ReproducibiliTea 2020-10-01

# ReproducibiliTea 2020-10-01  [Discussed paper](https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.2006022) [Summary Slides](https://t.co/xaDeNRyqcB) During the ReproducibiliTea Meeting, a group of 8 Early Career Researchers (ECRs) came up with actionable ideas on how to introduce skills training to all students (Bachelor, Master and PhD). This collaborative document summarizes the ideas for implementation and the discussion. ## Discussion Questions  * Should courses on open science practices be field specific? * Different challenges from different kinds of data * Different relevance of computer code * Different prior knowledge * Different research output that has to be made open e.g. * Questionnaires * Experimetal Setup * Description of field * ... * Learning from each other and accross field is an important skill, too -- Think about team science approaches and interdisciplinary research; this includes realizing that different fields use different sets of tools * Does a reproducible workflow make your own life easier? * Some aspects of reproducibility or version control might * Adding comments/meta data takes significant time without (immediate) pay-off for yourself. * Some aspects pay-off in the (very) long run, and might pay-off only/more when you stay in academia. ## Ideas for Implementation ### Research portfolio Open science practices result in the dissemination of a lot of citable research items. This research output demonstates a multitude of **transferrable skills**. Collecting them in a portfolio broadcasts those skills/employability for both academia and industry. - Slides - Data - Posters - Videos - (Analysis) Code - Course Materials To support PhD students in curating such a portfolio it would be helpful to have a **portfolio coach/mentor** (potentially with HR or field related corporate experience) that assists in identifying and optimizing porfolio items. Such **portfolio checks** could be offered by carrer services (comparable to CV/linked in checks). ### Improve something that is already here Instead of play projects or textbook assignments skills could be demostrated by improving existing material. - Code - Dataset - Onboarding document - Wiki/Tutorial ### Reproducibility Checks Reproducibility checks or participating in a ReproHack can be used to gain first hand experience on the difference open science practices can make. Reproducing the analyses of your colleagues can raise awareness for issues addressed by open science practices such as: - Undocumented or unstructured code - Intransparent data cleaning - Missing labels or meta-data - A missing README file - -... If performed for public/third party research Code check reports/ verification reports can also be a research output with a doi in and by itself. ### Skill related Internships Current internship guidelines often demand that students carry out an entire research project answering a research question. This entails that they have to go through the entire research process in a very short amount of time without enough attention to the details. Skill related internships would allow a focus on a certain part of the research process with a higher chance to master the relevant skills instead of rushing them. - **Data internship** - Data stewardship - Data cleaning - FAIR principles - Data Archives/Repositories - **Code internship** - Write the analysis code for existing data - Style guide adherence - Coding practices - Version Control - Code Review - **Design Internship** - Research Questions - Research Design - Sampling - Recruitment - Preregistration - **Reporting Internship** - Academic writing - Reference Management - Style guides - Visual communication of results - Slides/Posters ### Replacing the existing with open science Small changes to assignments and grading structure allow to introduce working with open science tools to existing courses. Without having to change course names and getting approval for such major changes. For example: - Don't ask for a research proposal <-> but ask for a preregistration - Learn Peer-Reviewing (code, other peoples' thesis proposal, even across years, i.e., the first years peer reviewing the second year's proposals) ### PhD Introduction day/On-Boarding process During the Introduction Day (university wide and the faculty events), new PhD students are introduced to the university. This is a good moment to make a statement, set the tone and make students aware that starting-off with an Open Science workflow will make for better and robust science but also allows them to collect items for their research portfolio. - Introduce training program - Introduce the idea of a research portfolio (and how it is beneficial within and outside academia) - Introduce the general notion of robust, transparent and responsible scholarship as a core value