# The WORCS workflow, version 0.1.6

## WORCS: Steps to follow for a project

This vignette describes the Workflow for Open Reproducible Code in Science, as introduced in Van Lissa et al. (2020). The paper describes the rationale and principled approach on which the workflow is based; this vignette describes the practical steps for R-users in greater detail. Note that, although the steps are numbered for reference purposes, we acknowledge that the process of conducting research is not always linear. The workflow is illustrated in the graph below, with optional steps displayed in blue nodes:

### Phase 1: Study design

1. Create a new remote repository on a ‘Git’ hosting service, such as “GitHub”
• For inexperienced users, we recommend making this repository “Private”, which means only you and selected co-authors can access it. You can set it to “Public” later - for example, when the paper goes to print - and the entire history of the Repository will be public record. We recommend making the repository “Public” from the start only if you are an experienced user and know what you are doing.
• Copy the repository’s SSH link to clipboard; this link should look something like git@github.com:username/repository.git
2. In Rstudio, click File > New Project > New directory > WORCS Project Template
1. Paste the remote Repository address in the textbox. This address should look like git@github.com:username/repository.git
2. Keep the checkbox for renv checked if you want to use dependency management (recommended)
3. Select a preregistration template, or add a preregistration later using add_preregistration()
4. Select a manuscript template, or add a manuscript later using add_manuscript()
5. Select a license for your project (we recommend a CC-BY license, which allows free use of the licensed material as long as the creator is credited)
3. A template README.md file will be automatically generated during project creation. Edit this template to explain how users should interact with the project. Based on your selections in the New Project dialog, a LICENSE will also be added to the project, to explain users’ rights and limit your liability. We recommend a CC-BY license, which allows free use of the licensed material as long as the creator is credited.
4. Optional: Preregister your analysis by committing a plain-text preregistration and tag this commit with the label “preregistration”:
• Document study plans in a preregistration.Rmd file, and optionally, planned analyses in a .R file.
• In the top-right panel of ‘RStudio’, select the ‘Git’ tab
• Select the checkbox next to the preregistration file(s)
• Click the Commit button.
• In the pop-up window, write an informative “Commit message”, e.g., “Preregistration”
• Click the Commit button below the message dialog
• Click the green arrow labeled “Push” to send your commit to the ‘Git’ remote repository
• Tag this commit as a release on the remote repository, using the label “preregistration”. A tagged release helps others retrieve this commit.
• Instructions for ‘GitHub’ are explained here
5. Optional: Render the preregistration to PDF, and upload it as an attachment to a dedicated preregistration server like AsPredicted.org or OSF.io
• In ‘RStudio’, with the file ‘preregistration.Rmd’ open, click the “Knit” button above the top left panel
• When the PDF is generated, go to one of the recognized preregistration services’ websites, create a new preregistration, and upload it as an attachment.
• Optional: Generate a DOI for the preregistration through the OSF or a service like Zenodo
6. Optional: Add study materials to the repository.
• Only do this for study materials to which you own the rights, or when the materials’ license allows it
• You can solicit feedback and outside contributions on a ‘Git’ remote repository by opening an “Issue” or by accepting “Pull requests”

### Phase 2: Writing and analysis

1. Create an executable script documenting the code required to load the raw data into a tabular format, and de-identify human subjects if applicable
• Document this preprocessing (“data wrangling”) procedure in the prepare_data.R file.
• This file is intended to document steps that can not or should not be replicated by end users, unless they have access to the raw data file.
• These are steps you would run only once, the first time you load data into R.
• Make this file as short as possible; only include steps that are absolutely necessary
2. Save the data using open_data() or closed_data()
• WARNING: Once you commit a data file to the ‘Git’ repository, its record will be retained forever (unless the entire repository is deleted). Assume that pushing data to a ‘Git’ remote repository cannot be undone. Follow the mantra: “Never commit something you do not intend to share”.
• When using external data sources (e.g., obtained using an API), it is recommended to store a local copy, to make the project portable and to ensure that end users have access to the same version of the data you used.
3. Write the manuscript in Manuscript.Rmd
• Use code chunks to perform the analyses. The first code chunk should call load_data()
• Finish each sentence with one carriage return (enter); separate paragraphs with a double carriage return.
4. Regularly Commit your progress to the Git repository; ideally, after completing each small and clearly defined task.
• In the top-right panel of ‘RStudio’, select the ‘Git’ tab
• Select the checkboxes next to all files whose changes you wish to Commit
• Click the Commit button.
• In the pop-up window, write an informative “Commit message”.
• Click the Commit button below the message dialog
• Click the green arrow labeled “Push” to send your commit to the remote repository
5. While writing, cite essential references with one at-symbol, [@essentialref2020], and non-essential references with a double at-symbol, [@@nonessential2020].

### Phase 3: Submission and publication

1. Use dependency management to make the computational environment fully reproducible. When using renv, you can save the state of the project library (all packages used) by calling renv::snapshot(). This updates the lockfile, renv.lock.
2. Optional: Add a WORCS-badge to your project’s README file and complete the optional elements of the WORCS checklist to qualify for a “Perfect” rating. Run the check_worcs() function to see whether your project adheres to the WORCS checklist (see worcs::checklist)
• Reference the WORCS checklist and your paper’s score in the paper.
• Optional: Complete the additional optional items in the WORCS checklist by hand, and get a “Perfect” rating.
3. Make the ‘Git’ remote repository “Public” if it was set to “Private”
• Instructions for ‘GitHub’:
• Go to your project’s repository
• Click the “Settings” button
• Scroll to the bottom of the page; click “Make public”, and follow the on-screen instructions
4. Create a project on the Open Science Framework (OSF) and connect it to the ‘Git’ remote repository.
• On the OSF project page, you can select a License for the project. This helps clearly communicate the terms of reusability of your project. Make sure to use the same License you selected during project creation in Step 3.
5. Generate a Digital Object Identifier (DOI) for the OSF project
6. Add an open science statement to the Abstract or Author notes, which links to the ‘OSF’ project page and/or the ‘Git’ remote repository.
• Placing this statement in the Abstract or Author note means that readers can find your project even if the paper is published behind a paywall.
• The open science statement should indicate which resources are available in the online repository; data, code, materials, study design details, a pre-registration, and/or comprehensive citations. For further guidance, see Aalbersberg et al. (2018). Example:
In the spirit of open science, an online repository is available at XXX, which contains [the data/a synthetic data file], analysis code, the research materials used, details about the study design, more comprehensive citations, and a tagged release with the preregistration.
7. Knit the paper to PDF for submission
• In ‘RStudio’, with the file ‘manuscript.Rmd’ open, click the “Knit” button above the top left panel
• To retain essential citations only, change the front matter of the ‘manuscript.Rmd’ file:
Change knit: worcs::cite_all to knit: worcs::cite_essential
8. Optional: Publish preprint in a not-for-profit preprint repository such as PsyArchiv, and connect it to your existing OSF project
• Check Sherpa Romeo to be sure that your intended outlet allows the publication of preprints; many journals do, nowadays - and if they do not, it is worth considering other outlets.
9. Submit the paper, and tag the commit of the submitted paper as a release, as in Step 4.

### Notes for cautious researchers

Some researchers might want to share their work only once the paper is accepted for publication. In this case, we recommend creating a “Private” repository in Step 1, and completing Steps 13-18 upon acceptance by the journal.

For a list of sample worcs projects created by the authors and other users, see the README.md file on the WORCS GitHub page. This list is regularly updated.