How to fix academic research and publishing
I recently saw the following article about academic publishing being broken.
https://www.stuff.co.nz/science/300325106/academic-publishing-is-a-broken-mess
Something that has resonated with me was Naval Ravikant talking about the principal agent problem https://nav.al/principal-agent.
A principal is an owner; an agent is an employee
If we can align our incentives then the minor details will work themselves out. Slightly related to this is Goodharts law which is summarized by
When a measure becomes a target, it ceases to be a good measure
https://en.wikipedia.org/wiki/Goodhart%27s_law
In this article we will look at the incentives of each party and how that results in a broken system, then what to do in order to fix it.
Incentives and metrics
At the moment here are the main drivers for the parties are
- University — Funding and prestige
- Researcher s— Number of published papers
- Publishing platforms — Profit, trust and reach
The result of misaligned incentives
Seth Godin has a TED talk on the 7 kinds of broken. If you take this one level higher you can see that all of this is the result of misaligned incentives.
In my opinion the funding and incentive model is the root cause of all of the following
- Why bad science spreads
- Scientific fraud in America
- Flaws in peer review
- Publish or perish an example of Goodhart’s law
- Gaming the system
- The evolution of bad science
- Taking academic corruption to the next level
Objectives and key results
So if we were to go back to objectives what is it that the owner / principal is after?
- Benefit to society — The purpose of research is for the betterment of humanity
- Trust and quality — Provide high quality trustworthy research
- Reach — Provide a large reach
- Cost — Make the cost of using the system low
- Ease of use — Make using the system frictionless
How to fix this?
Going back to the Principal Agent problem, if we make the Principal (the owner) and the Agent the same stakeholders then this should reduce the symptoms.
Easy to use, self correcting incentive, publishing platform and review model that provides a positive feedback cycle that selects against fraud.
Distributed Autonomous Organizations that have an incentive model that aligns with its stakeholders.
A transparent product based funding model and virtual economy based on crypto tokens ‘research tokens’ and tokenomics (token economics).
There are already initiatives set out to solve this issue such as
- https://plos.org/
- https://www.cos.io/
- https://sci-hub.se/
- https://1729.com/crypto-sci-hub
- https://wikipedia.org/wiki/Preregistration_(science)
and this will be an interesting space to watch in the near future.