Saturday, 5 December 2009

Climate change and the publication of data

The University of East Anglia’s Climate Research Unit has been in the news recently. Data that had not previously been made available ended up on Wikileaks, along with some e-mails that are easy to interpret as evidence of some kind of cover-up, although that interpretation might well be mistaken.

I congratulate Wikileaks and whoever supplied the information. Wikileaks is one of the guardians of our liberty. The site has scored many other victories over those who would hide information.

I make no comment on whether this episode should change our views on climate change or on human beings’ role in it. But there is, as other commentators have noticed, an important point here about openness. If all data and all algorithms had been published, in full, as the research went along, there would not have been this embarrassing row. And we could have had more confidence in whatever conclusions were reached, because we would have known that the sceptics would have had every opportunity to make their case. There would have been other, smaller, rows as we went along, with specific results being challenged. But that would not have been a bad thing. If challenges are justified, then they should be taken seriously at the earliest opportunity. If they are not justified, then the weight of commentators will be enough to squash them.

Some people have, before this episode, spoken of researchers’ intellectual property in the data that they have collected, or in their algorithms. Excuse me? If the planet really is at stake, then it is simply immoral to put your intellectual property rights, or your chance to get to a significant result before a competing researcher, above the effective conduct of research.

How far should we take this call for openness? It strikes me that there are three main grounds for insisting that research data, methods and results be made public immediately.

The first ground is that something very serious is at stake. We increase our chances of getting the right answers if everyone can comment on data and results. And if a public response may be required, as with global warming, only complete openness is likely to build public confidence. Politicians are now going round saying that we should not be put off taking global warming seriously by the East Anglia affair. But public confidence has already been lost. It cannot be re-built merely by the pleas of politicians.

The second ground is that we will have to bear the consequences. This relates to research that is used to formulate legislation or government policy. New policies are often the subject of public consultation. But there are also the private data that are collected and that are used in the taking of policy decisions. Except where national security is at stake, the data should all be published, well before the decision is taken. The data may not give one as much confidence in a policy as the politicians who promote that policy suggest we should have. That strikes me as all the more reason to publish the data, along with full details as to how they have been used. It would do politicians good to have to admit that they take many decisions on balances of probabilities, sometimes quite fine balances. That could even help the reputation of politicians. We could better understand that it is inevitable that some policies will go wrong. Alternatively, we could argue that when things are finely balanced, the state should stay out of the matter and not have a policy at all.

The third ground is that the taxpayer has funded the research. I am not sure that this would justify a requirement to publish as one went along. But it could very well justify a requirement to publish, and to give a general and royalty-free permission to use the research and its results, within a short time after the end of each piece of work, or every three years or so if the project was a long-term one. In some fields, there is a counter-argument that the taxpayer can get his money back by having the results patented and licensed. That argument is to be taken seriously, at least if one believes that having a system of patents is a good idea. But in other fields that argument clearly does not apply, and free, open publication is likely to be the best way to promote the further development of our knowledge.

No comments:

Post a comment