The hacking of emails written by scientists at the Climate Research Unit has produced lots of comment and copy about the efficacy of climate science and seems to have done damage to the reputations of the scientists and sciences involved. However the response to this incident – an opening up and peering of science that should be the response to this incident, will benefit us all.
I’ve always felt that science is in essence, quite a democratic idea – this may seem the opposite of how some perceive scientists to be; cold, logical and aloof people getting their perspective from laboratories far from the real world. When I did science at school (and to be fair I was not that great at it) the concept that you could do the experiments to find things out for yourself rather than simply accepting what was written in the text books was, well…punk. The scientific method has at its core some very democratic principles: First is that when something is correct – it is so because that’s what the evidence says and this is regardless of the status of the person advocating and/or the view of society to the point being made. When Copernicus suggested that the earth rotated around the sun, it made those in power very unhappy – but the democracy of the evidence pushes the point regardless of the discomfort of the Catholic Church. Second is that you have to back your claims with both evidence and method – this in turn drives transparency and peer-involvement. It is the essence of peer-review; that others need to be able to see behind the curtain of your claims. This is also a democratic principle because when making claims in science you can’t rely on your position in society, your wealth or your power; you have to lay bare the foundations of your ideas for inspection and in doing to open them to challenge.
Now these are all lofty claims – the reality is often very different. The scientific method has noble ideas and claims, yet where it meets human nature, it often gets a little lost. Scientists are very human; they can, and do, get things wrong. They can, and do, make mistakes – and then sometimes try to cover-up those mistakes. Politics and not data can corrupt findings either by influencing the process or burying unfavourable outcomes. This is both a limitation and the main point of these democratic ideals. A limitation because it imposes by human flaws into how the system operates. The main point, because the requirement to share method and data – to open it up to your peers – is one of the ways we seek to iron out these flaws.
There is another process within science that is also important to understand; it’s interpretation. When I was taught science in school we were told that the idea behind the layout of an experimental write-up (abstract, introduction, method, results, conclusion etc) was structured as such so that anyone with a good grasp of language and maths could read it and understand it. Again, this is another noble and democratic ideal that often flounders in reality. Most areas of science are so complex that you need a better grounding in the specific area than the paper’s introduction may offer in order to set the correct context. The maths used in research has developed significantly since Copernicus and are the methods available require much more than basic maths to understand. As such most people outside the field rely on an interpretor to locate important findings, translate them into normal speech and summarise them in an accessible way – this function is most often done by the media. But just as scientists are very human; they can, and do, get things wrong – so are journalists. They can, and do, make mistakes – and then sometimes try to cover-up those mistakes. The politics of a media outlet can and does corrupt findings either by influencing the selection of reports and interviewees or by twisting outcomes to that media outlet’s narrative into favourable shapes.
This is where the changes in communication technology come into the story. First, science now has the opportunity to add many more peers into the process; it is much easier than ever before to collaborate, exchange data and ideas and discuss methods. We can all become peers; reviewing papers, performing experiments. We can control for the problems in the complexity of the process by aggregating resources and allowing people to focus on the aspects that they are most equipped to do. An example of this is the [email protected] project – where people can aggregate their efforts in the search for extraterrestrial intelligence. There are various levels of involvement to the project from the most basic entry level of installing the software (that uses downtime of people’s computers to analyse data) to assisting with the development of the software itself (called BOINC). This is not a perfect example of a fully democratic project for a number of reasons, for example the code for BOINC is not open source and getting direct involvement with the small handful of people coordinating the project working for Berkeley is hard; but it is a start.
Indeed BOINC has grown beyond [email protected] and is now used for a number of science research programs where mass computational power is required from helping in the development of particle accelerators to analysing proteins.
Second is that changes in communication technology allow much more scope for scientists to communicate without the middleman of the media. We’ve seen a proliferation of science blogging – some amazing while others are poor – but all allow direct communication on a level not possible a couple of decades ago. This is key, because from my experience of chatting to scientists they are not cold, logic, aloof people getting their world-view from laboratories – but human beings who are totally engaged in the real world – and know how messy it is. It is often the media interpretation of their work that likes to cull the complexities to make the research fit an ongoing narrative rather than report it, warts and all. Climate science is a classic example of how this is happening in practice; while those unhappy with the consensus on climate change have plenty of outlets and blogs – so we see more and more and more of the scientists themselves weighing in to speak directly with the public. Realclimate is a great example of this, where not only do the climate scientists post articles about the latests research and news items, but they also answer people’s questions in the comments sections.
However there is a downside to this; it does mean that people will try to use the data to arrive at pre-determined positions that are unscientific. They will do this anyway and so this is a danger, yes. It means people will abuse the data and process in all sorts of horrendous ways to suit thier political agendas. This is also a danger. Yet the same peer-principles that allow this abuse are also the key to de-bunking it. Let me give an example – the Global Historical Climatology Network (GHCN). This is a datset of temperature records from 1880 to the present day. It is simple to download the data and then do as you will with it. One user published a paper that purported to show that global warming was not happening as was claimed and that the adjustments made were distorting the picture – while another responded to this claim with another use of the same dataset displaying the changes applied by the scientists to the raw data in a simple graph to show how little change was going on. This graph is a simple and powerful refute to the obfuscation of the facts and is all the more powerful because it was developed independently of those who generate the data. Without this data being available and without the peer analysis of the data, the claims and counter claims lack any objective – and verifiable objectivity to them.
What has become apparent to me in the aftermath of the hacking incident is that is that the broad response (but sadly not all) to the hacking has been to both open-up the data and source-code being used to drive climate science, but also to also force more scientists involved to spend more time engaging with people in regard their work. It is going to be harder and harder for scientists to keep data private by default – which is good – and I would hope we move to a position where this is both best practice and the norm.
A climate example of some of these ideas in action is Climate Prediction – a series of experiments in climate modelling that uses the BONIC system. This site is the portal for a number of experiments, where hundreds of people are using their spare computing power to study of the possible effects of a prescribed slowdown of the North Atlantic meridional to modelling the climate of the Mid-holocene. What is encouraging about these projects is that they contain lots of information about the experiments themselves and a lively bulletin board for discussion of the experiments and the science behind them plus an offer to share the results of the experiments with others – a triple engagement. (Though as the moment the sharing of the data is aimed a scientists and not lay-researchers).
Hat-tip to Ryan and Samuel for extra information.