Collaborative research on Wikiversity

Collaborative research in Wikiversity Submitted to Wikimania 2008 Cormac Lawler University of Manchester cormaggio@gmail.com

Abstract
This paper examines the domain of research in the context of a wiki: how a wiki can facilitate research, and some challenges and opportunities of such work. This is set into the context of Wikiversity, both in its development of a research policy, and in two case studies of existing research projects on Wikiversity. It is argued that participative action research is an appropriate model for doing research on a wiki.

Keywords: Wikis, research, methodology, action research.

Introduction: wikis in work and research contexts
The rise of wikis has produced new ways, or at least modes, of working. Collaboration is not new, of course, but the ways in which wikis are being applied in particular contexts may represent qualitatively new work forms for the contexts themselves. These contexts can include schools (both as tools for students [1], and teachers [2]), businesses [3], [4], and non-profit organisations [5] – and aims of their use can vary from improved military intelligence gathering [6] to emancipatory “self-determination” [7]. The impact of wikis is most apparent in Wikipedia’s revolutionary take on the encyclopedia [8], and is reflected in Time magazine’s recognition of “you” as its 2006 person of the year, with Wikipedia as a prime exemplar of a societal shift, mediated by the internet, or so-called “web 2.0” [9]. This paper focuses on wikis in the context of research – surveying their usage in research contexts, and exploring how wikis and research contexts can influence each other – specifically focusing on the context of Wikiversity. Through this, I wish to raise some important challenges and affordances of wikis in supporting and facilitating research, as well as to suggest particular means of researching wikis themselves.

Much of the literature on the use of wikis in research discusses a wiki as a ‘knowledge base’ (eg. [10]) – a strand of literature which is echoed in other contexts [11], and often extending the traditional repository to incorporating more social elements [12]. Hill et al. [13] describe use of a wiki to discuss research ‘best practices’; Knox, DePasquale & Pulimood [14] used a wiki as part of students’ group work about doing research; and Frumkin [15] hints at possibilities afforded by a wiki for researchers in annotating and documenting usage of a digital library. Of course, there are different wiki platforms with different affordances [16]; and Buffa, Sander & Grattarola [17] detail their search for an appropriate tools for their research needs, and describe some of their experiences. Overall, however, the literature is scant in detail of the specific implications of wikis for research, and how they can be used to organise collaborative research. I feel that this question can be illuminated by a look at Wikiversity, how it has been set up to facilitate research, and its applications of and implications for research.

Wikiversity
Wikiversity is a Wikimedia Foundation project dedicated to developing learning materials, learning activities, and learning communities. Officially launched in August 2006 (though it had been previously ‘incubated’ in Wikibooks), there are nine individual language projects, as well as a multilingual hub for inter-project coordination.

According to its scope, Wikiversity is envisaged as both a repository of resources, and a space for learning. The first of these fits directly with the work of other Wikimedia Foundation projects – creating, compiling, and aggregating content made available under a free licence; but the second is substantially more problematic, raising the question of how learning can be provided for in this wiki space. ‘Wiki pedagogy’ is a still-emerging field [18], but a strong theme within discussions of Wikiversity’s learning model is around ‘learning by doing’. Though this is not a new phenomenon (cf. [19]), there are unquestionably new ways of learning opened up by the online-networked world, perhaps characterised in the similar concepts of ‘learning through participating’ or ‘learning to be’ [20]. Tapscott [21] cites ‘Wikiversity’ as a new way of doing education, involving opportunities and challenges for students and teachers. Part of this challenge for the Wikiversity community involves ‘learning about learning’ – and part of this requires a research component, which I will now discuss in its historical and wider context.

Developing a research policy on Wikiversity
In the main, Wikimedia projects explicitly exclude original research – see, for example, [22]. Wikipedia has long used this policy as a buffer against unpublished theories – it is one of the central “pillars” of Wikipedia. Policies developed on Wikipedia have become influential across Wikimedia projects, or sometimes adopted de facto – however, it should be borne in mind that each project, and each language community, have always been granted a certain autonomy in self-determining how they work. The arrival of Wikinews challenged this policy, insofar as needing to allow for specific modes of research as an inherent part of the journalistic process. However, Wikiversity’s envisaged scope (as both a learning repository, and a university-like learning ‘space’) brought a wider dimension to the question of research – and hence raised a number of significant challenges to be faced. Thus, Wikiversity was set up with two major questions over its subsequent development – as already mentioned, how to build a model for learning that is appropriate to the wiki way; and how (or whether) to allow for research, with a recommendation on its launch to develop a set of guidelines around research on Wikiversity.

In a meeting on IRC in March 2007 about these research guidelines, and through various Wikiversity pages, a number of questions around research were raised – mainly centred around how the Wikiversity community would deal with “bogus” or “fringe” research, exactly what would constitute “bogus research”, and how it would be determined. Relevant pages are in the references [23], [24], [25], and a transcript of the IRC meeting is at: http://beta.wikiversity.org/wiki/Wikiversity:IRC_meeting_about_research/log

Proposed solutions include that:
 * All research-related content and activity would be tagged as research - perhaps with specific tags, like ‘research paper’, ‘research data’, ‘research project workspace’, etc. (note that ‘tagging’ in the Wikimedia context currently equates with ‘categorising’)
 * There would be a research review board/team set up to advise on aspects of the research that might be problematic (eg. ethical or methodological questions)
 * All research activity would be subject to regular wiki review processes, like ‘requests for comments’, or ‘votes for deletion’.
 * Research that promoted partisan or commercial interests would be disallowed
 * Research would be limited to ‘educational research’ (ie that which supports its learning materials/activities)

However, none of the above solutions, nor the formulation of problems that they address, are unproblematic – evidenced by heavy discussion in the IRC meeting, particularly around the second and third solutions, ie whether there would be a review board or a review process (or, indeed, whether there is a contradiction between these proposals). To illustrate this, and to aid further discussion, I will quote from the IRC meeting (where, in the first comment, the quotation was taken from Wikiversity’s proposed ‘Review board’ policy [24]:


 *  "The Review Board helps the community distinguish between valid research methods and bogus research methods." - how can a review board know what is valid in all disciplines?


 * [2 comments cut]


 * the RB (review board) are not expected to know about the specific content of the research, just to analyse what is submitted in front of them


 * [2 comments cut]


 *  WiseWoman: '... who have demonstrated expertise in particular subject areas ...' - The idea being, that the experts in a particular area can take care of it, I assume


 * for example, removing research that breaks the (proposed) ethical guidelines, or has a method flawed beyond repair


 *  gboyers, I do a lot of review of project proposals for various money-dispensing organizations. Even if they are in my field, it takes me quite a lot of time (for which I am normally paid by the MDO) to understand the proposals and then to formulate my problems.


 * but does that mean we need a computing referee, and a biology referee, and a history referee etc? we can't expect to have experts on all the subject areas that this research will involve


 * research methodology should be a learning project in itself...

The major questions here and overall were around whether a group of selected individuals could arbitrate on any research/proposal, and how such a group would be selected. Inevitably, this led to raising the issue of ‘expert’ credentials, and whether such a group would be formed on this basis (with the equally inevitable problems that this poses in an online context). This notion is nuanced by shifting in focus to “expertise”, which could be derived from wiki editing experience, as was already noted in the policy [24]; though there was unresolved discussion on whether expertise would be solely based on wiki-editing experience, or whether someone could claim expertise by pointing to work done elsewhere. However, in this issue and overall, tensions arising from this discussion seem to be inherent, ever-present and never unproblematic features of this context, calling for an open, discursive, developmental, and (self) critical process of wiki policy-making [26], [27].

Equally, on the question of whether such a process would require people of expertise (or, at least, competence) in ‘all areas’, it seems logical to assume that evaluating research does require some expertise – though I would argue it is difficult to answer definitively what the criteria and boundaries of this expertise are, aside from problems of evaluating expertise in such an environment. In any case, it sets up a real challenge to the Wikiversity community in evaluating research which ‘might’ be problematic. Indeed, in WiseWoman’s stated opinion, in defining what is ‘unethical’ or ‘illegal’ research, “you end up reaching into a can of worms”. In this context, I find the following (later) comment useful:


 * [..] determining "good and bonafide research" is a very hard problem, and a core driver for the existence of NPOV on other projects.

Gmaxwell’s raising of ‘NPOV’ is especially acute in this context. ‘NPOV’, another ‘pillar’ of Wikipedia, requires that articles be written from a “neutral point of view”, in which multiple perspectives on a topic are acknowledged and discussed, and no one is privileged [28]. However, it has long been discussed whether this is a useful policy for Wikiversity to adopt de facto, or whether it needs to modify it or develop an alternative to it for its own needs. This has given rise to an acknowledgement in Wikiversity of the value of ‘NPOV’, its echoes in the codes of ‘scholarly ethics’, and an invitation to base materials and/or activities on a point of view, while explicitly acknowledging the nature of that ‘POV’ [29], [30]. Of course, the Wikipedia ideal of NPOV is also problematic and difficult to achieve – and is both a source of much conflict, but also a means of conflict resolution (or a “heat shield”, as Reagle [31] put it). In other words, striving towards NPOV constitutes a process of discussion and negotiation, often between different epistemic or world views – and the NPOV policy document is a working resource to facilitate a community-based process. Similarly, in this IRC discussion, it was advocated by Sj that Wikiversity would adopt “a review process […] with specific procedures but not specific people” – in the same community-based way as is done in discussing the neutrality or deletion of articles [32]. There was critique from JWSchmidt as to how the proposal of a community-based discussion/deletion process differed from a community-based selection of reviewers process; however, I would take from other comments that he saw the latter proposal in the context of a community-based discursive process, to be augmented by some form of recognition of expertise.

What I hope to have done here is to portray the subtleties and complex, problematic nature of many of these issues. (Note, however, that “problematic” does not for me equate with something necessarily negative – often quite the opposite, in that it can provoke in-depth discussion of the issue at hand.) I would also like to emphasise that the agenda of research on Wikiversity has always been seen in the context of learning about doing research – as echoed in Yeoman’s comment above about methodology being a “learning project in itself”. This attitude is extended by Yeoman’s plea to shift focus from a “flood of crackpots”, to “exploring ways of attracting serious researchers” (and to which I would add: “ways of developing senses of good research practice in researchers”). And clearly, this is to be developed in a manner which is appropriate to not just the mission but the climate of Wikiversity – as distinct from Wikipedia and other projects – as is seen in the following comment:


 * so let Wikiversity make the mistakes - let WV be the place that has POV and discussion - let WP keep the verified facts

Finally, I would note that the proposals I initially outlined above do not constitute a combined policy system (i.e. they will not necessarily all be utilised, or all in the same way) – and I also note that the policies in place are mostly a proactive structure, while leaving room for reactive developments. In other words, these proposed solutions have to be seen in the context of dealing with a largely imagined problem (“flood of crackpots”), rather than one which has been experienced directly (except perhaps, and in a different context, in Wikipedia). Solutions will only be tangibly shaped through real cases (i.e. real problems), which will result in the usual developmental model of wiki policy-making. So, in an attempt to relate this discussion to real experience, I will now proceed to situate this developing policy (and discussion) within the context of research practice in Wikiversity.

Wikiversity research case studies
Given that research is a component of Wikiversity, I will outline two case studies - “Bloom Clock” and “Developing Wikiversity through action research” - to illustrate how Wikiversity is being used to conduct research, and then proceed to use these case studies to illuminate the wider context of such research.

Bloom Clock
Bloom clocks are used by gardeners and horticulturists to keep track of what is blooming at a given time in a specific location – often the gardener’s garden – and this data is generally kept from year to year to track changes in plants’ regional blooming seasons. The Wikiversity bloom clock [33] is an attempt to create a distributed bloom clock – ie by involving multiple participants across multiple locations. It works through having a ‘log’ page for every plant, and asking participants to add their wiki signature to a plant’s page on (or close to) a day when the participant has seen the plant blooming – which constitutes ‘logging’ the plant as in bloom. In order to correlate the time of blooming (in the timestamp of the signature) with where the plant is blooming, the participant is asked to add their (user)name and location to the participants’ page.

In order to help people identify plants, there are sets of ‘keys’, working in a similar way to traditional plant/animal identification books – eg. “flowering in late Spring” >> “with white flowers” >> “is herbaceous” – giving a list of plants in all these categories, with photographs to aid identification. If the keys do not help (ie are not yet fully comprehensive), there is also a page for identifying unknown plants, where participants can upload a photo or describe the plant/flower, and other participants will try to identify it.

As with everything in a wiki, the bloom clock is a work in progress, and will improve as more people in more parts of the world begin to participate (though even if only a few people contribute consistently, the bloom clock will become more accurate over time for the areas and plants covered). The bloom clock’s methodology is envisaged to being expanded to other similar distributed data collection projects such as an “entomology clock” (on insects), and on particular organic farming methods - and the templates used in the bloom clock have been adapted for use in a project to glean demographic data about Wikimedians.

Developing Wikiversity through action research
This project [34] has been instigated by myself to develop an understanding of Wikiversity’s work, its conceptual underpinnings, and its wider context (amongst other things), and to use these understandings to develop systems and resources to help participants carry out their work on Wikiversity. I’ve proposed that this be undertaken within an ‘action research’ framework, which I see as a particularly congruent methodology for studying wikis [35]. This is echoed by Atkins, Brown & Hammond [36], who, in outlining enablers for what they call an ‘open participatory learning architecture’, call for ‘action-based research’ to be carried out that would generate “deeper basic insights into human learning (both individual and community)” (p. 36). The project aims to collaboratively build knowledge about Wikiversity in the process of developing Wikiversity, and hopes to generate insights that are meaningful to its participants, and perhaps to participants of similar projects.

Action research is difficult to pin down as a specific methodology [37], but it is essentially concerned with studying a context or process through changing it, and attempting to improve it. It is a developmental, cyclical process – involving planning, taking action, observing and reflecting on the action, and using observations to form further plans, actions, etc. [38], [39]. As part of a ‘participative paradigm’ [40], it is research with as opposed to on people [41], and influenced by the ‘empowering’ work of Paulo Freire [42]. This type of ‘involved’ research has both challenges and opportunities for the researcher [43] – and this particular project I am outlining is further expanded by dealing with an online context [44]. However, my major rationale for equating wikis and action research is based on their dedication to continual systemic improvement through an egalitarian, discursive model [7], [45].

Specifically, this project is addressing the basic questions around a ‘wiki learning model’ that Wikiversity was set up to answer. For example, how ‘learning by doing’ works in this context is only partially and sketchily understood amongst the community – and this project seeks to clarify what it means, and how it can be utilised. Another main question is around building and sustaining learning communities – and yet another is about making Wikiversity work as a ‘personal learning environment’.

Discussion
If Wikiversity were only to allow educational research, it would seem that the Bloom Clock project would not fall within that scope – unless the project was seen as a learning project about plants (a function it clearly serves). This suggests to me that limiting research on Wikiversity to educational research would be blinkered to the fact that all research-related activities (doing research, writing about research, and reading about research) are inherently learning processes, and conversely, that all educational work will at some point lead to a research component – hence its prominence in the modern university. It would also be unclear where to draw the boundaries in such a policy.

Wikiversity’s bloom clock is, so far, a quiet, though still-early success – but the prospect of editable databases for doing research does pose some obvious questions. How can we safeguard large volumes of data in an editable environment? (Scientific data worldwide is thought to be generated and stored in the order of petabytes – where 1 petabyte = 1million gigabytes [46].) In light of this, should we allow for ‘protected’ spaces, and if so, how? What would be a productive and pragmatic ‘boundary’ between closed and open editability? It is also unclear as to the public use of data which is released under the GFDL licence which all Wikimedia projects use [47] – for example, there are other initiatives to provide legal frameworks around publicly available data, such as the Public Domain Dedication and Licence [48] and, perhaps less explicitly, Creative Commons’ ‘CC Zero’ licence [49]. Furthermore, and as raised in discussion, an irony of Wikiversity using a free-content licence is the possibility that research reports written on Wikiversity could be difficult to submit to particular journals, where those journals require copyrighting the published work.

The scientific data mentioned will involve any number of formats (text, image, video, etc), but volumes of words alone which are generated in online discourse are sizeable (as anyone involved in such discourses will testify!) – which pose both opportunities and challenges to the researcher [50]. Also, particularly in the type of collaborative action research outlined above, there are clearly ethical questions around using all participation as data [51], as well as difficulties that any collaborative (and particularly qualitative) research team will have in coming to common conclusions about a particular piece of data [52].

Seeing as academics and others are increasingly relying on sophisticated technological tools, or sometimes ‘virtual research environments’, to facilitate their research processes [53], it is a valid question to see how wikis fit in with this technological array, and whether a wiki could itself constitute such a research environment. The obvious question of choosing a wiki platform appropriate for specific needs must be addressed [16] – for example, semantic functionalities would be clearly important for some groups [54]. Mediawiki, the technology that is used on all Wikimedia projects (and many others), has many affordances, but it is also limited in some respects – and there seems to be a need for researchers to discuss what technical features would be necessary for them to be able to use Wikiversity productively. As raised in discussion, this might involve a social dimension – for example, whether groups could ‘protect’ (or lock) pages from editing, and if so, by what mechanisms and through what processes this would be carried out.

It is quite clear -– from discussions here and from general experience – that ‘research’ does not mean one thing, nor spring from one tradition; epistemological, theoretical and methodological strands of academia and their followers will continue to work in ways that are substantially different and often completely opposed to each other. While, as already discussed, Wikiversity allows for materials and activities based on particular points of view (and this is a necessary freedom for doing research), Wikiversity could extend the tradition of NPOV, in developing an open, scholarly peer review system of critical feedback on research projects from differing traditions. This would be a major component of having research as a communal learning activity within Wikiversity – whether or not there evolves some form of ‘review board’.

Conclusion: opportunities and challenges
As has been hinted at, many of the issues and implications that a wiki raises for practice (particularly research and education) are both challenges and opportunities. For example, there are obvious – and legitimate – fears about students’ perhaps uncritical use of Wikipedia in doing research [55] – but there are also those who argue that Wikipedia offers a chance to gain critical information literacy skills [56]. My own bias here is in favour of the latter, while incorporating a deep and honest acknowledgement of the former’s fears. It is also partly in this context that I have advocated the critical action research model – where participants of a project are encouraged to reflect on the mutual influence between the wiki and the context (research, practice, project), making the process of doing research more transparent.

Even though wiki-based work is new in many ways, it does not necessarily represent a ‘clean break’ with existing ways of working – in particular, research methodologies [57]. I have tried to indicate here how a wiki can be thought of in methodological terms, and the implications it raises for the applications it is put towards. However, there needs to be much more in-depth academic discussion about the use of wikis in research – my literature review for this paper has revealed mainly descriptions of their use as a knowledge base, but with little discussion or analysis of issues related specifically to research. I hope this paper will have at least raised issues that can be introduced into this ongoing discussion.

Though it can certainly be used as such, a wiki is not simply a repository for knowledge to be chunked and managed, but a social space where knowledge is actively constructed [58]. If Wikiversity is to realise, sustain and drive the ideas behind the OER movement and the development of a “meta-university” [59], it needs to take account of what knowledge it is generating and how – both at the level of ‘learning/researching’ and ‘learning about learning/researching’. There is clearly room for many types of research to take place on Wikiversity, but I argue that, in order to address the meta issues of research in a wiki context, such research needs to be in the latter category. Furthermore, I think that research done ‘the wiki way’ must involve both wide participation and critical self-reflection (by the individual and the community), and, even though I by no means see it as the only way, I cannot think of a better model for ‘wiki research’ than action research.


 * Note on data in this paper: In quoting from the IRC meeting on research in Wikiversity, I’ve taken the liberty of making minor spelling corrections.