Perverse incentives: how the reward structures of academia are getting in the way of scholarly communication and good science
by Sir Timothy Gowers
Abstract:
The internet has been widely used for the last 20 years and has revolutionized many aspects of our lives. It has been particularly useful for academics, allowing them to interact and exchange ideas far more rapidly and conveniently than they could in the past. However, much of the way that science proceeds has been affected far less by this development than one might have expected, and the basic method of communication of ideas — the journal article — is not much different from how it was in the seventeenth century.
It is easy to imagine new and better methods of dissemination, so what is stopping them from changing the way scientists communicate? Why has the journal system proved to be far more robust than, say, the music industry, in the face of the new methods of sharing information?
The dream that all information is available on-tap, accessible through a few clicks. We’ve got a bit of that via Google, Wikipedia, YouTube, map sites, travel sites, news sites. But a lot of content is for subscribers only: you can find content in Google Books – but only a page before cut off due to copyright.
Of course copyright holders need incentives to create, to cover costs, etc. Academics aren’t directly paid though – actually the barriers to content are the bigger problem. Covering costs? maybe “if you insist on antiquated methods of publication”.
What could we share? The not-yet-complete idea, that others can build on. OTOH if everyone shared everything they thought it could end up a complete mess. But we can make order from chaos.
In maths the revolution has started: his library transitioning from painfully closed stacks to open stacks coincided unfortunately with not even needing to go to the library for content anyway. Wikipedia for basic concepts; arXiv.org for preprints (don’t need journals at all in maths); OEIS (database of sequences of whole numbers along with formulae for generating them); MathOverflow – for questions at the research level – usually get a useful answer within a few hours.
Traditional way of doing things is the “lone genius” model. But thought it’d be interesting to solve a problem in the public. So posted some initial thoughts on his blog and invited contribution. Traditionally there’s a fear of getting scooped – but doing it completely in the open, timestamps mean no-one can take credit; in fact it rewards putting your comment up quickly before someone else can. Problem was solved in just 6 weeks.
Perverse incentives in maths:
- personal ambition
- reward for being first (not for being inspiration, or for being second but with a better solution)
- primacy of journal article while expository and enabling activities are downplayed – when you start writing textbooks instead of journal articles this is seen as your career slowing down
- little recognition for incomplete ideas
These are obstacles to efficiency.
Paradox of paywalls: mathematicians write, peer review, edit; dissemination costs almost nothing; almost all interesting recent content is on arXiv (which can include final accepted manuscript – and anyway it’s not much different from the preprint); and still libraries pay huge subscription fees. The problem is the internet came along very quickly while we’re still doing things the old way.
Some initiatives:
- his blog post about personal Elsevier boycott which inspired someone to set up a pledge which thousands signed
- Open Library of Humanities set up when Journal Lingua left Elsevier and became Glossa
- Discrete Analysis (arXiv overlay journal) set up as proof of concept for cheap journal publication, with US$10 submission charge – and a nice user interface
- No.Big.Deal – trying without success to get Cambridge and JISC to bargain better
- Freedom of Information Act requests to UK universities for how much unis are paying Elsevier (contra confidentiality clauses)
Perverse incentives are held up by the whole network of publishers, editors, writers, readers (subdivides into people actually reading it, and people scanning it to judge the writer eg hiring committees), librarians (who have the power to cancel – but subject to academic criticism), scholarly societies (who often derive income from publishing journals), consortium negotiators, funders (in a good position to create mandates) – creating a situation where it’s very difficult to change things.
Feels like has had little effect, but it’s important to have lots of little initiatives which together build to pull the wall down.