Challenges in Assessing Contributions to Reproducible Research and Open Science

Challenges in Assessing Contributions to Reproducible Research and Open Science

This talk was given in a panel organised by Declaration of Research Assessment (DORA): Addressing Roadblocks in Research Assessment Reform. Slides are available on Zenodo and can be cited as “Sharan, Malvika. (2020, September). Challenges in Assessing Contributions to Reproducible Research and Open Science. Zenodo. http://doi.org/10.5281/zenodo.4013300”

Reproducibility is when the same steps are applied to the same data set, they always produce the same result. Reproducibility along with transparent reporting provides a good indicator of quality research.

This might be a simple definition, yet it offers us a promising way to assess the quality of a research work that requires multiple and complex decision-making processes. Starting from the selection of a project idea, planning, data collection, processing, analysis, publishing, archiving and ultimately by ensuring that others can build upon our work, it promotes collaboration over competition.

Nonetheless, conducting research with reproducible workflow can be daunting, both resource-wise and skill-wise. Moreover, it is seldom mandatory for publication; the ultimate reward researchers aim to obtain. Hence, most efforts in this area are often led by well-intentioned, self-motivated and already overworked researchers. Here are some visible work and tangible outcome researchers may achieve in conducting and sharing their work openly:

  • People designated to lead the work receive all the credit for the project’s success.
  • Numbers of publications and successful grants (mind you, there are way more grant applications that are rejected).
  • Growth in community size, for example, if its a software project you can think of the users and collaborators as a community.
  • Open source code and documents that are available for reuse under a license.
  • Digital Object Identifier that most of these research outputs can be attached to and the citations can be counted.

There is obviously more to research work that goes unacknowledged and unrecognised through the traditional system. In this post, I want to draw your attention to those hidden and invisible work in open and reproducible research. A few examples of hidden labour in open source communities are the following:

  • Shared leadership efforts and pro-active volunteer (unpaid) work.
  • Public engagement by communicating science to a larger audience through citizen and participatory science.
  • Supporting role, for example, people in core staff positions.
  • Accessibility and inclusion efforts. Folks who are most affected by a lack of support in this direction end up carrying the burden of work done to change the culture.
  • Archiving and maintenance roles. Librarians and members who continue to take care of our research objects even after the research is over.
  • Teaching and mentoring efforts that are completely led by usually poorly supported individuals - beyond their work.

Here are some highlights of the challenges that exist in recognising the contributions of both visible and invisible labour in reproducible and open research:

  • Most of these contributions are not fairly incentivized.
  • They require specialized skills, that are often not formally taught.
  • They need resources which are not equitably available to everyone, especially in low-income organizations, communities, and countries.
  • Specifically talking about invisible contributions, they are very hard to assess, and many volunteer-led efforts are happening there.

Assessment metrics

There are several metrics of assessment that are put in place but are not yet widely adopted. A few incredible initiatives are DORA, UK’s Research Excellence Framework, the newly established HiddenREF that aims to recognize hidden labour, CRediT Taxonomy, JOSS for publishing software and CHAOSS metrics that also takes into non-conventional metrics such as diversity and inclusion into account.

When adopting a metric system for research assessment, we should involve all stakeholders of research in identifying what kinds of recognition are meaningful for them and where in the chosen system necessary changes can be introduced to ensure that it will fairly incentivise everyone’s work and make them feel supported while promoting healthy research culture. For every project, we need to define what ‘incentives’ are valued, who are they designed for, who are involved in designing them and in what ways this will make their experience in the academic system rewarding. Answers to these questions will vary from project to project and hence, we should understand, accept, promote and celebrate diverse ways different contributions are recognised.

Examples from my work

To explore this, I would like to give examples of two independent Open Source projects that I work on: The Turing Way and the Open Life Science.

The Turing Way: Online community-led book project on Data Science.

Let me start by listing the visible metric of success (quantitative assessment) that we have in this project: within 2 years, we have 30 chapters and 134 subchapters co-authored and supported by over 250 contributors and read and share worldwide by 1000s of people.

Then we also have a qualitative assessment process in place for acknowledgement of different types of contributions, which I have listed under five aspects:

  1. Eligibility: Who is eligible to participate? Anyone interested in contributing to the project is welcome. I believe each of us has some skills that are useful for someone else, and The Turing Way aims to be a vehicle for that.
  2. System of acknowledgement: To dismantle the traditional evaluation metric system, we promote a culture of collaboration. Opposite to individual ownership and competitive nature of academia, we promote shared ownership of the project outcome.
  3. Recognition: Often in open communities, volunteers come and invest their time, skills, and energy in developing the project.
    • Most open-source projects will deem the “opportunity for people to work” as a value or reward which is not tangible at all. Specifically when these contributors can not contribute anymore. Therefore we intentionally create an opportunity for people to record their contribution in a manner that is useful and valuable for them, which is highlighted in the book and is available for people to share in their CV, for example.
    • Everyone who develops, not only writes, but also reviews, maintains, and mentors are named as an author of the book.
  4. Quality control: We have a peer-review system for collaboration and improvement of the contributions made by our contributors.
    • We also support individuals by providing support, creating helpful guidelines that can streamline their work, and multiple opportunities for community-led efforts, collaboration, and engagements.
  5. Community: where we are deeply committed to providing a welcoming and safe environment and involving people from diverse backgrounds and identities.

Open Life Science: 16-week long training and mentoring program.

Qualitatively, we can look at statistics from running two cohorts successfully. For example, we received 50% more proposals for the second round than the first one, we managed to receive funding to run two cohorts in 2021 and our post-cohort surveys have been overwhelmingly positive. We are grateful for the support from an incredible community of open science leaders who are willing to support the next generation of open practitioners.

Using the same 5 aspects for quantitative assessment in this project, we can state the following:

  1. Eligibility: We involve folks who have a research idea and they want to lead it in their communities.
  2. System of acknowledgement: We want to democratize leadership skills for a free or low cost so that it is widely accessible for everyone and not just those who have resources and mean to acquire those skills.
  3. Recognition: We offer a global platform for our members, both for the project leads or mentees to create and highlight their work, and for mentors to get recognition for their inspiring work in the open science communities.
    • Of course, visibility without fair compensation is not enough, therefore we are in the process of establishing an honorarium system for all our members.
  4. Quality control: The quality assessment depends on the commitment and not by the extent someone has developed their projects. There is also peer-networking, mentor guidance, check-ins and certification at the end of each cohort.
  5. Community: we aim to connect a community of community builders who are empowered to design inclusive, not only open, research projects.

Here I provided two independent systems that have evolved to recognise people for different roles that have emerged separately in these projects. I want to make it clear that we can not and should not standardize the way we recognise labour (visible or invisible) across different research domains and open source communities. There is no silver bullet or a system that can equitably recognise diverse roles and contributions that emerge independently and sporadically in all projects.

Closing remark

“Time for a Culture Shift”, The Turing Way Community, & Scriberia. (2019, July 11). Illustrations from the Turing Way book dashes. Zenodo. http://doi.org/10.5281/zenodo.3332808

We have been hearing this for a long time that it’s time for a culture change. The culture change can not truly happen if we fail to recognise less traditional research contributions. By thoughtfully and collaboratively setting-up multiple pathways for recognising all kinds of work in our projects, we will ensure that we do not promote a system that disproportionately undermines important work carried out by different members, especially those who have been previously and continuously left behind. By incentivising collaboration and welcoming the diversity of ideas, we can move away from the competitive and individualistic nature of academic evaluation.

With that, I would like to acknowledge my colleagues and community members from The Turing Way, especially Kirstie Whitaker and Open Life Science, particularly Yo Yehudi, Berenice Batut and our volunteer members. All our resources are available under cc-by 4.0, and you are welcome to join and help us improve how we recognize visible and invisible labour in science.

Cover image by @sam_truong on Unsplash.

comments powered by Disqus