Last changed 11 Aug 1997 ............... Length about 4,500 words (28,000 bytes).
This is a WWW document by Steve Draper, installed at http://www.psy.gla.ac.uk/~steve/ECSCW97.html. You may copy it. How to refer to it.

CSCW for education

Contents (click to jump to a section)


by
Stephen W. Draper
Department of Psychology
University of Glasgow
Glasgow G12 8QQ U.K.
email: steve@psy.gla.ac.uk
WWW URL: http://www.psy.gla.ac.uk/~steve

Preface

This is my pre-circulated statement for the workshop "CSCW for learning: physician heal thyself?" at ECSCW'97, Lancaster, 7 Sept. 1997.

Introduction

The point of this workshop is to have some discussion across the different backgrounds that people may come from when they move into applying CSCW in education. I shall therefore begin by stating my background, and hence the attitudes and perhaps bias that you may perceive in me. I will next state my own goals for what I hope to gain from the workshop. I then offer sections on a number of subtopics, any of which I would be happy to have discussed in the workshop. These topics are essentially my answers to the question "what are the various things I know or think about applying CSCW to education?".

My background

With Julian Newman and many others, I am part of the MANTCHI project. This is looking at collaborative teaching, particularly tutorials as opposed to primary exposition, between institutions over the internet in the subject area of HCI. The theoretical idea behind the project is to use Answer Garden type software to support "vicarious learning" or "learning by onlooking": that is, the idea that students, like lurkers on an email discussion group, might benefit from being a silent audience to others' learning dialogues. My own current list of the features of the MANTCHI project is:

You can find more on all these things through links on my web page.

Personal aims for the workshop

My own personal "learning objectives" for this workshop are to learn something about other perspectives on CSCW, and to discuss any or all of the issues I mention here. Please note that what I write here are things on my mind: opinions I am interested in debugging. I have not cleared them all with fellow workers, who may disagree with them and are certainly not responsible for generating or defending them. In fact Julian Newman may well disagree with them, as well as others in the MANTCHI project.

Pedagogically-led, not technologically-led, design

In reviewing the applications of CAL I have seen, I have come to the opinion that the few big successes, where large gains have been achieved, have been where the design was driven by educational needs, not by technology. In the usual case, designers took their brief to be "replace human teaching by technology", and the result is what you would expect from that: use of technology, adequate teaching quality, and increased expense. If you want increased quality, then the focus must all the time be on how to improve the quality. It is apparently just not true that applying technology brings improved learning as an automatic side-effect. To get improved learning, you have to design an improved learning and teaching process.

I argue this at greater length for CAL in Draper (in press). I would expect it to apply to CSCW: technology-led applications will not bring much educational benefit, only extra costs.

NetSem

One of the applications at Glasgow that our group evaluated was a type of CSCW: the introduction of email seminars in the Music department. It was, on the whole, one of the big successes, which I believe is because it was introduced to fix a serious existing problem with the seminars. However this equally means that we couldn't just apply it elsewhere (in other departments or courses) and expect the same big gains. Note too, that although the technology was cruical, equally important was the skill and effort of the human tutor.

Traditional seminars in the music department had been abandoned as too poor in quality. The format required one student per session to have prepared a paper which they read to the others. A discussion was supposed to follow, but in practice only the tutor made any contribution. Other students either did not attend, or said nothing. One reason for this situation was that students from different faculties and different years would be taking the same course. A first year engineer might be sitting next to a third year music or literature major, and naturally feel they were not in a position to debate. Email seminars were introduced, in which both the papers and the discussion were done by email within small groups. Six out of eight of these groups were markedly successful by various measures (number of contributions, quality of contribution, student opinions), in contrast to almost no success in the old face to face format. More details are given in Duffy et al. (1995).

This success is probably due to at least the following factors. Students could meditate and take time to formulate contributions and responses, rather than having to think and articulate them on the spot in "real time". The tutor mediating the email seminars, whose skill was a crucial factor, applied a marking scheme that rewarded all contributions with bigger rewards for better contributions (or conversely, could be thought of as penalising silence). It would not be practicable to apply this reward scheme in a face to face seminar. Thus while the technology was very simple, it could overcome crucial problems in the traditional delivery that it replaced. While to some extent these advantages would apply as an alternative to any use of face to face seminars, this department has, as noted, a particularly unpromising situation for fostering relaxed discussion: so the software solution matched this case better than it might others.

Cooperative learning: theories of what works

As you will have gathered, I don't think technology is the key issue. Human-human interaction can promote learning, technology may only help secondarily where it helps human interaction. Thus success in CSCW, especially in promoting learning, will depend on whether the humans organise the interaction well. I think NetSem is an example, where it crucially depended upon the skills of the tutor to prompt and organise the interaction. Simply telling students to discuss does not usually work.

As I understand it, although it is moderately helpful for learning (particularly learning facts rather than complex concepts) to listen to someone, more learning occurs when you talk rather than listen: the effort of re-organising your thoughts in order to speak or write causes learning. Thus a powerful technique in cooperative learning is teachback, where one learner tries to teach another. The most learning (although also the most effort) occurs for the "teacher" rather than the "pupil". Thus writing this section will probably be of most use to me, the writer, although of course I hope it will be of some use to some to you, the readers.

However it needs skillful (human) organisation for this to occur for several reasons: firstly many learners do not believe that they will learn most from trying to tell others, but instead have a naive theory that listening will cause them to learn. Secondly, it is important to have a willing audience. Thus I have not written these words before, because I have not had this audience before: it took ECSCW97 to provide the occasion for us to find each other, and I looked at most of your position papers before adding this section which I think may connect with some of your interests. For students, they need to know that other students will be a willing audience (which they often won't be without teaching intervention, because other students believe they can only learn by listening to teachers). The skills of a good discussion facilitator are to get students to contribute, partly by convincing them it is worth their effort, partly by making them feel they have a willing audience for their contributions. Neither usually happens spontaneously.

Almost always people can hear or read and understand a lot faster than they can organise and utter sentences. It is thus quite surprising that it is worthwhile for people to discuss things face to face, as most of the time is wasted with several people waiting for the speaker to produce sentences; although perhaps that is why most people find that less effort than reading. This of course provides one of the great advantages of technology: asynchronous communication, where receivers do not have to waste time waiting for the slow production of contributions but can do other things, or produce their own messages. Thus asynchronous communication is an advantage in overcoming one great natural problem (difference in generation and reception speeds of human language). However it makes worse the other issue of persuading learners they have an interested audience. This already requires special skills in a human tutor, but most of these face to face skills do not work with asynchronous communication. In fact, it seems likely they do not even work over synchronous video links. In email (or BCSCW or whatever) we need the equivalent of lots of looking the speaker in the eye to show interest, nodding, saying "Uh huh", etc.etc.; we need very fast ways of transmitting a message to say "yes, that's interesting, what do the rest of you think?", which with email not only takes far longer to send than it does to say it, but then clutters up the list of messages where it is clearly much less important than the messages with real topic content.

An evaluation of video conferencing: whose values matter?

As part of the MANTCHI project, a small trial that I witnessed was done of using Superjanet video conferencing between two groups of students at different universities. Both groups did this as part of an HCI course that covered CSCW as a topic. Interviews with both the students and the teachers concerned (one of whom was Julian Newman) showed that the participants regarded that session as a disaster. My own view is different. Since the primary aim was to give the students some actual experience of CSCW, it seemed to me that the session was essentially like a lab. exercise rather than a tutorial where they should learn conceptual material. It seemed to me that the ways in which the session didn't go well were useful demonstrations of the problems and limitations with that type of video conferencing. (For instance, there wasn't a very well defined agenda or a well-chosen concrete task. The video equipment wasn't easy enough to operate, and furthermore was different at the two ends so the participants couldn't help each other: all the money spent on hardware, but no-one on hand to help, and no warning for users that they had to have a previous session to train themselves. The monitor screens were probably too small: large by domestic standards, but given the room layout, the images of the remote heads made them in effect seem very distant compared to those physically present.)

As an evaluator, this divergence of opinion is the first good example I have experienced, where it seemed important to report on 3 different views: learners, teachers, and evaluators. Thinking about it, I can also think of a few other occasions that felt bad, yet resulted in learning gains.

CSCW for teachers

In MANTCHI we have recently decided to explore a mode of teaching collaboration in which small units of teaching are supplied by a teacher at another institution. There will probably be many variants of this, from personal attendance on site, to posting materials (by paper mail would probably often be as effective as electronically, whether for handouts or videotapes). However one of the aspects of this is how enjoyable it is to hear about other people's teaching approaches: not just the content, but the particular exercises and activities that they have worked up successfully. Should this take off in a big way in future, we would end up giving courses in which we delivered other people's material a lot (even more than we now do if we use a textbook), but provided just one or two units ourselves. The advantage for the recipients seems clear, but the advantage for the authors (besides being paid in kind through a barter system) would be to see a unit delivered several times in one year, which will obviously allow it to be debugged much faster. This seems to be a kind of collaborative work among teachers, and the feedback is likely to be done over the internet (i.e. computer supported) and to be vital in that debugging of the materials.

Learning by onlooking

If learners could learn without participation, just from observing (or viewing records of) others' interactions, then CSCW would have a whole new importance for education. Here are some notes on this idea.

Terry Mayes calls it "vicarious learning", and discussed it in the MANTCHI grant application and in related projects. What he is referring to is the possibility that people can learn by witnessing other learners' dialogues with teachers.

There is an earlier line of work in psychology. The first use of the word "vicarious" may have been by Muenzinger (1938) who used the phrase "vicarious trial and error" to refer to animals doing thought experiments (or a kind of mental modelling) as an alternative to repeating actual actions and their consequences while learning a contingency.

The first use of "vicarious learning" may have been by Berger (1961) who used the phrase to refer to learning by observing other's actions and their consequences, unlike Muenzinger's usage. This topic (though not the term "vicarious") goes back at least to Thorndike (1898). In this usage the adjective "vicarious" is applied to a number of related terms as in Berger's own title "Incidental learning and vicarious reinforcement". This line of work went on from animal studies to social psychology studies of humans, often associated with the name of Bandura, and with related terms such as observational learning and social learning.

In such situations of learning through observing others, the onlooker might note and learn:

In practice all of these can and do occur at times. People can sometimes learn actions by observing them (a: learning by imitation), and learn contingencies (d: what rewards and punishments follow certain actions): most judicial punishment has this possibility as a main justification (deterrence: "as an example to others"). When we see people all looking in one direction, for instance up to the sky, then we too look in that direction (e). However it seems clear this is pretty limited: there are many things we can see (e.g. expert sport, craftsmanship) which we can't then reproduce. And we often see people enjoying themselves without believing we would enjoy that. Still, perhaps most of what we learn is that there is something to try out (or avoid) for ourselves; but exactly what it is (what we should do) and whether we ourselves would enjoy (or hate) it has to wait for that personal experimentation. In other words, in these cases of social and observational learning it seems likely that it is attention that is the main direct learning content: other learning may then later follow as a consequence if the change in attention leads to other learning actions. Thus, as some of the social psychology work suggests, vicarious learning as an instructional method suffers from the fact that what is learned by onlookers depends upon them and what they attend to, which is often not what the instructor intended. (This is consistent with a reasonable interpretation of constructivism: that other people are very important for directing our learning by directing our attention, yet that is only the first step, and personal action and experience must then follow for learning to occur.)

Higher education

Higher education presents a different situation, and a different application of the term "vicarious". In this, as in much else, HE seems a very different kind of learning, and learning by observing the learning of others plays a different role, and operates differently.

In higher education, the distinctive mode is really learning by overhearing or looking on, rather than "vicarious" or "social". It is social, but so is a learner reading a book in solitude. That is, the basic enterprise of HE is that of the social transmission of cultural knowledge: of social learning as opposed to discovery learning, so we can't distinguish one mode as "social". Similarly all HE learning is vicarious in that it is not primarily through personal experience: the whole enterprise of science and scholarship is about reporting other people's experience / experiments and learning from them.

Instead to distinguish a special mode where in HE learners benefit from being an audience to others' learning, we need to look at cases where the learner is an onlooker of an event concerned with learning, as opposed to playing one of the main two roles (teacher, and learner being addressed by the teacher). Examples are hearing question and answer between another learner and the teacher; observing the teacher solve a new problem; sitting in on the project supervison of another student. If one learner is witness to a dialogue between a teacher and another learner, do they benefit? This is as opposed to a) being addressed by the teacher themselves; b) speaking or writing themselves; c) peer interaction i.e. personal interaction with another learner.

Christine Howe and Andy seem to have found (in their studies of peer interaction promoting conceptual learning) that groups of 4 work as well as pairs; and also looked at the dialogue analysis and who spoke most; and found from this that listening without talking is as powerful a way to benefit from "dialogue". It must be that processing other angles/ views is the key added value.

Margaret Brown told me that a student in one of my HCI labs had in fact commented:

  1. That seeing me talking to another student encouraged them to ask me for help. So at the social level of whether they feel it is worth the risk of asking me strongly affects the level of questions actually asked.
  2. Overhearing another student's questions made them think of something they would like to ask.

My terminology

Thus there is a distinctive mode of learning to study. I shall call it "learning by onlooking" (LBO), although it might also be called learning by overhearing, by auditing, by witnessing, by being an audience, by non-participation. The essential position is being an audience to action done by others, as in conventional theatre. The audience is, if the production is successful as a theatrical event, actively engaged in their minds, but they do not participate in the material or communicative action. Calling it "learning by non-participation" would stress this, and contrast it most explicitly to theories of active learning and of personal instruction (two other popular models). The essential metaphor is of learning by being an audience. The best contrast to other theories of learning in HE is to call it learning by non-participation. I will call it learning by onlooking (LBO) for convenience.

More points on the nature of LBO in HE

2 things to note.
a) It is not essential, either in HE or in direct observation of action and contingencies, that the protagonist being observed is learning. The learning benefit for the onlooker does not depend on this, only on observing action or dialogue.
b) Learning from others' direct actions can and does still occur and be of some importance in HE, even though it is not the most important mode.

It is interesting to note the common formats for presenting learning material on TV. One type of documentary shows illustrative material while giving a direct exposition to the viewer on the sound track. Another has a presenter talking to camera (directly to the viewer), like a lecturer. Other popular formats, however, are essentially LBO in nature: the audience observes a dialogue between people (e.g. presenter and guest, interviewer and interviewee) neither of whom addresses the audience directly. Whether on news programmes, documentaries, or educational programmes, LBO is about as common and widely accepted as direct instruction.

Although the central type of LBO in HE is being an audience for a dialogue with a teacher, perhaps because of the centrality of dialogue in HE, it is interesting to ponder another case. Consider observing a chess match. Most would probably agree that witnessing chess matches can be a useful learning experience for chess players. Dialogue plays no role in chess. On the basis of superficial definitions, this would make it an example of the second type of vicarious learning. But although actions (moves) and contingencies (pieces being taken, who won) are there to be observed, that is probably not the important kind of learning. It is the thinking stimulated in the onlooker by the positions and the moves, even though that thinking is probably rather different from that of the participants. This is not "vicarious" (in the sense of learning what the actor is learning) because the onlooker may well not be looking at consequences but at the protagonists' choice of actions.

This is one prototype for the argument that onlooking may have some special advantages, benefits greater in some respects than those of participating. McKendree et al. (1997) argue that listening without participating in a dialogue may free you to do more and different processing: in other words, the social and processing pressures of participation may in some respects act to reduce learning.

Why is LBO important?

It is important to understand LBO in order to justify and optimise common practices. For instance seminars and tutorials should be 1:1 unless either the teacher is going to lecture, or in fact LBO is effective. Many students believe it is not effective, but many experiences in fact show that it is: e.g. that students benefit from sitting in on segments directed wholly at another student. If it really is effective, then we should not only use it, but demonstrate to students that it is effective to increase their knowledge of how to learn.

Summary of types of "vicarious" learning

  1. Thought experiments: further learning by using memory to imagine or recall consequences.
  2. Learning from seeing other's direct actions (action and reward or punishment).
  3. Learning from witnessing others' dialogues: getting information from a learning dialogue, rather than a learning experiment (trial and consequence), done by another.

In fact, there are (as Laurillard says) two basic kinds of learning content: learning from one's own experience, and learning public, formal descriptions. Each of these has a corresponding kind of LBO. I regard "onlooking" as a more accurate description than "vicarious", as the onlooker may in fact learn something somewhat different from the participant: whether attention in learning from action or inferences and relationships in learning from others' dialogue. Both kinds of learning, and both kinds of LBO, occur and have some importance in HE; but learning public descriptions is what makes HE different from other learning, is what humans do that animals don't, is ultimately the focus and test of learning in HE.
In education, LBO may be seen as surprising because it is learning by non-activity, non-participation; and even more surprising because there may be special advantages (e.g. don't need to have the cognitive load of doing dialogue yourself.) that allow it to offer benefits not offered by active learning.

Content types for LBO

LBO can probably apply to all types of content. It could however be particularly important for two types not often focussed on. The first is learning how much should (and should not) be understood. At times, children irritate adults by asking "why?" repeatedly in response to answers. It annoys us because we don't have answers, and feel that "sensible" people know that many things don't have reasons. But actually a learner must learn this about any new subject; and in addition, most courses contain simplifications on the grounds that the full story would not be beneficial at this stage.

The second type of content is related: it is information about the status of other information. Some things are facts in the sense that everyone agrees, others are "usually thought to be", others "one theory is" etc. In other words, every fact and theory rests on something, and knowing what those grounds are, and the degree of support from the community for that view is part of the knowledge.

References

Draper,S.W., Brown, M.I., Henderson,F.P. & McAteer,E. (1996) "Integrative evaluation: an emerging role for classroom studies of CAL" Computers and Education vol.26 no.1-3, pp.17-32 and Computer assisted learning: selected contributions from the CAL 95 symposium Kibby,M.R. & Hartley,J.R. (eds.) (Pergamon: Oxford) pp.17-32 [WWW document]. URL http://www.psy.gla.ac.uk/~steve/IE.html

Draper, S.W. (in press) "Niche-based success in CAL" Computers and Education [WWW document]. URL http://www.psy.gla.ac.uk/~steve/niche.html

Duffy,C., Arnold,S. & Henderson,F. (1995) "NetSem - electrifying undergraduate seminars" Active Learning, no.2, CTISS Publications, University of Oxford (June 1995). Reprinted in Musicus (CTI Centre for Music), Vol. 4, June 1995.

McKendree,J., Stenning,K., Mayes,J.T., Lee,J., & Cox,R.. (1997) "Why observing a dialogue may benefit learning: The vicarious learner" Proc PEG'97 [WWW document]. URL http://www.cogsci.ed.ac.uk/hcrc/wgs/graphics/vicar/VicarPapers/PEG97.RTF