This response addresses several Letters to the Editor (LTEs) that provided feedback on the 2019 AJPA Commentary “Data Sharing in Biological Anthropology: Guiding Principles and Best Practices” authored by C. J. M. and T. T. All of those who participated in the 2019 Workshop on Data Sharing in Biological Anthropology were shown a draft of the current document and had an opportunity to provide feedback before publication. We include the “2019 Workshop on Data Sharing in Biological Anthropology” as a coauthor on this response and we list all workshop participants and their affiliations to demonstrate the diversity of voices represented at the workshop that resulted in the original AJPA Commentary by C. J. M. and T. T.

This response has the following goals to: (a) emphasize that Mulligan and Turner (2019) was only meant to be the beginning of more detailed work that is needed to institute satisfactory data archiving and sharing practices in the field of Biological Anthropology in the United States, (b) summarize and analyze the perspectives of the letter writers, and (c) address the potential next steps in this discussion based on input from the letter writers.

The LTEs unanimously supported the importance of more open, consistent, and standardized data sharing in Biological Anthropology. They saw the commentary from 2019 as important and timely. Wagner's LTE pointed out that Mulligan and Turner (2019) was initially locked behind the journal paywall. As quickly as possible, C. J. M. and T. T. worked with the journal and the publisher to provide free access to the article as they had always intended. Given that the article focused on the need for more open sharing of data, we recognize the mixed message this error may have sent. This experience highlights the need for authors to ensure that access to their articles, or datasets, is implemented in the way they intend .

We are grateful to all the LTE writers for highlighting priorities for continued discussion on data access and sharing in Biological Anthropology. We sympathize with Wagner's frustration that Mulligan and Turner (2019) addressed issues at a high level without specific actionable or enforceable suggestions. The primary goal of Mulligan and Turner (2019) was to signal to the Biological Anthropology community that it is important to examine its data archiving and sharing habits and to develop policies aligned with emerging best practices exemplified by other data‐driven sciences where possible. The workshop helped reaffirm that each subdiscipline has certain unique data sharing issues. More specific guidelines will need to reflect these differences, but the 2‐day workshop was not enough time to articulate, let alone understand and integrate, all of this variation. We therefore realize that continued dialogue within the field is needed to resolve the specifics. We also recognize the validity of Wagner's concern that Mulligan and Turner (2019) could mistakenly be treated as a complete solution rather than the start of an intensive process for the development of community standards and cultural change among numerous groups of stakeholders. We stand with her in calling to action the Biological Anthropology research community as we come together to begin this process.

We believe the research community will be responsive to our call for expanded data access and sharing because the fundamental purpose of this effort is to accelerate the advance of robust scientific knowledge in an inclusive way. Data sharing should lead to accelerated advances because it will result in better and broader data access. There are at least two mechanisms by which better data access can accelerate scientific progress. First, researchers can better evaluate the relative robusticity of new or competing scientific hypotheses when they have access to underlying data. Clearer insight into the relative strengths and weaknesses of hypotheses will guide researchers in the field to ask and study the most critical questions sooner, without getting sidetracked by ideas that fail to find strong support. Second, as data sharing becomes more common and effective, there will be an increase in the ability to assess new questions by combining and repurposing previously published data. The use of previously published data to address new questions means that the relevance of previously published data can expand beyond the research scope envisioned by the original authors. Use of previously published data will allow new studies to be conducted without collecting more data and will also will help researchers who collect novel datasets to interrogate any surprising implications of their observations more rapidly and thoroughly. This capacity will empower reviewers and editors to enforce higher standards of evidence in the research they accept for publication in the journals they oversee.

Wagner presented a list of unanswered questions that are an excellent starting point for the next phases and we welcome volunteers to help us address each one. In particular, developing guidance to help members evaluate the strengths and weaknesses of repositories is an important future issue. We recognize that more funding may be required to ignite the next phase of work and encourage interested individuals to come forward to discuss how this might be structured and what resources it will require .

Some of the LTEs emphasized particular challenges and risks that need to be addressed before more effective sharing can be accomplished. In particular, McDade's LTE points out the risk for “deductive disclosure,” whereby names of individuals represented in anonymized data can be deduced from the unique combinations of certain remaining variables in the dataset. There are various options for reducing such risk, but the decisions on how to approach deidentifying data can also substantially limit the utility of the data for anything other than testing reproducibility of the original study. Discussions on de‐identifying data and post hoc re‐identification of study participants are already underway in other research contexts such as genetic databases and biobanking. A set of recommendations for the standard approaches for de‐identification and the scenarios in which to choose one over another would be a useful resource that a subcommittee could research and compile .

An issue that was viewed differently by different LTEs was the impact of more comprehensive sharing and archiving on early career scholars. McDade worried that the extra effort required to effectively archive data could hinder progress of early career investigators, while Leigh saw the outcome of more effective data sharing as a boon to the same cohort by leveling the playing field and ending the days when senior scientists could monopolize data and use access to it as a currency for authorship. Elton also commented on the tendency of researchers to hold back their data due to the perception that they deserve the benefit of any remaining scholarly value because they spent time, effort, and money to collect the data. Elton sees this tendency as a cultural norm that will take time to shift, but which can shift due to technology and digital infrastructure that makes sharing more pragmatic.

Returning to McDade's concern about the burden that strict archiving requirements put on emerging scholars, it is our belief that what makes this task most burdensome is the lack of clearly outlined standards for, and training in, effective data curation practices for a project life cycle. In other words, as discussed in Mulligan and Turner (2019), data management plans should be designed before a project even starts in order to anticipate what data will need to be archived, what metadata will be necessary for data to be interpreted and reused, and in what repositories data can be stored. Having a plan that addresses these issues is in line with current federal funding agency requirements. Articulating, and then following, a protocol for how data will be curated can markedly reduce the burden that comes when the article is about to be published. Data management plans also benefit the research group in cases where data collection may be halted for years and then resumed or where there is a substantial delay between completion of data collection and analysis/interpretation. One of the participants (A. M. P.) has an affiliation with the Inter‐university Consortium for Political and Social Research (ICPSR). ICPSR has a repository providing leadership and training in emerging archiving practices. Based out of the University of Michigan and supported by more than 780 member institutions, the ICPSR has developed a robust technical and curatorial infrastructure to handle sensitive personal and health data. They also offer training opportunities and materials for scholars, including scholarships and other forms of support for acquiring data management skills. Some of these training opportunities are specifically targeted at emerging researchers. While this repository is not suitable for all Biological Anthropology datasets, it offers a model of knowledge support and infrastructure that should be on the minds of our colleagues as data sharing initiatives develop. As another example of a data repository that incorporates project life cycle features and support for early career scholars, MorphoSource is a Duke University‐affiliated, NSF‐supported, open access repository of 3D datasets with scan data from tens of thousands of individuals and thousands of species of primates and other mammals (as well as other vertebrates, invertebrates, plants, and protists). MorphoSource was launched in 2013 and is directed by the corresponding author of this response (D. M. B.). Over 1,200 MorphoSource users are graduate students. Already a number of dissertations and Masters theses have extracted substantial amounts of measurement data from specimens downloaded from MorphoSource. Students using MorphoSource for dissertation data can now analyze hundreds or thousands of high‐fidelity scans downloaded over a number of days/weeks instead of spending months and thousands of dollars traveling to museums or generating their own datasets as was the case less than a decade ago. Digital data sharing magnifies inclusiveness in data access in a way never before possible .

Another critical need is comprehensively reasoned criteria for choosing repositories that reflect subfield needs and issues related to specific datasets. We recognize the shortcoming of the link to repositories provided by Mulligan and Turner (2019). However, developing a well‐researched list and guidelines for repository selection was beyond the scope of the workshop. While there exist a number of repository registries (e.g., 3data.org, FAIRsharing.org), these still present a large number of options. Interpreting the features of each repository is complicated and is not straightforward without guidance. MorphoSource and ICPSR are examples of “domain‐specialized” repositories. However, there are also generalized repositories. It is not always clear when to use a specialized or a generalized repository. Dryad is a nonprofit, member‐supported generalized repository that was started by researchers at the University of North Carolina ‐ Chapel Hill and now has global adoption. Dryad is trying to help clarify the best place for preserving data by creating connections between repositories with different strengths for more rich and linked data preservation. In particular, Dryad recently announced collaboration with Zenodo, another noncommercial, relatively generalized achiving platform.

While Trustworthy Repositories Audit & Certification (TRAC) and other certifications can be indicators of repository quality, there are currently no simple solutions and the landscape for data archiving/sharing resources is changing quickly. For data sharing to be most effective, subfield representatives need to research and recommend repositories according to various criteria (potentially using TRAC as a guideline). Investigation of repository services must go hand‐in‐hand with research into and recommendations for subfield‐specific workflows for data management that guide handling and curation of data from the observation/measurement stage to the endpoint of archiving, publishing, and sharing.

We think early career scholars are well poised to help design and usher‐in the new culture of data sharing. With less investment in the old ways, more to gain from open access, and a better handle on emerging technologies, their vision and design ideas are needed and welcome.

The need to facilitate cultural change is not just a matter of the research community making up its mind that data sharing is good and necessary. It is about understanding the practical limitations of available infrastructure and the incentive structures that sustain the current culture. Only when grounded with this information can we begin to advocate for resources that enhance both the ability and tendency to share data. Useful policies and sharing platforms must address issues that hinder data sharing and provide incentives to amplify it. In the case of MorphoSource, not only researchers, but also museum curators, educators, and the general public have to be considered as active users of the platform. The platform is therefore designed to allow all stakeholders to collaborate in contributing and managing data according to ethical expectations and requirements for effective preservation of data. More resources like this are needed for the data of other subfields. Various stakeholders can help design and implement incentive structures including repositories, journals, and universities. The transition to a culture in which data sharing is supported, expected and rewarded is beginning to happen. Promotion committees are beginning to recognize that data sharing and reuse of data by third parties are important evidence of an individual's scholarly impact, but there is still a long way to go .

To close, we reiterate our gratitude toward those contributing LTEs on Mulligan and Turner (2019) for their help in further focusing the discussion toward next steps, to NSF for recognizing the importance of this discussion through its support of our workshop, and to the Biological Anthropology community for the steps they have started to take in the right direction. We look forward to striving toward better data sharing practices with our colleagues and partner institutions to advance science and the human condition.

该回复针对几封致编辑（LTE）的信函，这些信函提供了由CJM和TT撰写的2019年AJPA评论``生物人类学中的数据共享：指导原则和最佳实践''的反馈意见所有参加了2019年AJPA数据共享研讨会的人向生物人类学展示了当前文档的草案，并有机会在出版前提供反馈。我们将“ 2019年生物人类学数据共享讲习班”作为对此回应的合著者，我们列出了所有讲习班参加者及其所属团体，以展示参加讲习班的声音的多样性，从而产生了由CJM和TT撰写的原始AJPA评论

该回应具有以下目标：（a）强调Mulligan和Turner（2019）仅是开始更详细的工作的开始，需要开展这项工作才能在美国生物人类学领域建立令人满意的数据归档和共享实践各州，（b）总结并分析了信作者的观点，（c）根据信作者的意见处理了本次讨论中可能采取的下一步措施。

LTE一致支持生物人类学中更加开放，一致和标准化的数据共享的重要性。他们认为2019年的评论很重要且很及时。Wagner的LTE指出，Mulligan和Turner（2019）最初被锁定在期刊付费专区的后面。CJM和TT会尽快与期刊和出版商合作，以免费提供他们一直想要的文章。鉴于本文关注的是需要更开放地共享数据，我们认识到此错误可能发送的混合消息。这种经验凸显了作者必须确保以预期方式实现对他们的文章或数据集的访问。

我们感谢所有LTE作家强调了优先重点，以继续讨论生物人类学中的数据访问和共享问题。我们同情瓦格纳的沮丧，即Mulligan和Turner（2019）在没有具体可行或可执行建议的情况下高水平解决了问题。Mulligan和Turner（2019）的主要目标是向生物人类学界传达信号，重要的是检查其数据存档和共享习惯，并在可能的情况下制定与新兴最佳实践相一致的政策，这些最佳实践是其他数据驱动科学的例证。讲习班有助于重申每个子学科都有某些独特的数据共享问题。需要更具体的准则来反映这些差异，但是为期2天的研讨会尚不足以阐明，更不用说理解和整合了，所有这些变化。因此，我们认识到需要在领域内继续对话以解决具体问题。我们也认识到瓦格纳（Wagner）担心将Mulligan和Turner（2019）误认为是一个完整的解决方案是正确的，而不是为众多利益相关者群体制定社区标准和文化变革的密集过程的开始。在我们共同开始这一过程的过程中，我们与她一道呼吁采取行动的生物人类学研究界。令人担忧的是，Mulligan和Turner（2019）可能被误认为是一个完整的解决方案，而不是为众多利益相关者群体制定社区标准和文化变革的密集过程的开始。在我们共同开始这一过程的过程中，我们与她一起呼吁生物人类学研究界采取行动。令人担忧的是，Mulligan和Turner（2019）可能被误认为是一个完整的解决方案，而不是为众多利益相关者群体制定社区标准和文化变革的密集过程的开始。在我们共同开始这一过程的过程中，我们与她一道呼吁采取行动的生物人类学研究界。

我们相信研究界将响应我们关于扩大数据访问和共享的号召，因为这项工作的根本目的是以包容性的方式加速强大的科学知识的发展。数据共享应导致更快的进步，因为它将导致更好和更广泛的数据访问。至少有两种机制可以使更好的数据访问加速科学进步。首先，研究人员在访问基础数据时可以更好地评估新的或竞争性科学假设的相对健壮性。对假设的相对优势和劣势的更清晰了解将指导该领域的研究人员尽快提出和研究最关键的问题，而不会因无法找到有力支持的想法而步履维艰。第二，随着数据共享变得越来越普遍和有效，通过合并和重新利用以前发布的数据来评估新问题的能力将会提高。使用先前发布的数据来解决新问题意味着，先前发布的数据的相关性可以扩展到超出原始作者所设想的研究范围。使用以前发布的数据将使进行新的研究而无需收集更多的数据，也将帮助收集新颖数据集的研究人员更快速，更彻底地审视其观察结果的任何令人惊讶的含义。这种能力将使审稿人和编辑能够在接受的研究中执行更高的证据标准，以便在其所管理的期刊中发表论文。通过合并和重新利用以前发布的数据来评估新问题的能力将有所提高。使用先前发布的数据来解决新问题意味着，先前发布的数据的相关性可以扩展到超出原始作者所设想的研究范围。使用以前发布的数据将使进行新的研究而无需收集更多的数据，也将帮助收集新颖数据集的研究人员更快速，更彻底地审视其观察结果的任何令人惊讶的含义。这种能力将使审稿人和编辑能够在接受的研究中执行更高的证据标准，以便在其所管理的期刊中发表论文。通过合并和重新利用以前发布的数据来评估新问题的能力将有所提高。使用先前发布的数据来解决新问题意味着，先前发布的数据的相关性可以扩展到超出原始作者所设想的研究范围。使用以前发布的数据将使进行新的研究而无需收集更多的数据，也将帮助收集新颖数据集的研究人员更快速，更彻底地审视其观察结果的任何令人惊讶的含义。这种能力将使审稿人和编辑能够在接受的研究中执行更高的证据标准，以便在其所管理的期刊中发表论文。使用先前发布的数据来解决新问题意味着，先前发布的数据的相关性可以扩展到超出原始作者所设想的研究范围。使用以前发布的数据将使进行新的研究而无需收集更多的数据，也将帮助收集新颖数据集的研究人员更快速，更彻底地审视其观察结果的任何令人惊讶的含义。这种能力将使审稿人和编辑能够在接受的研究中执行更高的证据标准，以便在其所管理的期刊中发表论文。使用先前发布的数据来解决新问题意味着，先前发布的数据的相关性可以扩展到超出原始作者所设想的研究范围。使用以前发布的数据将使进行新的研究而无需收集更多的数据，也将帮助收集新颖数据集的研究人员更快速，更彻底地审视其观察结果的任何令人惊讶的含义。这种能力将使审稿人和编辑能够在接受的研究中执行更高的证据标准，以便在其所管理的期刊中发表论文。使用以前发布的数据将使进行新的研究而无需收集更多的数据，也将帮助收集新颖数据集的研究人员更快速，更彻底地审视其观察结果的任何令人惊讶的含义。这种能力将使审稿人和编辑能够在接受的研究中执行更高的证据标准，以便在其所管理的期刊中发表论文。使用以前发布的数据将使进行新的研究而无需收集更多的数据，也将帮助收集新颖数据集的研究人员更快速，更彻底地审视其观察结果的任何令人惊讶的含义。这种能力将使审稿人和编辑能够在接受的研究中执行更高的证据标准，以便在其所管理的期刊中发表论文。

瓦格纳（Wagner）提出了一系列悬而未决的问题，这些问题是下一阶段的良好起点。我们欢迎志愿者帮助我们解决每个问题。尤其是，制定指南以帮助成员评估存储库的优缺点是未来的重要课题。我们认识到，可能需要更多的资金来启动下一阶段的工作，并鼓励有兴趣的个人站出来讨论如何组织这项工作以及将需要哪些资源。

一些LTE强调了在完成更有效共享之前需要解决的特定挑战和风险。麦克戴德（McDade）的LTE特别指出了“演绎披露”的风险，据此，可以从数据集中某些剩余变量的唯一组合中推断出匿名数据中所代表的个人的姓名。有多种方法可以降低这种风险，但是关于如何确定身份的数据的决策也可能极大地限制数据在测试原始研究的可重复性以外的用途。在其他研究背景下，例如基因数据库和生物库，已经开始进行关于去识别数据和事后重新识别研究对象的讨论。一组有关取消身份识别的标准方法的建议以及一个选择另一个方案的方案将是小组委员会可以研究和汇编的有用资源。

不同的LTE对问题的看法不同，这是更全面的共享和归档对早期职业学者的影响。McDade担心有效存档数据所需的额外努力可能会阻碍早期职业调查人员的进步，而Leigh则通过公平竞争环境并结束了高级科学家可以垄断的时代，看到了更有效的数据共享的成果，这对同一组人来说是一个福音。数据并将其访问权用作创作的一种货币。埃尔顿还评论说，由于人们认为他们应该花时间，精力和金钱来收集数据，因此他们应该获得任何剩余的学术价值，因此他们倾向于保留其数据。埃尔顿（Elton）将此趋势视为一种文化规范，需要一段时间才能转变，

回到McDade对严格归档要求给新兴学者带来的负担的担忧，我们相信使这项任务最繁重的是缺乏针对项目生命周期的有效数据管理实践的明确概述的标准和培训。换句话说，正如Mulligan和Turner（2019）所讨论的那样，应该在项目开始之前就设计数据管理计划，以预测需要存储哪些数据，解释和重用数据所需的元数据，以及可以在哪些存储库中存储数据。制定解决这些问题的计划符合当前联邦资助机构的要求。阐明然后跟随，有关如何整理数据的协议可以显着减少文章即将发布时的负担。如果数据收集可能会中断数年然后又恢复，或者在完成数据收集与分析/解释之间存在相当大的延迟，则数据管理计划也会使研究小组受益。参与者之一（AMP）与大学间政治与社会研究联盟（ICPSR）有从属关系。ICPSR拥有一个存储库，可提供有关新兴归档实践的领导力和培训。ICPSR立足于密歇根大学，并得到780多个会员机构的支持，已开发出强大的技术和管理基础架构来处理敏感的个人和健康数据。他们还为学者提供培训机会和材料，包括奖学金和其他形式的数据管理技能支持。其中一些培训机会专门针对新兴研究人员。虽然此存储库不适用于所有生物人类学数据集，但它提供了知识支持和基础架构的模型，随着数据共享计划的发展，我们的同事应该考虑该模型。作为结合项目生命周期功能和对早期职业学者的支持的数据存储库的另一个示例，MorphoSource是杜克大学附属，NSF支持的3D数据集的开放访问存储库，其中包含来自成千上万个人和成千上万个人的扫描数据灵长类动物和其他哺乳动物（以及其他脊椎动物，无脊椎动物，植物和原生生物）的物种。MorphoSource于2013年启动，由该回复的相应作者（DMB）指导。超过1200个MorphoSource用户是研究生。已经有许多论文和硕士学位从从MorphoSource下载的样本中提取了大量的测量数据。使用MorphoSource进行学位论文数据的学生现在可以分析在数天/周内下载的数百或数千个高保真扫描，而不用像十年前那样花费数月和数千美元前往博物馆或生成自己的数据集。已经有许多论文和硕士学位从从MorphoSource下载的样本中提取了大量的测量数据。使用MorphoSource进行学位论文数据的学生现在可以分析在数天/周内下载的数百或数千个高保真扫描，而不用像十年前那样花费数月和数千美元前往博物馆或生成自己的数据集。已经有许多论文和硕士学位从从MorphoSource下载的样本中提取了大量的测量数据。使用MorphoSource进行学位论文数据的学生现在可以分析在数天/周内下载的数百或数千个高保真扫描，而不用像十年前那样花费数月和数千美元前往博物馆或生成自己的数据集。数字数据共享以前所未有的方式扩大了数据访问的包容性。

另一个关键需求是用于选择可反映子字段需求和与特定数据集有关的问题的存储库的综合推理标准。我们认识到Mulligan和Turner（2019）提供的存储库链接的缺点。但是，为仓库选择制定详细研究的清单和指南超出了研讨会的范围。尽管存在许多存储库注册表（例如3data.org，FAIRsharing.org），但它们仍然提供了大量选项。解释每个存储库的功能很复杂，而且没有指导也不容易。MorphoSource和ICPSR是“域专用”存储库的示例。但是，也有通用存储库。何时使用专用或通用存储库并不总是很清楚。树精是非营利组织，成员支持的通用存储库由北卡罗来纳大学教堂山分校的研究人员启动，现已在全球范围内采用。Dryad试图通过在具有不同强度的存储库之间建立连接以更丰富，更链接地保存数据，从而帮助澄清保存数据的最佳位置。特别是，Dryad最近宣布与Zenodo合作，Zenodo是另一个非商业性的，相对通用的成就平台。

尽管可信赖的存储库审核和认证（TRAC）和其他认证可以指示存储库的质量，但是目前还没有简单的解决方案，并且数据归档/共享资源的格局正在迅速变化。为了使数据共享最有效，子领域的代表需要根据各种标准（可能使用TRAC作为指导原则）研究和推荐存储库。存储库服务的调查必须与针对数据管理的子领域特定工作流的研究和建议紧密结合，以指导从观察/测量阶段到归档，发布和共享端点的数据处理和管理。

我们认为，早期职业学者已经做好准备，可以帮助设计和引入新的数据共享文化。以旧方式进行的投资减少，可以从开放访问中获得更多收益，并且可以更好地利用新兴技术，因此，他们的愿景和设计思想必不可少且受到欢迎。

促进文化变革的需要不仅仅是研究界的事，它断定数据共享是良好且必要的。它是关于了解可用基础架构的实际局限性和维持当前文化的激励结构。只有以这些信息为基础，我们才能开始倡导能够增强共享数据的能力和趋势的资源。有用的策略和共享平台必须解决阻碍数据共享的问题，并提供激励来扩大数据共享。就MorphoSource而言，不仅研究人员，而且博物馆策展人，教育工作者和公众也必须被视为该平台的活跃用户。因此，该平台旨在允许所有利益相关者根据道德期望和有效保存数据的要求进行协作，以贡献和管理数据。其他子字段的数据需要更多此类资源。各种利益相关者可以帮助设计和实施激励结构，包括存储库，期刊和大学。开始向支持，期望和奖励数据共享的文化过渡。促进委员会已开始认识到，第三方共享数据和重用数据是个人学术影响力的重要证据，但还有很长的路要走。

最后，我们再次感谢那些在Mulligan和Turner（2019）上为LTE做出贡献的人，他们的帮助将讨论进一步集中到下一步，感谢NSF通过支持我们的讲习班和生物学来认识到这一讨论的重要性人类学界对于他们已开始朝正确方向采取的步骤进行了了解。我们期待与我们的同事和合作伙伴机构更好地共享数据，以促进科学和人类状况的发展。