Data and Methods Transparency

All Organization Science submissions received on or after August 1, 2025, will be governed by a set of policies and guidelines designed to increase research transparency and strengthen trust in the veracity of the papers we publish. We’ve designed these policies and guidelines to support our mission of advancing our scientific understanding of organizations, giving both authors and readers improved confidence that they can accurately evaluate data and analysis. We seek to balance these transparency objectives with the recognition that the diverse mix of empirical approaches that we publish have different practical and ethical constraints on disclosure. A key element of transparency is the acknowledgement that no empirical research can be perfect and that we should embrace transparent imperfections as better guidance for theory and policy as well as a challenge for future advancements. Valuable knowledge accumulates more reliably and efficiently when others can fully understand, trust, and build on prior work. Adhering to best practices for transparent data description in your specific methodology is crucial. This involves clearly documenting and communicating the nature, sources, and limitations of the data used. It is our hope that these guidelines are also useful to authors in documenting their research process for future reexamination.

General Policies and Guidelines

Presentation of Data and Setting

Authors should provide all necessary details for reviewers and editors to understand the empirical setting, processes, characteristics and distributions of the data, and the results. Submissions lacking these details can raise uncertainty of successful revision and thereby reduce the likelihood that a revision might be requested or that the initial submission will even be sent out for reviews. Sufficient detail and references should be made so others can replicate the work. Intentional misrepresentation, falsification, or knowingly making inaccurate statements are unethical and unacceptable.

Authors of quantitative papers are encouraged to present data and variables beyond simple means and standard deviations and should consider visual representations such as kernel density or bin scatter plots that reveal skewness, outliers, and bunching. Simple scatter plots between dependent variables and key predictors can help reveal which observations might drive results. Visual representation of individual observations is encouraged when possible. Qualitative studies can present supporting tables with additional examples that highlight the prevalence of key points. Visualizations through figures and tables can be highly effective in helping the review team confidently understand the nature of the data and its potential. Transparency within the submitted paper, not merely supplemental files, is key.

In addition, all authors should disclose the use of AI and AI-assisted technologies in the text or footnotes of their submission whenever it has been used.

Research Sequence

Organization Science welcomes and publishes empirical papers that follow deductive, inductive, abductive, and other discovery processes, including those with strong evidence against tested hypotheses. We therefore expect authors to openly and accurately present the sequence of events and actions through which results were reached, while allowing author discretion in how the paper is structured. Studies need not be presented in chronological sequence, for example, so long as the paper is explicit about the sequence in which they were conducted. This includes but is not limited to: theory development, data collection, analysis, and follow-up studies. Studies conducted but not included in the paper should be disclosed along with the reason for the exclusion. Changes to the research plan should be communicated as well, including adverse events such as data loss, non-compliance, unanticipated harm or ethical concerns, or uncontrollable changes to organizational or institutional environments that might impact the validity of the research design.

Participant and Vulnerable Party Protection

Authors should disclose any institutional review board (IRB) or related approvals required and granted by involved institutions. We encourage authors to directly address any clear ethical concerns in their studies, as IRB standards vary widely across institutions and are not perfect screening mechanisms. We also strongly encourage authors studying vulnerable or traditionally disadvantaged populations to seek input from scholars with that lived experience, as even the most thoughtful and well-intentioned scholars might miss problematic language or mischaracterizations.

Data Overlap with Other Papers

Authors should disclose to the editors upon submission any related authored papers with overlap in data and/or research question, whether published, previously rejected, of in working paper form. This information should be included in the submission letter, and authors are encouraged to cite and differentiate from related papers in the manuscript when it can be done without compromising the double-blind process. We consider the simultaneous submission of very similar papers to different journals without disclosure to be a significant violation of publication ethics.

Policies and Guidelines Most Relevant for Quantitative Data

Sharing and Publication of Code

Shared code clarifies for readers the precise operations through which data were cleaned, variables were generated, samples were set, and analytical models were implemented. The process of documenting and sharing code also helps authors identify their own errors or bugs during the review process and to clarify the often-discretionary choices made in the methods. Consequently, authors are expected to freely share the statistical code that produces their empirical results during the review process, upon request of the editor, except in those rare cases where code might breach confidentiality or other legal requirements. Authors of theoretical papers are also expected to freely share the code that produced their results, such as in simulation, upon submission or resubmission. This code is treated similar to an analytical proof — available for reviewers and readers to verify logic and conclusions. All code should include instructions and comments describing the steps that progress from raw data to final analysis, allowing both the authors and readers to match reported results, samples, and measurements in the paper with the lines that produced them.

In their initial submission, authors should report the sample size for all models and studied subgroups.¹ All data-related decisions, including how missing data were handled and whether outliers were detected (and how they were handled) should be clearly discussed. Authors should also report all measures created or collected, experimental conditions implemented, and any data exclusions imposed. Effect sizes for estimated coefficients should be presented and interpreted. When they are unable to accommodate all important details regarding data collection, variable creation, model choices and robustness tests in the main text, authors should submit a supplemental appendix with their submission that does so.

Sharing and Publication of Data

We require authors of accepted papers to publicly share data through which their code produces all reported results in both the main text and any supplemental appendices, unless there is a strong reason for withholding. These data, along with accompanying code, should be archived in a third-party data repository such as OSF or openICPSR. Authors can request that such data be delayed or embargoed after publication to protect the authors’ other ongoing projects, but this delay only applies to public posting and not to requests from the editorial board. Unless otherwise stated, parties downloading data and code posted for replication purposes are expected to only use such materials for that intended purpose.

Authors seeking an exception to data sharing requirements must request one at the time of submission, or in subsequent resubmissions if new data or variables are introduced. The requests should include alternative transparency plans acceptable to the handling editor, such as synthetic datasets, detailed instructions for generating a proprietary dataset, distributional and summary statistics to populate a model, or a replication plan that allows researchers to legally access and verify results with the data provider. Reasons for such exceptions include but are not limited to:

• Formal legal documents such as non-disclosure agreements (NDA) with data providers
• Data where anonymity is crucial and cannot be achieved in shared data
• Disclosures that violate formal laws and regulations as determined by university legal entities
• Disclosures that might endanger researchers or human subjects

We acknowledge that in some cases, such alternative plans are infeasible for legal, ethical, or practical reasons. We encourage all authors to seek NDAs that both allow for the verification of results by a designated third-party and ensure the right to publish research regardless of the conclusions reached by the researchers. These elements raise confidence that results are not censored and that a viable path to replication exists.

Preregistration Guidelines

Organization Science does not require the preregistration of experimental or other empirical work but emphasizes that preregistration can substantially increase confidence in evidential value and any claims of deductive tests of true ex ante hypotheses. Preregistration can also raise confidence by replicating and confirming the authors’ prior study results or pilot studies and strengthen claims that empirical results are not artifacts of ad hoc specification and sample choices. Any preregistration should be presented in the submitted paper, as should any deviations from the preregistration. We emphasize that unexpected or surprising results can be valuable but should be presented as such with ex-post explanations that may be obvious or simply speculative. Editors might request replication of these results in additional preregistered studies to address possible Type I errors.

OSF, AsPredicted, and other sites host confidential or anonymous preregistrations that can be linked to in the manuscript.

Policies and Guidelines Most Relevant for Qualitative Research

Transparency Regarding Data Gathered

Authors of qualitative studies are expected to provide as much detail about the study setting as possible without violating any confidentiality, anonymity, or legal agreements, while also ensuring researcher and participant safety. These details might include the geographic location, the organization type, when the data were collected, all types of data that were collected and for what purposes, and who collected the data. The reader should understand what population of study the findings are based on, as well as the reasoning for choosing that population.

In addition to detailed text descriptions, supporting tables that organize and document the different types of data gathered can be helpful, especially for studies with complex data collection efforts with different sources over time. When data collection involves interviews, the interview protocol and an anonymized description or table of interviewees should be included in an appendix. Here again, transparency within the submitted paper, not merely supplemental files, is key.

Transparency Regarding Data Analysis

Qualitative methods include many data gathering and analytical approaches. Authors should clearly delineate those approaches and the reasoning for choosing them. The details of these methods are crucial for evaluation, including the precise sequence of data acquisition and methods choices such as the organization, coding, and interpretation of your data.

Given the many approaches to qualitative data and analysis, clearly explain the analytical approach you use and why it is appropriate for your data and research question. Explain in detail how you organized, interpreted, and coded data using examples of how your coding framework emerged from the data. The explanation of the analytic steps taken should have a relationship to how the findings are presented — so that readers can map the methodological approach described to the findings presented. Clarity on how the data were organized, analyzed, and interpreted can also sometimes be conveyed through the use of supporting tables and figures. Communicate honestly and concisely with journal readers to make your theorizing process transparent in terms of the moves taken to advance from data analyses to developing theory. Such an explicit account might explain how you developed insights or made discoveries and what drove the process of theorizing from the data analyzed.

Transparency Regarding the Extent to Which Stated Findings Are Supported by Data

Be as clear as possible about how the stated findings are supported by your data, presenting a clear path from the raw data collected to your interpretations. We encourage sharing rich data such as full quotations and event descripts in the manuscript text, with additional (but not duplicative) examples shared through supporting tables. Confidential identifiers or pseudonyms for people, occupations, groups, and organizations can provide data transparency without compromising confidentiality. Such identifiers help the reader delineate how data are distributed within and across people, places, time periods, or field visits. If mixed methods are used, it is important to specify which methods address which research questions and ensure that the rationale for using multiple methods is clear and justified.

---

¹ For example, if a regression estimates how the effect of a treatment varies by whether a CEO is a White man, then the authors should report the number of CEOs who are White men vs. who are not.