Are you promoting good practices for open research?
Created by Scriberia for The Turing Way community (CC-BY)
This post is part of the CW20 speed blog posts series.
Many researchers are interested in learning about good practices for open research but do not know where to start and what to promote to colleagues. Staff in central roles such as Research Software Engineers (RSEs) and library research support staff aim to promote reproducible research practices and do not always reach their audience. This blog post summarises a discussion at CollabW20 on promoting good practice effectively. It provides a summary of existing resources on good practice, discusses failures in communicating or using these resources before providing a list of actions for reproducible research champions to spread the word and better practice. Our outlook provides special consideration of how the new form of working due to COVID-19 might present some opportunities to tackle existing communication issues to facilitate changing practices.
Reproducible and open science
There are many resources available that aim to equip researchers with tools, knowledge, training and confidence to follow reproducible open science good practice. Some examples include:
- The Turing Way
- Good enough practices in scientific computing
- Code is science
- Best practices code in open source badge
- R open science reproducibility guide
In addition to platforms and resources, social practices like code clubs, or positive community behaviours such as SatuRdays and Tidy Tuesday activities become more widespread. If you start a new activity or tool: puns work, so do stickers.
But how come researchers often don’t know about these? What are the possible reasons why they might not engage?
Barriers to following good practice
Industry partners are not always aware of the benefits of open publishing or reproducibility. Even tech companies like Google and Microsoft who have the technical expertise choose to not always be open with their research. Some grants already require statements on good practice but highlighting companies that are doing this well to incentivise further good practice might be useful.
There are also no clear incentives for researchers at the moment. There is a correlation between open practices and citations, but those advantages are not advertised widely.
Another barrier is the psychological barrier created by the lack of training in these practices for many researchers. They are domain experts, but not coding experts, so they may worry about people seeing that they don’t have “proper code” due to just learning on the job, and feel ashamed. Academics are pressured by this academic fallacy that they need to be brilliant at all aspects of the job, which means they are reluctant to share something they’re not brilliant at.
Promoting good practice within your research community
You don’t cease to be a researcher, even if your role may have a different title (such as Research Software Engineer). A degree of shared identity can help both parties to discuss gaps in learning or understanding more openly. Many Early Career Researchers feel that they don’t have the power to change practice but every little helps. Here are a few things that you can start with:
Look to see whether there's some central or departmental service. Ask them for help and things to promote - they may have resources you can use to spread the word across the whole university!
Join with like-minded people and start a code, open research, or reproducibility club
Some ideas of what to do: learn a technical skill together, try to replicate each other’s paper (upload the results to a platform like ReScience [e.g - Reproducibility Hackathons])
Thank the people who are doing a great job: email them with a thanks or broadcast to the world how useful that open-resource has been for you (twitter, facebook, postcard,...), also let them know how to fix problems if you encounter them and be nice if you ask for help.
Ask for help on where to improve. You've got a repository already with your code? Add a readme with a section asking for contributions and thanking them.
Promote the use of badges on GitHub/GitLab repositories to show people how much has been done, for example: Is the data and/or code also in Zenodo with a Digital Object Identifier (DOI)? Add it there! Have you got a good practices badge? Add it there too!
There's always room for improvement - don't do all at once. Start with what seems easy first, then start to add new practices to your routine. Learning git, <programming language>, <web platform>, <testing framework>, <open data policy> all at once is overwhelming. Go bit by bit and ask for help from the community.
Encourage institutions to provide funding for academics to do this work. Sometimes publishers will charge to make papers open access so put pressure to ensure that this money is made easily available.
The most successful channels may not be those you predicted...
In many institutions, central groups exist providing professional services support on good research practice. Those stretch from ethics committees that can help with some aspects of research design to research software engineers that can support good coding practice to support publishing data and outputs openly (often based in libraries). Many central services provide their support on their web pages, which might be a good starting point. Face to face training sessions are often run at times when researchers find it difficult to attend or the offered content is not relevant to them. Thus, virtual learning opportunities are useful where academics can access the support when needed and follow along in their own time. For additional queries, drop-in sessions could be hosted. It is important that those are kept consistent even if on some days, no one shows up and they feel like a waste of time. Feedback is often vital for central services to tailor provision, so the visibility and use of these channels (constructively) when possible benefits everyone.
Each collaboration or shared effort starts with a single act of communication. There are many barriers to changing research practices, including limitations on time, on resources, and on confidence. RSEs or other support staff/evangelists within organisations can provide training on open and reproducible processes and might help you with creating resources or running sessions. While CollabW20 is a great example of an opportunity to get in touch and pair up with others, any time is a good time to to exchange ideas and promote best practices.
The current lockdown comes with a set of challenges for many of us, but it may also provide a pause in the normal flow of research. An opportunity to consider new channels of communication, or to reimagine existing ones. It may be this time of change reaffirms shared goals, leading to new digital learning resources and new channels for support.