudinal studies should address missingness along the entire measurement time course. This should be examined at both the individual variable level (how many subjects are missing the variable information) as well as the overall study level (how many key variables have missing data). When missing data are present, statistical analysis methods should be described and limitations and assumptions of that method should be defined. Notably, missing data does not appear under Sampling. While complete case analyses are common in clinical and translational research, this is a measurement and analysis choice by the analyst rather than a sample or population characteristic; therefore, missing data should generally not be an inclusion or exclusion criteria.

Statistical analyses should reflect the nature of the measurement of the variables; for example, continuous variables should be analyzed with methods appropriate for continuous data. Continuous data should not be discretized or dichotomized unless there is a clear mechanistically based rationale, or the categories are well accepted and validated externally. When considering statistical inference, in general, discretization of continuous data does not represent a biologically plausible relationship. On the other hand, ordinal categorical data (examples include tumor stage, Likert scale survey response) should generally not be analyzed as continuous data. An additional consideration includes assessment of measurement error or misclassification error. The reviewer should evaluate whether the methods sufficiently allow for assessment of measurement error and, if applicable, whether statistical analyses are appropriate and whether limitations are discussed.

Suggestions for Revision

A thorough description of missing data and measurement is often lacking in research manuscripts. Reviewers should request justifications for decisions based on measurement of the data. Two common examples include use of complete case analysis without consideration of missing data mechanisms and arbitrary categorization of continuous data. Both may lead to biased interpretation or loss of precision [Reference Molenberghs, Fitzmaurice, Kenward, Tsiatis and Verbeke65Reference Thoresen70] and, if not fixable through revision, authors should be expected to provide discussion of the limitations applicable to these approaches.

Irreparable Issues

Incorrect assessments or measurements of the source data are rarely uncorrectable. Indeed, many statistical methods have been developed to explain signals in the presence of error and uncertainty in measurement.

Unhelpful Comments

When data are collected poorly or poorly measured at the source data, for example, a researcher using categories of age in decades (40-49, 50-59, etc.) rather than integer or more specific continuous age, it may not be feasible for the researcher to revisit data collection. As noted in Suggestions for Revision, the reviewer may request revision if applicable to the available source data. Beyond that, measurement issues might be addressed through statistical analysis and also with robust discussion of limitations and future research needs.

Analysis Methods


Statistical analysis methods should be described in detail so that another researcher with adequate training could replicate the analysis and obtain the same results if given the same dataset [Reference Lang and Altman36,Reference Weissgerber, Garcia-Valencia, Garovic, Milic and Winham71]. The analysis methods description should tie each aim or hypothesis described in the introduction to specific analysis methods. When applicable, reviewers need to consider implications of multiple comparisons corresponding to more than one hypothesis [Reference Rubin72]. Reviewers should expect analysis methods in the main manuscript methods section, but in those circumstances when the complexity of analysis is far greater than the typical study, a thorough and detailed methods section may be included as supplementary material if allowed by the journal.

Suggestions for Revision

When analysis methods are vague, reviewers should request specification and detail in a revised manuscript that allows for replication of the analysis. The reviewer should ensure that analysis methods not commonly used in the subject-matter discipline are described and cited.

Analysis methods should describe statistically valid and appropriate methods based on the Sampling, Intervention (if applicable), and Measurement of data to address study Objectives. If there are discrepancies, the reviewer can suggest that certain outcomes or objectives might be better analyzed using alternative approaches. Multiple statistical methods may often be valid, but there are potential tradeoffs that the researchers must consider; for example, some methods may carry fewer assumptions about the data and analysis model, some may be more statistically powerful in limited scenarios, or there may be nuanced differences in interpretation while still aligned with the overarching objective. The reviewer should ensure assumptions of statistical methods are described and assessed where appropriate and that the description of the methods used appropriately considers tradeoffs.

Irreparable Issues

Errors in the statistical analysis methods are rarely unfixable; however, analysis flaws can uncover major concerns with Sampling or Measurement of data.

Unhelpful Comments

The reviewer should be open to alternative analytic approaches. Consider, for example, that a binary outcome can be regressed on covariates using logistic regression, log-binomial or Poisson regression, linear probability models, or other approaches – estimating an odds ratio, risk ratio, or risk difference, respectively. While each method requires that different assumptions are satisfied, there are many situations where more than one approach is valid and others in which one or more approaches may not be valid. The reviewer should evaluate the appropriateness of the analysis method for the study aim, keeping in mind that there may be more than one effective approach. Reviewers can ask authors to provide clear justification, including an evaluation of whether the assumptions required for the proposed methods are well-described before concluding that the methods used are unacceptable.

Pre-specified analysis plans are commonplace in prospective interventional trials and are also strongly advised for observational research [Reference Thomas and Peterson73]. When possible, reviewers should respect the plan and avoid potential biases in selecting methods based on observed data. If pre-specified plans are inadequate, reviewers may suggest post hoc secondary analyses to supplement pre-specified methods. In rare instances, pre-specified plans may represent invalid approaches and reviewers may need to suggest only the post hoc analysis be reported with a clear rationale for readers.



Presentation includes summaries of the sample and analysis results in the text, tables, and figures. The presentation of results should follow directly from the Analysis Methods. For each analysis method described, there should be a summary of the corresponding result. The reviewer should ensure consistent and accurate presentation in the abstract, body of the manuscript, and in tables and figures. Variables, descriptions, and numeric results should be consistent throughout the manuscript. Tables and figures should have a descriptive title and, with accompanying footnotes and captions, should be clear and self-explanatory.

Suggestions for Revision

A good table or figure can convey a large amount of high-quality data whereas a poor figure can send mixed messages. The reviewer can make specific suggestions or point toward existing resources providing robust guidance for data presentation for authors [Reference Weissgerber, Winham and Heinzen16,Reference Weissgerber, Milic, Winham and Garovic74]. The reviewer should assess whether tables and figures are effectively formatted to interpret data for objectives and hypotheses and that presentation of results retains focus on the objectives. As an example, suppose a study evaluates an effect of randomized treatment vs placebo on an outcome with assessment of the outcome variable before treatment/placebo and after. A poor presentation would draw the reader’s attention to pre-post differences within each group which may be similar in both treatment and placebo study arms. Instead, the reviewer should request organization of results focusing on how treatment (vs placebo) affects the post-treatment/placebo outcome using an appropriate design and statistical method.

A focused presentation of results also suggests the reviewer should recommend removing ancillary data that are not directly related to the objectives or hypotheses. A study that develops a multivariable prediction model following applicable reporting guidelines [Reference Collins, Reitsma, Altman and Moons30] could provide a table of coefficients that can be used to output prediction, but may not need to include relative effect measures such as odds ratios with confidence intervals that are not relevant to the objective of obtaining a prediction. Similarly, if the objective is to estimate the association between an exposure and outcome adjusted for potential confounders in an observational study [Reference Shrier and Platt57], association estimates for each confounding variable do not need to be reported. Rather, retaining focus on the exposure of interest only will help readers avoid improper inference such as the Table 2 fallacy [Reference Westreich and Greenland75]. Reviewers should be skeptical of extraneous statistical tests and results unrelated to study hypotheses. For example, a table of baseline characteristics in a randomized trial should not include hypothesis testing comparing pre-randomization variables by randomized arm [Reference Moher, Hopewell and Schulz58], whereas in observational research such comparisons are valid but often unnecessary and irrelevant to the study objectives.

Irreparable Issues

Presentation errors are rarely unfixable with a revision. However, issues identified when reviewing results of a manuscript may uncover additional concerns with Sampling, Measurement, or Analysis Methods that necessitate further suggestions.

Unhelpful Comments

Problematic tables, figures, and results text require clear and constructive comments. Vague comments that critique the presentations without suggestions about how this could be improved are insufficient.



In the review of the interpretations of statistical analyses, there are three primary features that a reviewer should assess in a manuscript. The first is that interpretations of results should appropriately and precisely reflect what can be inferred directly from study results. For example, generalizations should only be made to the population that is reflected by the sample. The second fundamental feature of interpretation is that limitations of the study should be thoroughly noted. Reviewers should evaluate whether assumptions or incomplete knowledge that impact the data or conclusions are disclosed, including potential impacts of missing data or confounding. The limitations inherent in the data (e.g., missingness, surrogate variables, potential biases, etc.) should be clearly described. Reviewers should expect limitations in the data and analyses to be related with implications for interpretation of the results, rather than simply a list of limitations. When the likely impact is differential, the direction should be noted. Finally, assessing interpretation in the manuscript evaluates whether there is necessary context for conclusions. For example, the difference between statistically significant and clinically/practically significant results or the difference between confirmatory and exploratory results may need to be addressed.

Suggestions for Revision

Since interpretation is particularly fixable at the review stage, reviewers have an opportunity to provide meaningful feedback to improve a manuscript. Critiques of interpretation are most helpful when they are specific rather than general. For example, suppose that authors interpret a null finding from a traditional hypothesis test of superiority as indicating that there is evidence of no association or difference. It would not be helpful for a reviewer to vaguely state “Interpretation of hypothesis tests are not correct.” Instead, the following critique would be more helpful: “For hypothesis tests, statistically nonsignificant results should be interpreted as providing insufficient evidence of an association, rather than as providing evidence of no association.”

Irreparable Issues

Rarely will errors in interpretation be unfixable for a manuscript under review. It is possible for errors in interpretation to be inaccurate (rather than accurate but inappropriate), but these errors can usually be fixed through revision.

Unhelpful Comments

Reviewers should recognize that communication is highly individual; just because authors interpret results differently than the reviewer does not necessarily make the authors’ interpretation incorrect or inappropriate. Reviewers may benefit from assessing their own critiques of interpretation by asking themselves whether the issue they are commenting on is truly incorrect or inappropriate, or whether it is just a difference in style.


The inclusion of statistical reviewers has been shown to improve quality of published manuscripts [Reference Schor and Karten76Reference Gardner, Altman, Jones and Machin79], but adoption of specialized review by individuals with expertise and training in statistics and quantitative methodology in medical research journals is lacking [Reference Hardwicke and Goodman80Reference Goodman, Altman and George82]. Review of quantitative methodology often falls to subject-matter experts with varied competency in statistical methods, sometimes with expertise in only a narrow range of statistical methods specific to their research program [Reference Greenwood and Freeman39]. Recommendations for statistical review herein can be applied by expert biostatisticians and epidemiologists, as well as subject-matter practitioners with additional statistical training or competencies in quantitative methodologies.

Recommendations and Communication

Reviewers must score and communicate recommendations to the responsible editor. These recommendations generally include (i) comments and suggestions for the authors, (ii) private comments to the editors, and (iii) a recommendation to the editors. Ultimately, the responsible editor will collect comments from multiple reviewers, consider private comments and recommendations, and make a decision on the manuscript to either accept the paper, request a revision of the paper to address reviewer concerns, or reject the paper. Prior sections have focused on how reviewers can assess core components of a clinical or translational research manuscript and provide comments to authors guiding revisions. In private comments to the editors, the reviewer should articulate any potential conflicts of interest as well as any limitations to their review. Statisticians and methodologists are rarely expert on all statistical methods and may be asked to review manuscripts that use unfamiliar methods. Similarly, subject-matter experts with limited statistical training may often review and provide recommendations related to statistical methods. In private comments to the editor, reviewers should acknowledge their own limitations. Invitations to review do not always include sufficient detail to make a determination about methodological expertise up front, so providing these comments to the responsible editor can allow them to seek additional expert opinions as needed.

Editors also request, in addition to written comments, that reviewers make a recommendation for the manuscript: accept, revise and resubmit, or reject (or variations of these). The statistical reviewer should consider recommending acceptance when there are no concerns about the study conduct or conclusions – either from the statistical review as described previously or with respect to subject-matter concerns. A recommendation to the editor to reject a manuscript must include significant justification that there are major errors in the paper and those errors are unlikely to be fixed through revision (irreparable issues). Finally, manuscripts with flaws identified, but not irreparable, should be recommended for revision. Some journals subclassify revisions as minor or major based on how substantial revisions will need to be and how those revisions are likely to impact the overall messaging of the manuscript. Minor revisions are unlikely to require changes to the underlying data or analysis and comments to the authors may request changes that are editorial in nature. Major revisions will include comments to the authors that request significant changes to statistical methods and re-analysis of the data.


Statistical reviewers should make recommendations based on best practices in the field of collaborative biostatistics. Reviewers often read manuscripts that use incorrect or inappropriate methods, or poor study descriptions on the basis that prior papers have done so [Reference Weissgerber, Garcia-Valencia, Garovic, Milic and Winham71]. Some concerns may have little impact on conclusions, but the reviewer has responsibility to reduce propagation of incorrect methods. Inappropriate methodology with minor impacts on overall study conclusions can be addressed via comments and requests for minor revisions.

At the same time, the statistical reviewer must be flexible and adaptive to novel approaches or changes in best practices as methodologies evolve. It is critical for the reviewer to avoid dogmatic comments, especially those that suggest specific approaches or methodologies on the basis that prior papers have used them. When reviewers are not confident that suggestions reflect best practices, they may refrain from commenting on authors, make comments expressing uncertainty to the editor, or provide comments to the author that acknowledge their own limitations and uncertainty. Finally, reviewers should be open to rebuttals that do not implement those suggestions in a revised manuscript if authors can defend that alternative approaches represent valid and correct statistical approaches.

When possible, reviewers should consider implications of pre-specified study design and statistical analysis. Pre-specification avoids biases that may arise when making design and analysis choices based on observation of intermediate results. Reviewers should consider review of those pre-specified plans and whether authors adequately adhered to those plans. This includes review of registration for clinical trials.

Limitations of Statistical Review

Manuscript reviewers are tasked with evaluating whether the described methods are appropriate, and whether results are presented and interpreted correctly, but journals rarely require submission of data and programing code. As such, statistical reviewers cannot be accountable for evaluating whether implementation such as programing or software implementation was correct. In contrast with grant review or study section review [Reference Hug and Aeschbach8386], manuscript reviewers are not positioned to critique the credentials or training of the manuscript authors. The reviewer should not suggest authors “consult with a statistician” or make comments based on degree or position/appointment of authors. Rather, it is incumbent upon the authors to revise according to the objective and best practices-based comments made by the reviewer. This does not reflect a “free pass” or reduced criteria applied to those manuscripts without a biostatistician or methodologist among coauthors, but instead reflects varied approaches available for authors to implement revisions leading to pscientifically valid approach and conclusions.

Future Directions

Guidance provided in this paper represents an opportunity for training new clinical and translational science researchers, including biostatisticians and non-biostatisticians alike. Service and national recognition achieved through exemplary peer review and editorial board participation are favorable for consideration of faculty promotion and tenure [Reference Spratt, Fox, Shara and Mazumdar87,Reference Waller88]. Thus, education on the process of statistical review may improve not only the submitted clinical and translational research papers, but benefits may extend to readers and trainees learning to improve their review skills.


Reviewers rely on adequate descriptions of the analysis to provide a critique and as such, published poor-quality research is the fault of authors rather than reviewers and editors. Nonetheless, reviewers are asked to review study design, statistical methods, and presentation and interpretation of results, and can provide meaningful critiques that can improve statistical reporting and analysis in medical research. Recommendations provided in this paper are accessible to a wide audience of trained biostatisticians and quantitative methodology experts as well as subject-matter experts with varied formal training in statistics. In assessing each of these components of a manuscript under review, the reviewer may need to consider additional specialized guidelines and weigh evidence from different sources to best understand and recommend best practices.

Funding statement

This study was supported by the following grants from the National Institutes of Health: UL1TR001439 and R25HL161715 (HMS), UL1TR001998 (ES), UL1TR003096 (RAO), UL1TR002001 (DL), UL1TR001445 (JDG), U54 GM115458 (FY), UL1TR001420 (WTA), UL1 TR002377 (REC), UL1TR001086 (TDT), UL1TR002553 (CJL and G-MP), and 1UL1-TR001430 (HJC). Contents are the authors’ sole responsibility and do not necessarily represent official NIH views.

Competing interests

CJL receives research funding to his institution from NIH, DoD, CDC, Biomeme, Entegrion, Endpoint Health, AstraZeneca, bioMerieux, and stock options in Bioscape Digital unrelated to the current work.


Dance, A. Stop the peer-review treadmill. I want to get off. Nature. 2023;614(7948):581583.CrossRefGoogle Scholar
Publons. 2018 Global State of Peer Review. London: Clarivate Analytics, 2018. Scholar
Fox, CW. Difficulty of recruiting reviewers predicts review scores and editorial decisions at six journals of ecology and evolution. Scientometrics. 2017;113(1):465477.CrossRefGoogle Scholar
Altman, DG, Gore, SM, Gardner, MJ, Pocock, SJ. Statistical guidelines for contributors to medical journals. Br Med J (Clin Res Ed). 1983;286(6376):14891493.CrossRefGoogle ScholarPubMed
Harrington, D, D’Agostino, RB Sr., Gatsonis, C, et al. New guidelines for statistical reporting in the journal. N Engl J Med. 2019;381(3):285286.CrossRefGoogle ScholarPubMed
Lang, TA, Secic, M. How to Report Statistics in Medicine: Annotated Guidelines for Authors, Editors, and Reviewers. Philadelphia, PA: American College of Physicians Press, 2006.Google Scholar
Peat, J, Barton, B. Medical Statistics: A Guide to Data Analysis and Critical Appraisal. Hoboken, NJ: John Wiley & Sons, 2008.Google Scholar
Armstrong, RA, Davies, LN, Dunne, MC, Gilmartin, B. Statistical guidelines for clinical studies of human vision. Ophthalmic Physiol Opt. 2011;31(2):123136.CrossRefGoogle ScholarPubMed
Assel, M, Sjoberg, D, Elders, A, et al. Guidelines for reporting of statistics for clinical research in urology. Eur Urol. 2019;75(3):358367.CrossRefGoogle ScholarPubMed
Dupuy, A, Simon, RM. Critical review of published microarray studies for cancer outcome and guidelines on statistical analysis and reporting. J Natl Cancer Inst. 2007;99(2):147157.CrossRefGoogle Scholar
Hickey, GL, Dunning, J, Seifert, B, et al. Statistical and data reporting guidelines for the European journal of cardio-thoracic surgery and the interactive cardioVascular and thoracic surgery. Eur J Cardiothorac Surg. 2015;48(2):180193.CrossRefGoogle ScholarPubMed
Uter, W, Schnuch, A, Gefeller, O. Guidelines for the descriptive presentation and statistical analysis of contact allergy data. Contact Dermatitis. 2004;51(2):4756.CrossRefGoogle Scholar
Vickers, AJ, Assel, MJ, Sjoberg, DD, et al. Guidelines for reporting of figures and tables for clinical research in urology. J Urol. 2020;204(1):121133.CrossRefGoogle ScholarPubMed
Vickers, AJ, Sjoberg, DD. Guidelines for reporting of statistics in European urology. Eur Urol. 2015;67(2):181187.CrossRefGoogle ScholarPubMed
Weissgerber, TL, Winham, SJ, Heinzen, EP, et al. Don’t conceal: transforming data visualization to improve transparency. Circulation. 2019;140(18):15061518.CrossRefGoogle ScholarPubMed
Althouse, AD, Below, JE, Claggett, BL, et al. Recommendations for statistical reporting in cardiovascular medicine: a special report from the American heart association. Circulation. 2021;144(4):e70e91.CrossRefGoogle ScholarPubMed
Ou, FS, Le-Rademacher, JG, Ballman, KV, Adjei, AA, Mandrekar, SJ. Guidelines for statistical reporting in medical journals. J Thorac Oncol. 2020;15(11):17221726.CrossRefGoogle ScholarPubMed
Johnson, SL, Stone, WJ, Bunn, JA, Lyons, TS, Navalta, JW. New author guidelines in statistical reporting: embracing an era beyond p < .05. Int J Exerc Sci. 2020;13(1):15.Google ScholarPubMed
Wasserstein, RL, Schirm, AL, Lazar, NA. Moving to a World Beyond “p < 0.05”, vol. 73. Taylor & Francis; 2019:119.Google Scholar
Hardwicke, TE, Salholz-Hillel, M, Malički, M, Szűcs, D, Bendixen, T, Ioannidis, JPA. Statistical guidance to authors at top-ranked journals across scientific disciplines. Am Stat. 2023;77(3):239247.CrossRefGoogle Scholar
Smith, R. Statistical Review for Medical Journals, Journal’s Perspective, vol. 7. Hoboken, NJ: John Wiley & Sons, Encyclopedia of Biostatistics, 2005.CrossRefGoogle Scholar
JAMA Network. Instructions for authors. Accessed June 28, 2023.Google Scholar
AHA/ASA journals. Statistical reporting recommendations. Accessed June 28, 2023.Google Scholar
PLOS ONE. Submission guidelines. Accessed June 28, 2023.Google Scholar
Nature Scientific Reports. Submission guidelines. Accessed June 28, 2023.Google Scholar
Schulz, KF, Altman, DG, Moher, D, CONSORT Group. 2010 Statement: updated guidelines for reporting parallel group randomized trials. Ann Intern Med. 2010;152(11):726732.CrossRefGoogle ScholarPubMed
von Elm, E, Altman, DG, Egger, M, Pocock, SJ, Gøtzsche, PC, Vandenbroucke, JP. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med. 2007;147(8):573577.CrossRefGoogle ScholarPubMed
Page, MJ, McKenzie, JE, Bossuyt, PM, et al. The PRISMA 2020 statement: updated guideline for reporting systematic reviews. PLoS Med. 2021;18(3):e1003583.CrossRefGoogle ScholarPubMed
Collins, GS, Reitsma, JB, Altman, DG, Moons, KG. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. Ann Intern Med. 2015;162(1):5563.CrossRefGoogle ScholarPubMed
Bossuyt, PM, Reitsma, JB, Bruns, DE, et al. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. BMJ. 2015;351:h5527.CrossRefGoogle ScholarPubMed
Stevens, LM, Mortazavi, BJ, Deo, RC, Curtis, L, Kao, DP. Recommendations for reporting machine learning analyses in clinical research. Circ Cardiovasc Qual Outcomes. 2020;13(10):e006556.CrossRefGoogle ScholarPubMed
Percie du Sert, N, Hurst, V, Ahluwalia, A, et al. The ARRIVE guidelines 2.0: updated guidelines for reporting animal research. PLoS Biol. 2020;18(7):e3000410.CrossRefGoogle ScholarPubMed
Equator Network reporting guideline library. Accessed June 28, 2023.Google Scholar
Lang, TA, Altman, DG. Basic statistical reporting for articles published in biomedical journals: the “Statistical analyses and methods in the published literature” or the SAMPL guidelines. Int J Nurs Stud. 2015;52(1):59.CrossRefGoogle ScholarPubMed
Bailar, JC, 3rd Mosteller, F. Guidelines for statistical reporting in articles for medical journals. Amplifications and explanations. Ann Intern Med. 1988;108(2):266273.CrossRefGoogle ScholarPubMed
Altman, DG. Statistical reviewing for medical journals. Stat Med. 1998;17(23):26612674.3.0.CO;2-B>CrossRefGoogle ScholarPubMed
Greenwood, DC, Freeman, JV. How to spot a statistical problem: advice for a non-statistical reviewer. BMC Med. 2015;13(1):270.CrossRefGoogle ScholarPubMed
Sedgwick, P. Convenience sampling. Bmj. 2013;347(oct25 2):f6304f6304.CrossRefGoogle Scholar
Cochran, WG. Sampling Techniques. Hoboken, NJ: John Wiley & Sons, 1977.Google Scholar
Adcock, C. Sample size determination: a review. J R Stat Soc Ser D. 1997;46(2):261283.Google Scholar
Lachin, JM. Introduction to sample size determination and power analysis for clinical trials. Control Clin Trials. 1981;2(2):93113.CrossRefGoogle ScholarPubMed
Ryan, TP. Sample Size Determination and Power. Hoboken, NJ: John Wiley & Sons, 2013.CrossRefGoogle Scholar
Lenth, RV. Some practical guidelines for effective sample size determination. Am Stat. 2001;55(3):187193.CrossRefGoogle Scholar
Wang, H, Chow, SC, Chen, M. A Bayesian approach on sample size calculation for comparing means. J Biopharm Stat. 2005;15(5):799807.CrossRefGoogle ScholarPubMed
Kunzmann, K, Grayling, MJ, Lee, KM, Robertson, DS, Rufibach, K, Wason, JMS. A review of Bayesian perspectives on sample size derivation for confirmatory trials. Am Stat. 2021;75(4):424432.CrossRefGoogle ScholarPubMed
Riley, RD, Ensor, J, Snell, KIE, et al. Calculating the sample size required for developing a clinical prediction model. BMJ. 2020;368:m441.CrossRefGoogle ScholarPubMed
Riley, RD, Snell, KIE, Ensor, J, et al. Minimum sample size for developing a multivariable prediction model: part I - continuous outcomes. Stat Med. 2019;38(7):12621275.CrossRefGoogle ScholarPubMed
Riley, RD, Snell, KI, Ensor, J, et al. Minimum sample size for developing a multivariable prediction model: PART II - binary and time-to-event outcomes. Stat Med. 2019;38(7):12761296.CrossRefGoogle ScholarPubMed
Riley, RD, Debray, TPA, Collins, GS, et al. Minimum sample size for external validation of a clinical prediction model with a binary outcome. Stat Med. 2021;40(19):42304251.CrossRefGoogle ScholarPubMed
Degtiar, I, Rose, S. A review of generalizability and transportability. Annu Rev Stat Appl. 2023;10(1):501524.CrossRefGoogle Scholar
Chen, IY, Pierson, E, Rose, S, Joshi, S, Ferryman, K, Ghassemi, M. Ethical machine learning in healthcare. Annu Rev Biomed Data Sci. 2021;4(1):123144.CrossRefGoogle ScholarPubMed
Lakens, D. Sample size justification. Collabra Psychol. 2022;8(1):33267.CrossRefGoogle Scholar
Hoffmann, TC, Glasziou, PP, Boutron, I, et al. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide. BMJ. 2014;348(mar07 3):g1687g1687.CrossRefGoogle ScholarPubMed
Klein, JP, Van Houwelingen, HC, Ibrahim, JG, Scheike, TH. Handbook of Survival Analysis. FL: CRC Press Boca Raton,; 2014.Google Scholar
Shrier, I, Platt, RW. Reducing bias through directed acyclic graphs. BMC Med Res Methodol. 2008;8(1):70.CrossRefGoogle ScholarPubMed
Moher, D, Hopewell, S, Schulz, KF, et al. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340(mar23 1):c869c869.CrossRefGoogle ScholarPubMed
Mills, EJ, Chan, AW, Wu, P, Vail, A, Guyatt, GH, Altman, DG. Design, analysis, and presentation of crossover trials. Trials. 2009;10(1):27.CrossRefGoogle ScholarPubMed
Jones, B, Kenward, MG. Design and Analysis of Cross-over Trials. New York City, NY: CRC press, 2014.CrossRefGoogle Scholar
Senn, SS. Cross-Over Trials in Clinical Research, vol. 5. Hoboken, NJ: John Wiley & Sons, 2002.CrossRefGoogle Scholar
Bernal, JL, Cummins, S, Gasparrini, A. Interrupted time series regression for the evaluation of public health interventions: a tutorial. Int J Epidemiol. 2017;46(1):348355.Google Scholar
Wagner, AK, Soumerai, SB, Zhang, F, Ross-Degnan, D. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther. 2002;27(4):299309.CrossRefGoogle ScholarPubMed
Kim, K, Bretz, F, Cheung, YKK, Hampson, LV. Handbook of Statistical Methods for Randomized Controlled Trials. New York City, NY: CRC Press, 2021.CrossRefGoogle Scholar
Molenberghs, G, Fitzmaurice, G, Kenward, MG, Tsiatis, A, Verbeke, G. Handbook of Missing Data Methodology. New York City, NY: CRC Press, 2014.CrossRefGoogle Scholar
Harrell, FE. Regression Modeling Strategies: with Applications to Linear Models, Logistic Regression, and Survival Analysis, vol. 608. New York City, NY: Springer, 2001.CrossRefGoogle Scholar
Fedorov, V, Mannino, F, Zhang, R. Consequences of dichotomization. Pharm Stat. 2009;8(1):5061.CrossRefGoogle ScholarPubMed
Bennette, C, Vickers, A. Against quantiles: categorization of continuous variables in epidemiologic research, and its discontents. BMC Med Res Methodol. 2012;12(1):15.CrossRefGoogle ScholarPubMed
Morgan, TM, Elashoff, RM. Effect of categorizing a continuous covariate on the comparison of survival time. J Am Stat Assoc. 1986;81(396):917921.CrossRefGoogle Scholar
Thoresen, M. Spurious interaction as a result of categorization. BMC Med Res Methodol. 2019;19(1):18.CrossRefGoogle ScholarPubMed
Weissgerber, TL, Garcia-Valencia, O, Garovic, VD, Milic, NM, Winham, SJ. Why we need to report more than ‘Data were analyzed by t-tests or ANOVA’. Elife. 2018;7:e36163. doi: 10.7554/eLife.36163 CrossRefGoogle ScholarPubMed
Rubin, M. When to adjust alpha during multiple testing: a consideration of disjunction, conjunction, and individual testing. Synthese. 2021;199(3-4):1096911000.CrossRefGoogle Scholar
Thomas, L, Peterson, ED. The value of statistical analysis plans in observational research: defining high-quality research from the start. JAMA. 2012;308(8):773774.CrossRefGoogle Scholar
Weissgerber, TL, Milic, NM, Winham, SJ, Garovic, VD. Beyond bar and line graphs: time for a new data presentation paradigm. PLoS Biol. 2015;13(4):e1002128.CrossRefGoogle ScholarPubMed
Westreich, D, Greenland, S. The table 2 fallacy: presenting and interpreting confounder and modifier coefficients. Am J Epidemiol. 2013;177(4):292298.CrossRefGoogle ScholarPubMed
Schor, S, Karten, I. Statistical evaluation of medical journal manuscripts. JAMA. 1966;195(13):11231128.CrossRefGoogle ScholarPubMed
Gardner, MJ, Bond, J. An exploratory study of statistical assessment of papers published in the british medical journal. JAMA. 1990;263(10):13551357.CrossRefGoogle ScholarPubMed
Murray, GD. The task of a statistical referee. Br J Surg. 1988;75(7):664667.CrossRefGoogle ScholarPubMed
Gardner, MJ, Altman, DG, Jones, DR, Machin, D. Is the statistical assessment of papers submitted to the “British medical journal” effective? BMJ. 1983;286(6376):14851488.CrossRefGoogle Scholar
Hardwicke, TE, Goodman, SN. How often do leading biomedical journals use statistical experts to evaluate statistical methods? The results of a survey. PLoS One. 2020;15(10):e0239598.CrossRefGoogle ScholarPubMed
Hardwicke, TE, Frank, MC, Vazire, S, Goodman, SN. Should psychology journals adopt specialized statistical review? Advances in Methods and Practices in Psychological Science, 2019;2(3):240249.CrossRefGoogle Scholar
Goodman, SN, Altman, DG, George, SL. Statistical reviewing policies of medical journals: caveat lector? J Gen Intern Med. 1998;13(11):753756.CrossRefGoogle ScholarPubMed
Hug, SE, Aeschbach, M. Criteria for assessing grant applications: a systematic review. Palgrave Commun. 2020;6(1):115.CrossRefGoogle Scholar
Falk-Krzesinski, HJ, Tobin, SC. Let me count the ways: a comparison of research grant proposal review criteria across US federal funding agencies. J Res Adm. 2015;46(2):7994.Google ScholarPubMed
Publons. Grant Review in Focus. London, UK: Clarivate Analytics, 2019. Scholar
NIH Center for Scientific Review. Guidelines and Templates. Accessed June 28, 2023.Google Scholar
Spratt, H, Fox, EE, Shara, N, Mazumdar, M. Strategies for success: early-stage collaborating biostatistics faculty in an academic health center. Am Stat. 2017;71(3):220230.CrossRefGoogle Scholar
Waller, LA. Documenting and evaluating data science contributions in academic promotion in departments of statistics and biostatistics. Am Stat. 2018;72(1):1119.CrossRefGoogle Scholar
Figure 0

Table 1. Checklist guide of items to consider in biostatistical review of clinical and translational manuscripts