YES.
We take NHS data. We use clinical, epidemiology, stats and engineering skills. We turn it into DIGITAL TOOLS rather than boring papers. Massively higher impact. Same datasets. Same skills. No recognition in REF. No funding streams.
Academia is trapped in the 19th Century.
We take NHS data. We use clinical, epidemiology, stats and engineering skills. We turn it into DIGITAL TOOLS rather than boring papers. Massively higher impact. Same datasets. Same skills. No recognition in REF. No funding streams.
Academia is trapped in the 19th Century.
... A couple of people have suggested that academics producing software and tools for coalface staff ARE rewarded in REF because you could submit them as REF "Impact Case Studies". Let me explain why that is not in any sense parity of esteem, or practical...
.. Every academic is required to submit lots of papers to be recognised as excellent in REF (fine by me btw, I've got plenty).
But "Impact Case Studies"? Each organisation submits 10 for the first 160 full time staff, then one for every extra 50 staff..
ref.ac.uk/media/1034/ref…
But "Impact Case Studies"? Each organisation submits 10 for the first 160 full time staff, then one for every extra 50 staff..
ref.ac.uk/media/1034/ref…
So. One impact case study per additional fifty staff. 3 papers per researcher (or whatever it is this time). That's 150 times more papers than impact case studies. REF is not a system that recognises and rewards software/tools. It's built to reward an outdated model of research..
.. a sensible model for REF that truly rewarded impact, software, and tools would reward this real world output at the same unit-size as academic papers. If you need 100 papers for every one impact case study, then impact case studies can only be huge (mostly expensive) monoliths
...Now part of the problem here is workforce capacity. We don't have enough academics in senior roles who can comprehend what building a tool means, or entails. This, combined with latent snobbery and very-understandable maybe-admirable barricade-guarding makes progress hard..
..For example. This person is concerned that taking open source components "off the shelf" and integrating / deploying them to meet a user-need isn't regarded as new, progressive work. Let's think what it entails tho. You need domain knowledge of the...
... data you're working with. You need technical skills to knit tools together. You need an idea for something that's novel, deliverable, and worth the candle. It's all the same as making a paper. Then let's consider the process of making a paper. Let me let you into a secret...
.. If you go to @LSHTM to do the MSc in Epidemiology (it's amazing, sign up immediately) you do a course called ASME. This is basically the recipe book for cohort studies. Once you have that under your belt, every time you do a cohort study, you're not reinventing the wheel...
@LSHTM ... you're reaching for the cohort study recipe, off the shelf. You're using a regression library in stata, or R, or whatever, written by someone else. You're applying it to a new dataset, and a new question. But there is total parity with building a data-driven tool here..
@LSHTM ... "have an idea that's novel/possible/useful, understand domain and data, take tools, apply them to data, generate output". But if the output is a paper, take a bow in REF, CV. If the output is a tool that can be used at the coalface? No REF. No funding stream. No incentives..
@LSHTM .. This is very simple. It's the 21st century. Epidemiologists, researchers who understand data, can use software to turn data into real world tools, actionable insights. But we incentivise them to make only papers. So they do. So we fail patients, doctors, the NHS, and UK plc.
@LSHTM Now, yes, to be clear, by a strict reading of the REF rules, you can put something other than papers into REF as a research output. But this is high stakes stuff, universities fighting for recognition in a conservative system. REF is about papers. But I'm keen to hear:
@LSHTM So, medical academics, doing their REF dummy runs: has anyone every asked you for this other non-paper stuff to submit as a research output? For anything other than journal papers and REF impact case studies? ..
@LSHTM ...I would also LOVE to see an audit or summary of all the software outputs submitted for medicine in 2014... I find it very hard to believe that any individual or institution would put software outputs in, if they had papers in high impact journals as a safer bet.
Right. END.
Right. END.
Just for the record, I asked for any examples of medical academics submitting software rather than paper as a "research output" (as the rules theoretically permit). >100 RTs for the thread, thousands of reads, lots of discussion, but nobody has said "yes" or given an example yet.
