Octopus set to make sea change in research culture
I can’t believe it’s been five years since I first had the idea for Octopus. Finally, it’s come to fruition, thanks to funding from Research England and a partnership with Jisc, and it launches on 29 June 2022.
Back in spring 2017, I had recently started a new job in academia in Cambridge after 16 years working on BBC science programmes. As an ‘outsider’ to academia, I was struck by the similarities: researchers, just like media professionals, were looking for neatly packaged stories. This, I thought, was a huge problem.
Journalists and programme-makers expect to be judged on the popularity of their outputs. In research, though, where the aim is to carry out careful work (regardless of the eventual findings), a powerful story and popularity with readers are not appropriate metrics of success.
If researchers are being judged almost entirely on their abilities to ‘land a message’ and write popular journal articles and not on how well they were doing the underlying work, then ‘good stories’ is what we’re going to get – not necessarily good research practice.
The idea
This is where the idea behind Octopus came from. I started a document back then called 'A different way to organise scientific research’.
At the start, I listed concerns at what I was witnessing in scientific research groups, then moved on to my tentative solution: splitting up the concept of a scientific paper into eight linked publication types, allowing each to be judged on its own intrinsic merits and not on any overall constructed narrative.
After setting out the idea, I speculated about how this concept could affect not just the way people published research, but the way they approached the whole process of doing research.
A new approach
Each part of the research process could be carried out by different people, possibly at different times, encouraging professional specialisation. For example, analysts or data collectors could publish alone and get credit for what they do.
That could change the way researchers approach their careers.
It could also change the way research funding is allocated. I suggested in that first document that potential funders could browse the research trees and tag links in the chain that they would be willing to accept funding bids for.
They could use the open reviews of hypotheses and methods rather than the current secretive system of grant applications and grant peer reviews.
I listed a set of advantages that I saw for the new system:
- It should encourage more collaborative work, worldwide, regardless of geographical and political barriers.
- It should be much easier and clearer to allow replication and combination of data.
- Potential automatic translation portals/options, making research more accessible worldwide.
I also listed challenges, such as:
- Would ‘flag as inappropriate’ comments need a moderator?
- Does this model work for ‘all’ scientific publications?
- Could the funding be ‘gamed’ by people ‘liking’ their own or their friends’ protocols to gain funding options?
- How do we ‘seed’ the system?
These are all still active questions.
Hope for the future
We are still hoping that transparency will lead to self-regulation of the system, that the model can work for a large number of research subjects, and I am still sitting up late at night creating training datasets for the algorithm that someone else is skilfully teaching to ‘seed’ the system with existing research questions.
Perhaps the biggest question, though, is: “How would we overcome the commercial and current publishing pressures that keep the ‘old system’ in place?”
Simply creating a new system that is somehow ‘better’ overall isn’t enough, even if people understand why they should prefer it. What’s important is creating a new system that works better for individual researchers day-to-day, to the extent that the benefits from using it outweigh any risks of trying something new.
Simply creating a new system that is somehow ‘better’ overall isn’t enough, even if people understand why they should prefer it.
I am hopeful for the future if the risks are low (because people can still write a traditional journal article based on the same work), and we can keep building up the benefits (making it quick and easy to use, useful for researchers in finding as well as sharing work, and – most importantly – ensuring it gives them direct benefits in terms of credit with funders and institutions).
I truly believe that Octopus has the potential to change research culture for the better, to improve research quality, and allow us to make and use research findings faster and more efficiently for the benefit of everyone.
The tiny baby Octopus will soon be swimming free in the wild, wide ocean of digital research tools, but the currents of the digital research ocean are strong; no one now knows where Octopus will end up.
I end with a plea: every one of us in the research ecosystem can shape those currents in one way or another. If you want to see change of the type that Octopus promises, then throw it some food and help it grow. If enough of us do that, then maybe that potential will indeed be realised.