Michael Nielson Keynote on Open Science

He rightly points out that he is probably going to be talking to the converted, so his talk is aimed at looking for resources that can help us to find answers about how to make open science works. He starts talking about an example of failure in open science. his example is an open notebook science from Tobias J Osbourne. He built up a readership of about 100 readers on a highly technical field, but he was not getting much participation, and very little feedback. He was putting in a lot of effort but was not getting much value out of the exercise. His reason, we are in a local optimum and he felt we needed a global change.

In other areas of the world we have had examples of where making a collective change has been implemented. Sweden changed the side of the road that people drive on on one day in the 1950s.

The concept, is that there is some action where if everyone changed it would be better for everyone, but you need everyone to change at the same time. There are incentives for people not to participate because there is some cost involved in changing for the individual but if the individual does not change, they get the benefit anyway from everyone else changing. This is the same kind of problem that we have with the move to open data.

This is known as the problem of Collective Action, and the definitive work on this was published in the 1960s and is The Logic of Collective ACtion by Mancur Olsen.

The problem of open science is also an example of the difficulty of providing a public good (or for an economist a club good).

A case study is regulation in the airline industry, another example is the creation of trade unions. It is difficult to exclude non-union workers from gaining from the goods brought about by a union (for example introduction of better safety measures). This means that members have an incentive to avoid paying dues. How did unions form in the first place? Historically they didn’t form in one big go, but rather small groups tended to self organise within an organisation and started bargaining with management, and the second stage was agglomeration between these groups, and between groups across companies.

Incentives changed at different scales.

Another example is facebook, there is another public good problem here, the information that you share on facebook has many of the characteristics of a public good. The way that a successful social network grows is very similar to the way that a trade union grows. You see the same pattern in successful open science projects, the ArXiV displays the same characteristics. The ArXiV started in two very small fields, and slowly expanded into other fields.

Narrowness is a feature and not a bug when you are getting started.

Elinor Ostrom (nobel prize) in her work Governing the Commons really extended Olson’s work. She is interested in looking at how you manage the commons. She noticed that in most cases the Tragedy of the Commons tends not to happen, but rather communities tend to self-govern. She asked what principles were at work in these cases?

She has a list of rules that seem to emerge. Nielson highlights three of these, and provides an example from farmland and water usage outside in small farms in the vicinity of Valencia where one expects that water usage could be a significant problem.

They have been developing a system that deals with this commons for about 1000 years, from the middle of the 15th century. It is organised into 7 syndics each one originally related to a canal or water resource. Each of these has an elected head who is responsible for monitoring problems. These people meet every Tuesday and chat, and talk about problems that have arisen. These problems are resolved by vote, and where appropriate sanctions are applied.

There is an example that is similar in the ArXiV. There is a page telling people explicitly how you should cite ArXiV documents. (every open science project needs a page like this). This is a prerequisite for monitoring and sanctions.

Nielson gives a good example on how ArXiV citations are treated in physics, and he mentions that there are different expectations in different fields about how to use these citations, some fields strongly expect them, some are indifferent, some journals tend to be hostile.

In summary open science is a collective action problem. A lot is known on how to solve these problems in other contexts, e.g. initially focus on small groups, have a collective agreement in place, have sanctions in place for breaking these collective agreements.