D. R.
Newman
Queen's University
Belfast, Information
Management Dept.
<d.r.newman@qub.ac.uk>
Presented to ERCIM and W4G Workshop on CSCW and the Web.
This paper shows how we can design computer support for group, co-operative learning, in order to promote critical thinking and deep learning. From Garrison's theory of critical thinking we can evaluate the student learning style in face-to-face and computer supported group learning. From our pilot study results and this theory, improvements in CSCL software and teaching have been identified, and a possible way of integrating support for all stages of critical thinking in a World-Wide Web (WWW) based environment.
Other papers at this workshop cover the how of getting CSCW on the WWW. This paper addresses what we should build to support co-operative learning and problem-solving, and why.
Three years ago, I was challenged to write a chapter predicting what effect groupware might have on learning in the 21st century (Newman 1994). Here is an extract from the imaginary dialogue of the next century.
Two students, Nuala and Raphael, are discussing their assignments. They will soon have to go to class.
- Nuala
- Are you saying that in the '90s all students worked completely alone?
- Raphael
- Sure, they went to lectures, took notes, wrote essays and memorised everything for the exams.
- Nuala
- Seems like they would never have understood anything well enough to use it at work. But didn't they even discuss the subject?
- Raphael
- They had seminars in which half of them would turn up. Half of those would talk, when pressed to say something by the lecturer: the rest would hope that they would not be noticed in a group of 30 and waited for their release at the end of the class.
- Nuala
- So at least some of them did discuss the issues.
- Raphael
- Well, not quite discuss. One or two of them would present their summary of the paper everyone was meant to have read, and others would daydream until it was their turn to say a few trivial things. Within an hour it was over.
- Nuala
- Just an hour? And they say that education standards are declining. In one hour they could only have skimmed the surface. No wonder ministers and civil servants seem so stupid, they just leaned things superficially.
- ...
- Nuala
- OK. But we've got class now, better switch it on.
In their world, many of our common learning situations will seem surprising. Today many students commonly learn to study on their own, no cheating, then the teacher marks it. Afterwards, if they get a job, they will be expected to work in teams, sharing work, evaluating each others' contributions: i.e. do co-operative work. At work we solve problems. We need a deep understanding that allows us to apply our learning in new contexts, not a surface skim for the exam paper.
In this world of accelerating technological and economic change, we continually improve our work, rapidly learn new technologies, and rapidly develop new social institutions to embody these technologies (Ellul). For this we need deep learning. Thanks to educational research and learning experiments, we already know how to promote deep learning. In particular, increased student activity (not passively listening), rapid feedback and group work are needed. For example, we can break up large, face-to-face, classes into smaller active groups using rounds, line-ups, pyramids, projects, courts of enquiry, posters, brainstorming and so on. Students can carry out group research or development projects, case studies or presentations. Critical peer and self-assessment can be used instead of tutor marks to encourage critical thinking (Gibbs & Jenkins 1992).
The challenge is to use computer tools to support such active deep learning, or indeed, any critical thinking process during group work, in either of two ways:
These approaches are different from merely providing tools to facilitate single learning tasks such as report writing, and expecting learners to immediately make optimum use of them. It is not enough that groupware works for a given task: it must encourage critical thinking in the problem-solving process of which the task is a part.
In September 1993 I went to ECSCW Workshop on Evaluating CSCW, organised by Liam Bannon. Not a single CSCW developer turned up. It seems that only CSCW users and implementers in large companies and in universities were interested in evaluation.
Of those that came, many had measured anything they could, then came and asked what could they do with the data? This unscientific approach can lead to going round in ever-decreasing circles, like the Oozelum bird.
The serious approaches to evaluation start from theories in:
The last approach helps in evaluating the quality of work done. For the type of work we call learning, we can measure its quality, since we have testable theories of learning.
A number of theorists have considered critical thinking as a problem-solving process, rather than just a set of skills to employ in reasoning tests or in essays. Processes have stages. For example, Brookfield (1987) and Garrison (1992) identified similar five phase processes of critical thinking. A triggering event causes us to engage in self scrutiny and appraisal, then explore ways to make sense of the situation, and finally integrate the new way of thinking into our existing experience and knowledge. Garrison's model is a dynamic, cognitive one, similar to models of problem-solving used in cognitive psychology and artificial intelligence. Although first proposed as a model of individual learning, it requires shared understanding with others, making it well suited to the evaluation of group learning. To this end, Anderson (1993) has placed each stage in a social context, in his study of distance learning by audio conferencing in Canada. These parallels are shown in Table 1.
| Stage | Critical thinking (Brookfield, Garrison) |
Critical reasoning skills (Henri) |
Group learning (Anderson) |
Deep learning (Biggs, Entwistle, Ramsden) |
| G1 | Problem identification | Elementary clarification
observing or studying a problem, identifying its elements, observing their linkages |
Triggering event
within social environment |
Motivation to learn |
| G2 | Problem definition
define problem boundaries, ends and means |
In-depth clarification
analysing a problem to understand its underlying values, beliefs and assumptions |
Using experience
of others |
Discussion
building overall description of content area |
| G3 | Problem exploration
ability to see to heart of problem based on deep understanding of situation |
Inference
admitting or proposing an idea based on links to admittedly true propositions |
Using multiple perspectives
and experiences |
Discussion
use of previous knowledge and experience |
| G4 | Problem evaluation/applicability
evaluation of alternative solutions and new ideas |
Judgement
making decisions, evaluations and criticisms |
Social modelling
of problem |
Discussion
relate evidence to conclusions |
| G5 | Problem integration
acting upon understanding to validate knowledge |
Strategies
for application of solution following on choice or decision |
Validation of process
through group interaction |
Maintaining a stance
critical and objective stance |
| (problem-solving process) | (skills needed during process) | (social process) | (preferred attitudes and behaviours) |
Others have developed ideas of deep versus surface learning styles. When deep learning, learners attempt to reach an in-depth understanding of the subject. When surface learning, they memorise surface features for later recall (e.g. in examinations or medical emergencies). Ramsden (1983), Biggs (1987) and Entwistle (1992) have linked deep learning to interaction among learners and teachers. We can identify parallels between features of deep learning and Garrison's stages of critical thinking.
Henri (1991) identified five dimensions for analysing Computer-Mediated Communication (CMC): participative, social, interactive, cognitive and metacognitive. Questions of deep learning and critical thinking are in the cognitive dimension. She laid out 5 skills needed for critical reasoning, to be used one after the other, roughly corresponding to Garrison's stages.
In fact, when Henri set out to measure the cognitive dimension, she classified statements separately according to which stage of critical reasoning they came under, and how deep the learning style was, as shown by a set of indicators. So learners can adopt deep or surface learning strategies at each stage of critical thinking.
Finally, we can consider how learners' understanding of the problem develops as they go through a critical thinking process (the SOLO taxonomy)
Given these theories, we can:
When groupware developers talk about evaluation, they are often concerned with how well they get computers to communicate with each other, at the level of files, documents, clients and servers. Many of the papers at this workshop show how they succeeded in this despite everything, treating the WWW as a transport mechanism: most clearly in AlephWeb. The key issues are compatibility, reliability and performance.
The next level of evaluation is that of communication between a single human user and a computer. Part of the reason for the massive expansion in use of the WWW has been the simplicity of the standard browser interface. I spend 2 hours teaching my students to use the PowWow computer conferencing system, but only half an hour teaching them how to use Netscape for both browsing and searching. So an easy-to-learn interface is a necessary condition for the use of CSCL.
But good HCI is not sufficient to ensure that work done (learning in this case) is improved by the use of a computer technology. This is a cognitive and social question.
The ultimate measure is the long-term employment success of learners after finishing a course. Shorter term student assessment techniques like examinations do not measure critical thinking or deep learning alone, as they depend greatly on memory. But we can find out, for any type of work, the approaches used by the best workers and the worst, then look for indicators of which approach is being adopted while the work is going on. This is a technique for evaluating the work rather than assessing the workers.
Theories of critical thinking identify attitudes and behaviours that correspond to deep and surface approaches to learning. The attitudes can be measured by self-assessment questionnaires (Webb, 1994), the behaviours by content analysis (Newman 1995a). These techniques can then be used to measure differences in critical thinking when using different computer technologies or learning situations.
At Queen's University Belfast, Clive Cochrane had been using conventional seminars in his 2nd year undergraduate Information Society module for discussing controversial issues, so that our Information Management students could learn there is no one right answer to the questions. We hoped to be able to use computer conferencing to enable the same number of tutors to handle increasing student numbers, without affecting the quality of learning.
We set up a controlled classroom pilot study, in which the students of an Information Society module did half their seminars face-to-face, and half over our Network Telepathy computer conferencing system (Ashmount Research). This was a COSY system, but with a threaded browser and integrated editor. Students could go in each day and leave comments on continuing discussions.
We evaluated the learning in two ways. The students filled in questionnaires at the end of the semester on their learning experiences in face-to-face seminars, and in computer conferences. The questionnaire were analysed by factor analysis. The analyses, reported in detail by Webb et al. (1994), suggested that critical thinking was taking place in both environments and with similar emphasis. One difference between the two was found in the identification of personal experience. This was highly positively loaded in a factor under seminars, but highly negatively loaded in a factor under computer conferencing. From answers to open ended questions, it seems that this may be related to student's being put off by the difficulty of learning the Network Telepathy conferencing system. For this reason we later changed to a Windows version of the product (PowWow) and are now looking for WWW-based systems, since even PowWow requires the student to learn a lot of buttons and menu items.
Secondly, several seminars were tape recorded and transcribed. Using the methods described in detail in Newman et al. (1995a), the seminar and computer conference transcripts were scored for phrases indicating obviously deep or surface approaches to learning. These were counted, and a set of depth of critical thinking ratios were calculated, on a scale of -1 (all surface) to +1 (all deep). There were significantly higher overall depth of critical thinking ratio for the computer conferences, reflected a more serious writing style (half-way to essay) with fewer irrelevant distractions, at the expense of participation.
A detailed account of the content analysis results has been published in the electronic journal IPCT-J (Newman 1996a), and a summary paper covering our approach and the results of both evaluation techniques will appear in JASIS (Newman 1996b).
But to get an idea of how the results, have a look at the
graph on the right. We mapped the indicators from the content analysis on
to the
stages of critical thinking at which these behaviours would be
observed (e.g. new ideas in stage 3), and then plotted the
depth of CT ratios against Garrison's stages.
Over all three groups, it seems that computer conferencing encouraged deeper approaches to critical thinking in stages 1, 2, 4 and 5. But computer conferencing gave little, if any, benefit to creative problem exploration in stage 3. Also, stage 5 seems to depend more upon the subject being discussed rather than the technology: making this an issue for course rather than software design. However, other software designs than COSY conferencing system might better support problem exploration.
In the pilot study we did not attempt to optimise the technology and instructional context to promote critical thinking. We ran conventional seminars both face-to-face and on the computer conferencing system.
Now, however, from theories of critical thinking, and the results of our pilot study, it is possible to suggest some combinations of technology, teaching technique and learning task that would most encourage deep processing or deep learning styles in each stage of the critical thinking process.
It is possible to design groupware to support each of the stages of critical thinking. This is an advance on the common use of Computer Based Training to transfer information or practice skills (e.g. Portuguese grammar), leaving the deep learning to face-to-face sessions with a tutor (such as Portuguese conversation). Instead we can support critical thinking as a group process by using computers to mediate communications during group learning activities.
For some time, groupware developers have taken particular communication activities and written software to allow that activity to take place over computer networks. Some merely offer a tool, an equivalent to something we do face-to-face or on paper, and leave us to organise our work to take advantage of it. So we have support for unstructured discussions in computer conferencing and video conferencing, or support for directed, one-to-one, messages in electronic mail. Others are designed to support a range of activities that people do in completing a particular task, and provide ways of structuring our approach to the task. Engineering design tasks are supported in shared drawing space systems, and writing tasks in group editing systems.
But even the task-based systems do not effectively support all stages of the critical thinking process. The particular skills and behaviours corresponding to deep learning approaches change from stage to stage, as shown in Table 2, which is based on critical thinking theories and our experience of computer supported seminars.
| CT stage: | G1 | G2 | G3 | G4 | G5 |
| Acquiring outside information | X | ||||
| Contact with outsiders | X | X | |||
| Organising/structuring ideas | X | X | X | ||
| Structuring documents and other outcomes | X | X | |||
| Stimulating ideas, creative widening | X | ||||
| Group size changing | X | ||||
| Creating and using problem models | X | ||||
| Critical feedback | X | X |
This is a challenge to groupware designers. Software that is ideal for organising problem descriptions or possible solutions may hinder creativity during problem exploration (as PowWow does).
Furthermore, learners have difficulties learning many different computer systems. We found a negative factor in our analysis of student questionnaires that seems to relate to the difficulty of learning to use the Network Telepathy conferencing system. To overcome this drawback, the course designers and software developers need to work together to bring a series of tools supported different stages of critical thinking under a common user interface, preferably as simple as the click on something underlined interface of WWW browsers, without any complicated menus or drag and drop. (I have spent a long time trying to explain to students how to hold down a mouse button - no, don't let go - and then move the mouse - now you can let it go!)
In the following sections, is a discussion of how computers may support each stage of critical thinking.
Learners start by identifying a problem, perhaps by scanning the environment for unusual conditions, and then gather information on it. This is a fairly divergent and exploratory activity.
Information retrieval and access tools help at this stage. Lecture notes (Newman 1996) and discipline-based resource collections (such as the ISWorldNet, Ives 1995) on the WWW can be used to discover problems, their details and elements. By using graphics on the WWW, learners can be motivated a little more than when reading ordinary lecture notes. However, this is still a fairly passive process, with little opportunity for critical thinking. Setting different readings for every learner can help when they come to discuss them later, as the group will have been exposed to more perspectives. A simple sign-in script on the WWW could help: but I've yet to find one.
A bit less passive is searching for relevant information, using WWW indexes like Alta Vista. AlephWeb improves on this by using http to keep a distributed database of company data up to date. If you go a step further and use it in associations of learning organisations, you can store the results of anyone's problem-solving exercise locally, and make it automatically available to other problem solvers across the association. So learning accumulates, synthesising new knowledge out of low-level information.
But it is interaction with others that best arouses and sustains interest and curiosity in the problem: e.g. between the learners and people with the problems (problem owners).
Learners can enter communities of professional problem owners, such as Usenet newsgroups, mailing lists or WWW-based discussion lists such as some of the sites that are using Futplex (Holtman 1996). This can sometimes lead to resentment by the professionals, as happened when a class was instructed to take part in the IPCT-L discussion list, but helps learners quickly identify a wide range of issues and perspectives.
Or outsiders can be brought into contact with a group of learners, in seminars, local computer conferences, or in special synchronous sessions, using Internet Relay Chat, Worlds Chat or CU See Me. Some of the WWW-based chat systems demonstrated at the workshop could be used for this (van Welie and Eliëns 1996, Walther 1996). The synchronous systems should encourage more participation and interaction, leading to better problem identification, although they may hinder the next stage, organising what learners know into a precise problem definition.
We found that our students took a deeper approach in this stage when using computer conferencing. It appears to have helped them avoid distractions and keep to the point of clarifying the problem. They brought up proportionately more important versus unimportant points and more outside knowledge (from textbooks or their own experience) in the computer conferences.
To improve on this, we might use an Issue-Based Information System (IBIS). These let people taking part in a discussion map out the key issues or problems, alternative positions they take on each issue, and advantages and disadvantages of each position. There are commercial programs like CM/1 that let people draw out such maps from a computer, either during a face-to-face meeting, or at their convenience over a LAN. And an IBIS will form part of the mediation system to be used in the GEOMED system to support unco-operative work on planning decisions (Gordon 1996)
This IBIS example shows a possible WWW interface for an outline of an IBIS discussion on Internet pornography. Such a WWW-based IBIS was set up at MIT, under their Political Participation Project for one of the early US Government on-line discussions for federal employees.
For learning, there is a disadvantage of IBIS. It requires the participants to identify a position before putting arguments for and against it. A group of learners may not know all the perspectives on a problem at the start. Instead they can first identify the variables which are relevant to the problem, then develop arguments based on these, and finally fill in the positions these arguments support or oppose. However, any kind of clearly structured on-line discussion should help problem definition.
This is the most creative part of the critical thinking process, where learners explore the problem and possible solutions. They use both logical reasoning and lateral thinking to extend their understanding beyond the basic problem definition.
In this phase, we found little or no advantage in computer conferencing. In fact, given the lower participation, it seems that computer conferencing may inhibit the creative generation of new ideas.
Learners, like managers, can be helped to come up with new ideas by using creativity techniques. A typical creative problem-solving process follows these stages:
Most creativity techniques have been developed for either face-to-face meetings or individual use. High bandwidth video or audio synchronous CSCW could also be used, as could desktop conferencing, where groups type notes and doodle on shared whiteboards. Socially, Worlds Chat, where your avatars wander around imaginary rooms, and MUDs, where you can design your own rooms, can stimulate creativity.
It is harder to see how WWW-based systems could support creativity. Koen Holtman (1996) came across a case in which his Futplex shared document editing system was used by design students in the Netherlands and Texas for near real-time chat. Perhaps they were forced into this when trying to use a stage 4 and 5 tool at stage 3. The chat systems already mentioned might be more appropriate.
It may be possible to support creative ideas generation on-line with enough time, say some days or a week for each of the steps listed. See my creativity example of how one set of techniques, Synectics, could be supported on the WWW. NB: I have a licence to use these techniques with my students. If you want to do the same, email me, and I'll put you in touch with the licensers. For any creativity technique, the participants will need to be trained in its use first, before they can use it over computer networks.
Also useful during this stage are techniques to break larger groups into smaller ones. Unfortunately, most current computer conferencing systems do not make it easy to change group membership every 10 minutes, so this is usually achieved through classroom instruction. Perhaps Web4Groups (Roland Alton-Scheidl 1996) will support this.
It is in this stage where learners judge, justify and critically evaluate solutions, link ideas together, and check out how well solutions will work in practice. These are things we found were helped by computer conferencing, particularly linking ideas.
But computer conferencing systems do not help you organise your ideas, since they merely show the threads of comments to comments. An IBIS can help in choosing between solutions: particularly if augmented with a voting system to run preferenda (Emerson 1994). Beyond that, synthesising the ideas into a workable solution can be supported by group outliners or editors, in which participants can add, redraft and reorder points. I have not come across any group outliners on the Web, but there are now experimental group editors that work over the WWW, such as:
When decision-makers plan the implementation of a solution, or learners integrate it into existing knowledge, they need to validate the solution in their group and externally. This requires feedback. This is not always possible within some groups, as the privacy seminars showed, where the subject discussed had a greater effect on stage 5 depth of critical thinking ratios than the technology. In such cases, many learners are stuck at early stages of the SOLO taxonomy, perhaps seeing several perspectives, but not enough relations between perspectives and issues to bring them together into a conclusion. An explicit representation of the issues and perspectives (as in an IBIS) can help here; or external feedback.
There are three types of feedback needed here. Criticism from other group members can be supported via group editing software. Feedback from external experts can be facilitated either by live presentations of the students' work, or electronic communications (email, conferencing or WWW publishing).
And feedback from the tutor comes in the form of marks. To support the administration of assignments, you can use systems like WEST to deliver assignments, which individual students submit, and the teacher then has to mark. (WEST 1995) Given good marking, the students may be challenged to think more about their work and the subject, and start to deepen their learning. But it takes a lot of work to give such detailed feedback, and shorter courses give less time to run through cycles of improvement. So we need tools to allow feedback to be given on work in progress: perhaps by supporting peer assessment in group editors, or even simple CGI scripts for feedback.
| CT stage: | G1 | G2 | G3 | G4 | G5 |
| Information retrieval, e.g. WWW, CDROM | X | ||||
| Asynchronous discussion open to outsiders | X | X | |||
| Asynchronous discussion in closed groups, e.g. computer conferencing. | X | X | |||
| Idea organising, e.g. IBIS | X | X | |||
| Synchronous discussion, e.g. IRC, CU See Me | X | ||||
| Asynchronous creativity support tools of the future | X | X | |||
| Group Decision Support Systems | X | ||||
| Group editing | X | X | |||
| Feedback and assessment, e.g. WEST | X | X | |||
| Information dissemination, e.g. WWW | X |
There are two key lessons to be learned from our pilot study, in the context of critical thinking processes.
Firstly, we need groupware designed to support each stage of critical thinking. Table 3 shows the current types of groupware that can be used at different stages of critical thinking. In the over 300 systems listed in the Groupware Yellow Pages (YP 1995), fewer have been designed to stimulate creativity and new ideas, compared to the numbers that help structure communications or organise ideas. Where are the computer support tools for brainstorming, Synectics, brainwriting and generative graphics that Hiltz and Turoff (1993) suggest? We are still looking for good computer tools to support the problem exploration stage, although Andrew Burnett (1994) predicts that computer-assisted creativity is possible.
Secondly, we need to bring together groupware supporting different stages of critical thinking under a common interface, so that learners do not have to learn many different user interfaces, and learners and teachers can quickly and easily progress through the stages of critical thinking without off-putting changes in the learning environment. Whatever lies behind the interface, it should appear seamless to the user.
This is different to providing common interfaces to existing, communication-oriented tools. Netscape Navigator 2, for example, provides windows for electronic mail, Usenet newsgroups (discussion conferences), and browsing WWW published pages. Clicking on a reference (URL) in a Usenet discussion will bring up the relevant page. Run Netscape Chat at the same time, and learners and teachers can type messages at the same time, showing each other references which appear in their Netscape Navigator window. We have made use of this to teach a class of students where to find relevant references on the WWW. An off-line integrated equivalent is Ashmount Research's Virtual Access. It links to news servers, commercial conferencing systems and Internet email when you can afford the telephone calls, and provides a message browsing interface to all of these while off-line. Both of these could be used by a good teacher to support many parts of the critical thinking process. But neither has explicit support for the stages of critical thinking, or the deep learning styles we wish to encourage.
Instead, we find that there are a number of CSCL tools being developed that support parts of the critical thinking process, as discussed earlier. Many of these, like Futplex, WebShare or WEST, use a WWW interface. So it now is possible to integrate several CSCL tools behind a set of pages for a course, and direct the students to use the best tool for the problem-solving stage they have reached. In our new Information Technology and Society module, we may be integrating WWW-based group editing with on-line course notes to support group projects. Eventually it may be possible to provide a seamless interface to all the CSCL tools that support critical thinking. For example, we could find, under one WWW interface, tools to support each of Garrison's stages of critical thinking about an issue that needs seminar-like discussion.
At present, the current tools to support interaction via the WWW are quite slow. So developers need to address their performance. However, since others are working on WWW performance, we may well find that the current experimental tools will turn into robust and educationally useful systems to support learning. Tim Berners-Lee gave an idea of the pace at which software is advancing, when he said at the launch of the European branch of the World Wide Web consortium:
Within two years, the browser as we know it will not exist. It will become the interface to the local computer as well as the remote one it is currently used to access. (Prigg 1995)
Since I have already explained the implications of our work on CSCL for groupware designers, the conclusions will reflect on the systems presented at the conference as tools to facilitate critical thinking in learning and problem-solving.
Many of the presentations at the workshop were concerned with getting computers to communicate with other computers over the WWW efficiently enough that they can be used for co-operative work. Hooray, Gaspode the dog talks! Now if only we can get him to do what we want.
Well computers move data around, so lets use them to keep collections of data: a document repository. This is how Lotus Notes was first used in the Irish Ombudsman's office, to keep track of case notes, making them up to date and available all the time. There already are WWW-based document repository systems such as BSCW (Bentley 1996).
But keeping track of documents, and who has done what at what time, has little to do with problem-solving. Such systems are tools designed for routine work, where a superficial approach suffices. Although they can be used in less routine situations, such as keeping a collection of images for a newspaper library or advertising agency, we can get an idea of their intended context of use from the concerns of the designers. Who is it that wants to keep track of every single edit of a document and when it was done? Typically, it is the project manager. It is managers of routine work who wish to restrict what people can do to documents and when, according to rule sets, as Griera (1996) proposes.
In learning organisations, the emphasis is on continual learning by all staff. They are always trying to improve their work, always learning. So problem-solving becomes more important than routine work, and effective critical thinking counts for more than efficient workflow. I am a little surprised that the CoopWWW ESPRIT project (Appelt 1996) will be working with users in professional organisations. Its document repository base seems more suited to large hierarchical organisations. It is the human parts of the system that will have to organise themselves to handle group learning and problem solving, rather than the computer parts (unless the co-operative decision support tools can handle non-choice decision making).
Perhaps the best (only?) use of document repositories in learning is in systems that make it easier to use the computer to present documents to others, such as the Dress system at Eindhoven University (de Bra 1996). A factor analysis of our student questionnaires revealed a strongly negative factor associated with the difficulty of learning a computer conferencing system that had an interface that many Computer Scientists would take for granted.
Now, according to theories of critical thinking and deep learning, these are favoured in situations which actively engage the learners, and particularly in group problem-solving activities. There are several ways in which computers can support such active learning, three of which are:
The recent rapid increase in use of the Internet shows how some of the applications running over it have succeeded in making human-human communications easier: particularly the WWW, but also chat systems. At the Revelations Cybercafé around 80% of the customers use the computers to enter World's Chat, an IRC-like system which lets you choose an avatar (a cartoon body image) and move it around a virtual room to chat with people.
A number of new applications link the WWW to systems that allow two-way communications. These include feedback and conferencing systems like Hypermail and Hypernews and the Web4Groups project (Alton-Scheidl 1996); linking IRC and WWW browsers together so that people can take others on tours around the WWW (e.g. Netscape Chat); WWW hosts for chat systems which you use to make the connection, then chat using faster protocols (van Wellie, Walther, 1996). It seems that thanks to a race between commercial and shareware developers, there will soon be a lot of WWW-based tools that allow simple two-way or multi-way communication.
So what is left for CSCW researchers to do? Well, these tools support rather simple models of communication. The real-time systems have many of the same advantages and disadvantages for learning as face-to-face instruction. They can bring together learners and teachers who would not otherwise meet, as Robinson did in his in-service course on Education for Mutual Understanding (Robinson 1995). Video conferencing brought together schoolteachers from both Protestant and Catholic schools in different parts of Northern Ireland. But they don't help the learners keep track of ideas or organise them. The simple computer conferencing systems allow rudimentary organisation of statements (in threads), which showed up in our content analysis as an increase in linking ideas, but don't necessarily stimulate new ideas, or allow learners to try out their own ways of organising and linking issues and perspectives.
Several of the systems presented at the workshop do help organise ideas, mainly the group editors (Decouchant, Holtman, 1996) and the Issue-Based Information System used in Zeno and therefore GEOMED (Gordon 1996). There is a need for research into how to match tools for organising ideas to the different stages of critical thinking. The next step is to ensure common data representations so that work assisted by one tool can be used in the next stage by problem-solvers assisted by a different tool, and to integrate these tools behind a common WWW interface.
IBIS and group editors are not the only approaches to organising ideas. It is worth setting up action research projects to explore different tools and evaluate their effects upon critical thinking. Among the many approaches worth trying are:
However, these approaches, and the ones used in Group Decision Support Systems, are designed to support logical decision making. There is also so far unmet need for tools to support irrational, lateral, creative problem-solving. Without those, CSCL stops between stages 2 and 4, as the learners use superior (as measured by our content analysis) face-to-face techniques for problem exploration.
It is now time to deliberately design groupware to support critical thinking processes. We need tools to support each of Garrison's stages of critical thinking. And these need to be integrated behind a common interface, so that groups of learners can concentrate on organising their ideas rather than how to use the software.
In this design task, it is the social human-human interface that matters, in how problem-solvers organise each others' ideas, rather than the human-computer interface or computer-computer transport protocols such as nntp versus http. So the designers need to study theories of what improves the quality of particular types of work, such as learning, and the results of experiments based on these theories.
To paraphrase Bill Clinton's slogan:
It's the humans, stupid!
Send any comments to David Newman
