… so says the caption at the start of this video made by OpenScienceSummit. That statement is not necessarily meant to be a comprehensive description of science, of course, but it does presumably reflect the priorities of those who advocate Open Science. As such, it is striking that it omits to mention the idea that science is about producing reliable, empirically tested knowledge that confers practical benefits. Possibly, that is seen as something that can be taken for granted. It is the social nature of science that still has to be argued for.
While it is certainly possible for someone working in isolation to produce empirically tested knowledge that confers practical benefits, it is also fairly obvious that sharing of ideas and observations allows for a greater diversity of hypotheses to consider and a greater range of experience to test them against. Likewise, a greater diversity of perspectives and ingenuity can only result in greater overall practical benefit being derived from any given expression of scientific knowledge. As soon as there is any sense of competitive urgency or ambition about the scope of production of empirically tested knowledge that confers practical benefits, it is advantageous to work socially.
How can Open Science encourage or optimise the benefits of this social aspect of science?
Science, as we understand it today, is largely produced by professional scientists with specific kinds of education and training, working with equipment and facilities not usually found outside the world of professionalized science. To a large degree, the principal influences on their choice of research problem and the principal audience for their reporting of results come from within that world. If we are interested in allowing those from outside the world of professionalized science to realise the greatest overall practical benefit from scientific research, we need to see to it that the choice of research problems and the reporting of results are done in ways that take into account the perspectives of people from outside that world. One question for advocates of Open Science is therefore whether Open Science helps achieve that aim.
The term Open Science has been used in connection with a variety of concerns and its overall meaning arises as a summary of those various contexts. The short Wikipedia entry on Open Science describes it as “the umbrella term of the movement to make scientific research, data and dissemination accessible to all levels of an inquiring society, amateur or professional. It encompasses practices such as publishing open research, campaigning for open access, encouraging scientists to practice open notebook science, and generally making it easier to publish and communicate scientific knowledge”.
A few readily-found examples of initiatives branding themselves as Open Science include: the OpenScience Project, “dedicated to writing and releasing free and Open Source scientific software”; the Open Science Grid, which “advances science through open distributed computing”; and the Open Science Directory, “a global search tool for all open access and special programs journal titles”. A blog post at the Open Science Project answers the question “What, exactly, is Open Science?” with: “transparency in experimental methodology, observation, and collection of data; Public availability and re-usability of scientific data; Public accessibility and transparency of scientific communication; Using web-based tools to facilitate scientific collaboration”. Open Science Summit, answers the question “What is Open Science?” with “Science in the 21st century using distributed innovation to address humanity’s greatest challenges”.
The Open Science Federation aims “to improve science communications” with the participation of “open source computer scientists and citizen scientists, science writers, journalists, and educators, and makers of and advocates for Open Data, Open Access, and Open Notebooks”. The federation’s own contribution appears to mainly involve encouraging the use of blogs and online social networking media. From examples like these, it is possible to identify several specific areas on concern to Open Science advocates. Altogether, I have identified the following specific topics each of which seems to have a significant following among Open Science advocates:
Readiness to makes one’s data available to others is fundamental to good scientific practise. It bolsters confidence in one’s conclusions and allows alternative interpretations to be developed. The Open Data principle attempts to consolidate this into express obligations, first to make all data (including so-called “negative” data from studies or experiments that seemed to lead nowhere) available and, second, to make it available on terms that allow others to reinterpret and re-use it freely. This provides the possibility of allowing as many uses and interpretations of a given dataset as possible. There are limitations, of course. Some datasets will be proprietary and there is not always a clear boundary between “data” and anecdotal accounts of observations. More seriously, the effective and efficient use of datasets requires standardization of comprehensive metadata and some standardization in formatting of datasets themselves. These requirements are well advanced in some fields of science, but much less so in others where significant investment in standardization would be required to realise the benefits of Open Data, even if willingness to make data available is well established. Establishing standards for all types of data is a significant undertaking and may even confound the openness of scientific enquiry by placing constraints on the type of data to be collected. At the same time, they assist in establishing commensurability between datasets collected in differing contexts which could help diversify scientific enquiry. Until those considerations are addressed, professionalized scientists are likely to choose problems for study in much the same way whether they have a view to making their data open or not.
For some advocates of Open Science, the term is largely synonymous with an insistence on making scientific software Open Source. In fields of science where very large datasets are generated, data analysis may rely on specialist software. Publishing the source code of such software allows the manipulations it carries out to be properly understood by everyone with an interest in knowing, helps the discovery and elimination of coding errors and could accelerate the development of new or improved software to extend and diversify possible analysis. It is really just one aspect of the long-established scientific principle of disclosing one’s methods. While the term “Open Source” need not imply any more than disclosure of the source code, it is often accompanied by an expectation of freedom to use as well. This ensures that the above benefits are realised as broadly as possible. However, the Open Source principle does not in itself do anything to open up choice of what datasets are desirable in the first place.
Allowing free access to and distribution of scientific literature helps make the results of research more widely available. Although Open Access journals or, failing that, do-it-yourself web publishing must by now be possible for just about any professionalized scientist. Nevertheless, for many (most?), the priority remains is to publish in a ‘good’ journal even if that means libraries and individuals will be charged hefty subscription fees. This priority is professional, not scientific: a ‘good’ journal will attract readers and enhance one’s CV better than any of the open options. Much is said by Open Access advocates of how this professionalism limits access to science by those who could use it from outside the circles of professionalized science. Another possible effect of this ‘professionalist channelling’ of scientific publication into ‘good’ journals that is much less discussed lies in the uniformity of perception of the value of topics to be researched. Researchers gear their research priorities to what will get published in a ‘good’ journal. This effect will be pretty much the same whether the ‘good’ journals are open access or not.
In principle, the idea of opening up the drafting of research proposals presents an excellent opportunity for “lay” people to participate in deciding the direction of scientific research. It changes the relationship between publicly-funded professional scientists and “society” from one in which the existence of a class of professionalized scientists is seen as a public good in its own right to one in which the public good stems entirely from the extent to which professional scientists brings their specialist knowledge and expertise to bear on problems selected by the public. This does represent quite a shift from the way things are generally done at present, of course. Not least, it requires a change in the way that the world of professionalized science sees itself. Even scientists who are apparently committed to “the re-use and re-purposing … of scientific knowledge through collaboration between the scientific community and the wider society” go on to represent that collaboration like this. The research world is portrayed as something separate, even remote, from “society”. Research provides society with information through education and publishing and is itself influenced by society through “policy”, volunteer work and “citizen science” (more about that below). Behind this type of representation is a tacit assumption of research (professionalized science) leading an autonomous existence, almost as though science were itself a natural phenomenon outside the bounds of human society. In contrast to that, there is the view of science as a set of customs followed by a set of people in pursuit of relating to others and making a living for themselves within society.
In his book Reinventing Discovery, Michael Nielsen features the Polymath Project, initiated by Cambridge University mathematician Tim Gowers, as an example of crowdsourcing in science. Gowers used the reach of online social networking to swiftly form a virtual community of people interested in collaborating on developing a mathematical proof. This community was informal and non-professional. Amateurs could join in on practically the same basis as professionals. Individuals contributed only as much as they wanted to. A critical contribution might come from anyone at any time. They had a solution in only a few weeks.
Presumably, at each step of the way, the size of the community was enough to ensure that someone would already be thinking along the right lines to suggest the next step. One wonders whether an entirely professional collaboration, probably of fewer people and united as much by considerations of professional or expert status as by their interest in a specific mathematical problem, might have been keener to preserve orthodoxy in their thinking and would have taken longer.
A deprofessionalized crowdsourcing approach of the type exemplified by Polymath might also be used in experimental science to choose specific hypotheses for investigation, to design experiments that properly test a chosen hypothesis, or to evaluate hypotheses against data. Potentially, then, crowdsourcing could open science up to participation by non-professionals and, by that token, to some extent, direction by non-professionals. However, one has to question how far this could progress before running into difficulties. To what extent does Polymath’s success stem from it having been instigated by an individual of Gowers’ status? His reputation meant that a lot of people were already ‘watching’ (i.e. reading his blog) when he first started it. It also gave a kind of credibility to his selection of problem to work on. The collaboration formed quickly because a lot of people were aware of it as soon as it was announced and because Gowers’ involvement gave them confidence that the project would go somewhere. Gowers’ choice of problem seems to have been made on the basis that there was academic mileage in it. In other words, his career priorities would be served by it. That wasn’t necessarily true of other participants. Indeed, for non-academic participants there was little to gain other than the amusement value of participating itself. To be sure, Gowers gave up the kind of exclusivity of authorship he might have had from a more conventional way of working, but he retained “authorship” of the choice of problem to be tackled in the first place. To what extent can we expect a similar crowdsourcing approach to work for just anyone who has a problem they feel unable to solve themselves? Further difficulties become visible when one considers what might happen when the proposed problem is being solved in pursuit of some further practical purpose. How might a crowdsourcing collaboration go if working on a problem connected with potential responses to global climate change or mass vaccination proposals?
Science Communication/Public Understanding of Science
The world of professionalized science consists largely of networks of people who have undergone extensive formal science education and training and who talk to one another using specialized language. It’s difficult for an outsider who wants informed answers to specific questions to just dip into the primary scientific literature and get what they want. This is not only because there is a specialized vocabulary to learn, but also because the professional scientific literature follows the research priorities of professional scientists. If the questioner does not frame his or her question in relation to those priorities, it will be hard to relate what is found in the literature to the question, even if the vocabulary has been mastered. Science Communication and Public Understanding of Science are attempts to bridge this gap by training scientists and journalists to explain science in ‘ordinary’ terms. These initiatives could, in principal, foster a general understanding of science that could help members of the “lay” public articulate their interests into proposals for research. In practise, however, much of what is produced under these headings at present is either aimed at persuading the public that the projects of professionalized science are aligned with their interests and therefore worthy of public funding support, or are aimed at showing how existing scientific knowledge can inform government policy decisions. If Science Communication and Public Understanding of Science are limited to communicating the research priorities of professional scientists to the public and understanding how science can inform the decision-making of professional politicians, then they ‘open’ science only by providing a window through which the public can gaze as an essentially passive audience. They don’t open the way to direct public involvement in driving local research priorities.
Volunteer Science/Citizen Science
The term “Citizen Science” is used to mean different things by different people. To some, it means professional scientists recruiting members of the public to assist with data collection or data analysis. In some online research projects this has involved large numbers of informal volunteer researchers. Accordingly, I would prefer to call it Volunteer Science. While Volunteer Science certainly allows the public to get involved in research, it is rather like crowdsourcing in that most such projects so far seem to rely on direction by professionals. It remains to be seen whether online networks of ‘lay’ people who have a common civic interest or problem thought to be amenable to scientific investigation can recruit professional scientists.
Such engagement of professional scientists with those outside the world of professionalized science is described by Jack Stilgoe in Citizen Scientists – Reconnecting Science with Civil Society. Stilgoe’s Citizen Scientists are “people who intertwine their work and their citizenship, doing science differently, working with different people, drawing new connections and helping to redefine what it means to be a scientist”. The Citizen Scientist is motivated by a sense of engagement with civic interests that not only permeates, but actually drives his or her research interests. Research priorities are chosen not on the basis of what maintains one’s reputation and status within a professionalized science community linked by a professional interest in science, but rather on the basis of how they contribute to the needs and ambitions of a civic community linked by place or civic tradition.
How might Open Science contribute to the advance of Citizen Science?
Opening Up Open Science: The Possibility of Civic Research
We have seen that the concept of Open Science encapsulates a variety of initiatives, each of which encourages more openness, closeness or collaboration between the various parties involved in the scientific enterprise. Advocates of Open Science generally argue that the benefit of these developments is that they will accelerate the advance of science. In Reinventing Discovery, Michael Nielsen looks forward to “a new era of networked science that speeds up discovery” and assures us that this “will deepen our understanding of how the universe works and help us address our most critical human problems”. Inherent in that is the idea that science can and will (eventually) answer every question. Maybe so. But who decides which questions we tackle first? As I’ve tried to argue above, most of the initiatives of Open Science leave that question open to the status quo. In effect, that means professional scientists acting within the culture of professionalized science itself in conjunction with government or other organizations that sponsor them. There is a presumption that these are effective at deciding what “our most critical human problems” are and then translating them into the most appropriate courses of action for scientists.
Many civic organizations, associations, networks and individuals perceive issues or problems in relating their own particular interests to those of other members of society. While resolution of such issues is ultimately political, progress towards resolution may be advanced in some cases by some kind of scientific investigation. Such investigations we may call ‘civic research’. Groupings that might instigate such research could include NGOs, patient advocacy groups, consumer rights groups, local residents associations concerned about environmental contamination or pollution, farmers concerned with land stewardship and others. Because the concerns of such civic groups are ultimately political, because the types of scientific investigation they want often do not align well with the priorities of professionalized scientists and because such groups often do not have sufficient funds to engage the services of scientists on a professional contract basis, working with them is often not attractive for professionalized scientists. For engagement with such groups to become attractive, scientists have to be personally motivated by the political objectives, not just the ambition to pursue a scientific career. The scientist is a committed political actor whose contribution to the project happens to take the form of scientific knowledge. The science is as overtly political as the aims of the group. Nevertheless, to be effective in that role, to bring to it the benefits of as broad a range of scientific experience and understanding as it can use, scientists need to be connected to and to draw upon the world of Open science. Although that world is not aligned with any specific civic commitment, it can inform countless committed initiatives. Its openness also allows it to draw upon, integrate and grow from the submitted experience of countless scientists individually committed to overtly political programs of civic research.
Is the possibility of civic research an alternative to professionalized science? On the basis of the above, it seems that if Open Science can open up the prioritization of research problems to be addressed, it could sustain a lot more civic research than currently takes place. Moreover, if Open Science can create standards in dataset and metadata format, then the results of civic research projects could be more readily integrated into the Open Science knowledge base itself. Civic research could grow, be sustained by and eventually sustain the scientific knowledge base without the need for professionalized science. I intend to look more closely at this question in future posts.