Removing barriers to better data practices through capacity, opportunity, and motivation

通过能力、机会和动机消除更好的数据实践的障碍

• 批准号: 10199149
• 负责人: Andrew W Brown
• 金额: $ 10.49万
• 依托单位: TRUSTEES OF INDIANA UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-07-01 至 2022-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10199149
• 关键词: Adherence; Adopted; Awareness; Behavior; Behavioral; Biomedical Research; Biometry; Case Study; Chronic Disease; Code; Communication; Communities; Complement; Conscious; Consult; Data; Data Analyses; Data Science; Data Set; Discipline; Documentation; Education; Educational Materials; Ensure; Evaluation; FAIR principles; Faculty; Failure; Feedback; Funding; Generations; Goals; Health; Heterogeneity; Institution; Intention; Investments; Learning; Literature; Measurement; Methodology; Methods; Mind; Modality; Modeling; Motivation; National Institute of General Medical Sciences; Obesity; Outcome; Paper; Participant; Privatization; Process; Publications; Reading; Reporting; Reproducibility; Research; Research Design; Research Methodology; Research Personnel; Research Project Grants; Research Training; Resource Sharing; Science; Scientist; Self Assessment; Series; Standardization; Structure; System; Theory of Change; Thumb structure; Training; Training and Education; Trust; United States National Institutes of Health; Variant; Work; base; behavior change; data format; data management; data sharing; design; education research; flexibility; frontier; high standard; human subject; ideation; improved; insight; instructor; learning materials; lectures; mathematical sciences; pre-clinical; programs; success; trustworthiness; voucher; working group

PROJECT SUMMARY/ABSTRACT Transparent research is important to ensure scientific reports are accurate, to instill trust in the scientific process, and to increase the amount of scientific insight derived from a given research investment. Two important steps for transparent, trustworthy research is for scientists to share their data and their code, and to have another set of analysts verify their results. Many efforts to change investigator behavior around data and code sharing discuss complete workflows, best practices with standardized variables across the discipline, and high standards for FAIR principles. Unfortunately, much biomedical research in general, and preclinical or exploratory research in particular, is idiosyncratic. New outcomes are formed, new methods are created, and community standards may not exist. Furthermore, some researchers learn of data and code sharing after a research project has started, during a paper submission process, or only when someone asks for their data and code. Where possible, our educational materials will focus on best practices. However, given the heterogeneity of types and stages of research needing data and code to be shared, our proposed educational materials will also focus on encouraging better practices to avoid ‘making the perfect the enemy of the good.’ Educational materials will focus on making data and code publically available in a usable and interpretable format, even if not perfectly standardized or FAIR. Simple workflows will be devised for researchers to independently verify their own data and analyses to reduce errors in the literature. Because investigators need to engage with data and code verification at various stages of research, we will present the educational materials in three flows. 1) A complete start-to-finish workflow of data and code verification and sharing. 2) Special use cases, such as preparing data and code from a completed study, sharing code even when data are unable to be shared, and responding to data and code sharing requests after publication. 3) Discrete educational materials, including the lectures, case-studies, self-assessments, readings, and more, to be used outside of the module context as investigators may need. Each piece of educational material will be built upon behavior-change theory, namely the capability, opportunity, and motivation model of behavior change (COM- B). We will structure learning materials with the intention of changing investigator behavior by explicitly connecting materials to build investigator capability, outlining the many opportunities investigators have to verify and share their data and code, and enhancing motivation through positive and negative real-world case studies. Materials will be made in compliance with Sharable Content Object Reference Model (SCORM) standards for easy re-use and incorporation by other educators into their courses and trainings, and comply with accessibility practices. The audience will include investigators at all stages, with preliminary users being drawn from our existing NIH-funded educational programs in basic quantitative sciences and in behavioral chronic disease research, to enable a broad base of scientists to better share and verify code and data.

项目总结/摘要 透明的研究对于确保科学报告的准确性，灌输对科学的信任， 过程，并增加从给定的研究投资中获得的科学见解的数量。两 透明、可信的研究的重要步骤是科学家分享他们的数据和代码， 让另一组分析师验证他们的结果许多努力改变调查人员围绕数据的行为， 代码共享讨论完整的工作流程，跨学科标准化变量的最佳实践， 公平原则的高标准。不幸的是，大多数生物医学研究，临床前或 特别是探索性研究，是一种特殊的研究。新的成果形成，新的方法被创造， 社区标准可能不存在。此外，一些研究人员了解到数据和代码共享后， 研究项目已经开始，在论文提交过程中，或者只有当有人要求他们的数据时， 和代码。在可能的情况下，我们的教育材料将侧重于最佳做法。但鉴于 我们提出的教育研究类型和阶段的多样性需要共享数据和代码， 教材也将着重鼓励更好的做法，以避免“使完美成为好的敌人”。 教育材料将侧重于使数据和代码在可用和可解释的 格式，即使不完全标准化或公平。将为研究人员设计简单的工作流程， 独立核实自己的数据和分析，以减少文献中的错误。因为调查人员需要 为了在研究的各个阶段进行数据和代码验证，我们将提供教育性的 材料分为三个流程。1)完整的数据和代码验证与共享工作流程。（二） 特殊用例，例如从已完成的研究中准备数据和代码，即使数据 无法共享，发布后响应数据和代码共享请求。3)离散 教育材料，包括讲座、案例研究、自我评估、阅读材料等 研究人员可能需要的模块上下文之外的内容。每一份教育材料都将建立在 行为改变理论，即行为改变的能力、机会和动机模型（COM- B）。我们将组织学习材料，目的是通过明确地 连接材料以建立调查员能力，概述调查员必须 验证和分享他们的数据和代码，并通过积极和消极的现实世界的情况下，提高动机 问题研究材料将符合可共享内容对象参考模型（SCORM） 标准，便于其他教育工作者重新使用，并将其纳入其课程和培训中， 无障碍实践。听众将包括各个阶段的调查人员，初步用户包括 从我们现有的NIH资助的基础定量科学和行为科学教育计划中提取 慢性疾病研究，使广泛的科学家能够更好地共享和验证代码和数据。