Module 5: Sharing & Defending
Lesson Specification: Module 5.2
Lesson Specification: Module 5.2
Metadata
| Field | Value |
|---|---|
| Module | Module 5.2 |
| Title | Reproducibility & Open Science |
| Part | Part 1: Main Curriculum |
| Estimated Duration | 20 minutes |
| Nexus Tool(s) | None |
| Prerequisites | None |
Learning Objectives
By the end of this lesson, students will be able to:
- Define the replication crisis and its impact on academic credibility.
- Apply the FAIR data principles (Findable, Accessible, Interoperable, Reusable).
- Structure a research repository (code, data, and environment) for public release.
- Evaluate open-access publishing routes (Preprints, Gold OA, Green OA).
Key Concepts
Concept 1: The Replication Crisis
- A significant percentage of published studies across psychology, medicine, and economics cannot be replicated by independent researchers.
- Causes: p-hacking, publication bias (only positive results get published), and opaque methodologies (not sharing code/data).
- The Solution: Open Science.
Concept 2: The FAIR Data Principles
If you release data, it must be:
- Findable: Hosted on a persistent repository (like Zenodo or OSF) with a DOI, not just on a personal website.
- Accessible: Clear terms of use. Even if the data is restricted (e.g., medical records), the metadata must be accessible.
- Interoperable: Use standard, non-proprietary formats (e.g., CSV instead of an Excel file with macros).
- Reusable: Includes a clear license (e.g., CC-BY) and documentation (a Data Dictionary).
Concept 3: The Reproducible Repository
A reproducible code repository needs three things:
- The Code: Clean, commented scripts.
- The Environment: A
requirements.txt,environment.yml, or Dockerfile. (Code from 2020 won't run in 2026 without the environment). - The Data (or a toy dataset): If the real data is too large or private, provide a sample so the code can be tested.
Concept 4: Open Access (OA) Publishing
- Preprints: Uploading your manuscript to arXiv/bioRxiv before peer review. Fast, free, establishes priority.
- Gold OA: The author pays an Article Processing Charge (APC) to make the final paper free for everyone.
- Green OA: Publishing in a traditional journal, but self-archiving the accepted manuscript in a university repository.
Suggested Hooks & Motivation
- Pain Point: Trying to build upon a 3-year-old paper, but the author's code is a broken zip file linked to a dead Dropbox link, and the data is "available upon reasonable request" (but the author ignores your emails).
- Wow Moment: Seeing a perfectly packaged Zenodo repository that assigns a DOI to code, making the code itself a citable research artifact.
Sources & Further Reading
- Wilkinson, M. D., et al. (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data.
- Baker, M. (2016). 1,500 scientists lift the lid on reproducibility. Nature.