worldbank / dime-standards

Repository with resources for DIME's research standards and coding standards

DIME Standards

This repository includes DIME's commitments to high-quality research. In the spirit of transparency, we are developing and version-controlling DIME's standards on GitHub. This repository is maintained by the DIME Analytics team.

There are two components to this repository:

• DIME Research Standards
• DIME Coding Standards

DIME Research Standards

Pillar 1 - Research Ethics

• DIME Researchers must secure ethics approval from an institutional review board (IRB) and, if applicable, any relevant authority in the study location, for studies directly involving human subjects or using personally-identifying information.
• All DIME team members that handle personally-identifiable information must have up-to-date Human Subjects Research Certification.
• DIME staff must ensure confidentiality, privacy, and anonymity of study participants; study participants must have the opportunity to provide informed consent, and revoke that consent at any time.

For details and implementation resources, see the Research Ethics Guidelines

Pillar 2 - Research Transparency

• All DIME projects must be registered prior to receiving implementation funding.

For implementation resources, see the Research Transparency Guidelines

Pillar 3 - Research Reproducibility

• All DIME projects will use GitHub to document data work.
• DIME Research assistants will regularly participate in peer code review sessions.
• All DIME projects will have a master script that runs all the other scripts that are needed for the project, in order.
• Computational Reproducibility must be verified by DIME Analytics prior to publication for all DIME Working Papers and academic publications.

For implementation resources, see the Research Reproducibility Guidelines

Pillar 4 - Data Security

• All personally-identifiable data must be stored and transferred securely, including in communication with field staff. 
  • All servers for data collection must be encrypted both in transit and at rest.
  • All identified data must be stored only in securely encrypted locations, and must always be encrypted when shared, even if shared only within the project team
  • All DIME projects must follow the DIME Data Back-up Protocols

For implementation resources, see the Data Security Guidelines

Pillar 5 - Data Publication

• Data must be de-identified prior to publication, DIME Analytics can assist with assessing the risk of statistical disclosure.
• All DIME publications must include a replication package, including a citation or reference to the data in the World Bank Microdata Catalog and a GitHub repository containing the code required for replication of the paper.
• Data must be deposited into the World Bank's Development Data Hub no later than six months after it is acquired, with the provision to also later deposit any revisions or updates, as per the World Bank Procedure on Development Dataset Acquisition, Archiving and Dissemination

For implementation resources, see the Data Publication Guidelines

DIME Coding Standards

DIME Data Map

The DIME data map provides a high quality professional framework for planning data work.

• For a general discussion on the usages and benefits, see the Data Map article on the DIME Wiki.
• To download our Data Linkage Table template click here
• For an example of a Data Map with a Data Linkage Table, Master datasets and Data Flow Charts, see here

Style guides

DIME code should adopt the following guides:

• Python: PEP 8
• R: Tidyverse style guide
• Stata: DIME Analytics style guide (see the Appendix in the DIME Data Handhook)

Code review

• Department-wide peer code review rounds are held every quarter and all DIME and DIME-affiliated projects can sign up here.
• Checklists to prepare and review code for different tasks are available here.

dime-standards's Issues

[coding checklists] randomization code review

Suggested updates to data security guidelines

Following up the bootcamp, there are couple of things in https://github.com/worldbank/dime-standards/blob/master/dime-research-standards/pillar-4-data-security/data-security-resources/password-manager-guidelines.md you guys could update.

• Emergency Access is only available with LastPass Premium and LastPass Families

Also, in https://github.com/worldbank/dime-standards/blob/master/dime-research-standards/pillar-4-data-security/data-security-resources/veracrypt-guidelines.md, before point 8., you could include the recommended Encryption Options to use:

Finally, it would be nice to have some guideline on how to tailor the iefolder-like master do-files to point to the VeraCrypt volume and prompt the user to mount it any time she runs a do-file containing encrypted data.

Cheers!

---

Edit (3/13) by @kbjarkefur :
This does not happen in the browser extension, but on lastpass.com

I couldn't see the Remember Password button in LastPass, but I was asked by Google Chrome if I wanted to "save my password for this site" so perhaps you can mention this as something to always avoid by picking "Never"

Typo in the first paragraph of dime-data-security-guidelines.md

There is a typo in the second sentence of the first paragraph of the file dime-data-security-guidelines.md, which is repeated in the readme file of the folder dime-standards/dime-research-standards/pillar-4-data-security/. The text should be "To be fully compliant with the DIME Data Security Standards, a project must follow all applicable items" instead of "To be fully compliant with the DIME Data Security Standards, a project most follow all applicable items".

Link to DIME Security Guidelines in NDA is broken

@kbjarkefur , the current link (https://github.com/worldbank/dime-standards/blob/master/dime-research-standards/Research-Data-Security/DIME-Data-Security-Guidelines.md) is broken, needs to be updated

Link to "Data Linkage Table template" is broken

Possible misstatement in pillar 4 readme file

The second paragraph of the pillar 4 README says: The NDA can be accessed here but link will only work if you work at DIME. The link can actually be accessed without having to log in to GH, so it seems to work for non-DIMErs as well