Charles Crook

Loughborough University

Address: Department of Human Sciences, Loughborough University, Loughborough, Leicestershire, LE11 3TU. Email: c.k.crook@lboro.ac.uk

Running title: Deferring to resources

Undergraduate students were observed engaged in a species of collaboration rarely studied and yet which is grounded in an authentic form of normal study practice: namely, revising a course in preparation for an exam. Pairs of undergraduates were convened for recorded sessions in which they did this around either their own personal lecture notes or around a set of notes authored by the lecturer and made available as web-readable computer documents. Although the goals, motives and orientations of these pairings were similar, the nature of the collaborative resource effected the character and rhythm of the joint conversation. The computer-based documents led to less on-task collaborative talk. Moreover, these documents sustained conversation that was more fragmented around successive short topics. Observations are made regarding how certain discursive openings are more readily afforded when revision talk is mediated by a less singular, authorised, and directive form of document.

The best-engrained metaphors that have come to characterise computer use within education invoke notions of "personal tutor" (Suppes, 1966) or "personal toolkit" (Underwood & Underwood, 1990). Less well anticipated has been the inter-personal significance of this technology. In particular, its significance for supporting forms of collaborative learning. Indeed, the very topic of collaboration seems to have become much more prominent within psychology - as researchers have noticed the adoption of computers for shared activity in both classroom (Crook, 1987) and workplace (Bowers & Benford, 1991). However, research described in the present paper was prompted by two issues neglected within this new literature. First, recent studies have tended to overlook how the computer provides potential collaborators with a resource manifesting a particular format. For, as a context for joint activity, this technology has distinctive properties. Moreover, those properties may influence the form of a collaborative discussion in distinctive ways. Second, it is suggested here that research on computer-supported collaborative learning has dwelt too much on a certain standard task format – namely, circumscribed problems with well-defined goals. This has encouraged theoretical perspectives with too narrow a scope.

These two observations need expanding. To say that the computer has "distinctive properties" for joint activity is to say that its design tends to shape particular forms of coordination among collaborating users. The neglect of the computer as a distinctive context for joint activity has happened because researchers have grounded their interest in two other themes. The first is simply "collaborative effects". Here, the interest is in whether socially organised forms of learning are in general more successful than solitary arrangements. Success is usually taken to be a matter of post-intervention performance on suitable tests. The computer becomes a convenient site for supporting collaboration and designing experimental contrasts; thereby allowing researchers to compare different ways of configuring learning such that it becomes more or less social. The second significant strand of research on computer-based colloboration involves dwelling on structural features relevant to orchestrating the arrangements for the interaction, such as the size or composition of groups. In such research, outcome measures sometimes are complemented by process-oriented analyses of the joint activity itself. Typically, researchers will construct taxonomies of collaborative talk. Recently, there have been signs of interest in comparing such talk under different conditions of computer software use (Anderson, McAteer, Tolmie & Demissie, 1998). Yet it remains unusual for collaborative talk to be theorised in terms of how it might be shaped by the distinctive demands of tasks that are computer-mediated (Rochelle & Teasley, 1994; Teasley and Roschelle, 1993). This is an important issue to pursue, because often we may have choices regarding the media that might best be used as a catalyst for collaborative engagement. Such concerns suggest more focussed attention on the computer as a resource for collaborators and, thereby, revealing how it can afford or deny particular modes of social interaction.

The present study develops this perspective, although not in relation to the form of task most usually researched. The usual focus of research has been on people acting jointly in the interests of solving some well-formed problem. Accordingly, the term collaboration may often be defined in phrases such as "mutual engagement of participants in a coordinated effort to solve the problem together" (Roschelle & Teasley, 1994). Yet often we say that people are collaborating, even when it is not so clear that they can be said to be solving a problem - at least, not without allowing the notion of a problem to be very generously defined. My own preference is to modify the above definition of collaborative engagement to refer to "a coordinated effort to build common knowledge". Perhaps there is a price to be paid for shifting reference from the attractively visible "solving the problem" to the more abstract "building common knowledge". However, this is surely what collaborating most typically feels like (Schrage, 1990). Moreover, the outcome of collaborative effort may well not be an answer, or an artefact, or some similary accessible product. Often we enter these joint enterprises simply in order to consolidate, elaborate or refresh our understanding of some phenomenon of shared interest. In which case the payoff of collaborating may not be particuarly bounded and precise. It may not be closure on some problem.

Whatever that outcome/payoff, collaborators are also entitled to a self-interested perspective on it: sometimes the main feature of the product that matters will be what each individual person now feels he or she knows, rather than the communual nature of that knowledge. Nevertheless, for such individual rewards it is still the collaborating that counts. It is still the effort of building mutuality that lies behind the enrichment of an individual collaborator’s personal knowledge. The rewards will have been jointly won. Such successes hinge upon the manner in which discourse is actively organised to construct the valued outcome of common knowing. Constraints on the form of such discourse is an issue to be investigated below.

The task to be studied here is a species of activity that does seem to fall outside of what we would naturally term a "problem". However, it has considerable authenticity within contexts of learning and it is an activity increasingly likely to implicate computers. I am referring to something very familiar within student culture: that which is commonly termed "revising". Yet this is an activity rarely studied (cf. Entwhistle & Entwistle, 1992). In a broader plea for research authenticity, Geertz famously urged more "outdoors psychology" (1983, p.153). Perhaps there is a need to study "outdoors colloboration" as a counterpoint to the preoccupation of research with circumscribed and ad hoc problems. For example, in the context of higher education, increasing confidence with the content of some lecture course or other is a natural student ambition. Yet this is only a "problem" in the narrow sense of a "challenge". If anything gets achieved in students’ pursuit of this ambition, it may simply be a sense of progress, rather than a sense of solution. Researchers should nevertheless concern themselves with how students set about this challenge. It represents a potential arena for observing "outdoors collaboration".

If reviewing a course of study is to be researched as a collaborative activity, then a proper concern is with whether students actually ever do their studying in this social manner. Existing research suggests that this may be rather unusual (Crook, in press; Hounsell, 1987). The present study will determine how far there is a tradition of informal collaborative working within the university class from which the participants here are drawn. If this does turn out to be an infrequent activity, then this will not help any claim here for the authenticity of our collaborative revision observations. On the other hand, the present study is concerned with a computer-mediated version of such revision. This might render the research timely, even though the kind of collaboration being studied is currently infrequent. For new technology may become a lever on what is otherwise a failure by students to take advantage of collaborative opportunities. Yet while we may properly seek such a lever, the computer still must be shown to be a good candidate. This suggests that we compare the form taken by a joint activity - in this case collaborative revision - when mediated by a computer with the form taken when mediated by more traditional resources. Making just such a comparison is the central concern of the research reported below. Ideally, this should permit some consideration of the relative quality of collaboration and this, in turn, may help us locate this significance of new technology in this particular collaborative role.

The arguments made above have encouraged a study in which, first, there is some consideration given to baseline levels of the critical activity - students jointly reviewing some lecture course that has been studied in common. I then move to construct a comparison between such collaborations as resourced in two ways. First, where the talk is organised around the participants’ personal lecture notes. Second, where the talk is organised around the lecturer’s own notes, as marked up for reading in a web browser. Thus, students in each group have the same academic ambitions, they are studying in comparable environments, and they are orienting towards the same curriculum events. However, they are collaborating around different forms of record for these events.

An important point needs to be acknowledged about this comparison. Namely, that whatever happens among the students using the computer-based resource does not represent some unique "computer effect". This is a species of the familiar observation that media themselves do not have direct effects: it is always necessary to recognise the underlying forms of practice that are guiding the designer/author’s use of the media (Clark, 1994). In the present case, what might be distinctive about computer-mediated joint revision (compared to the alternative of one’s own notes) is that the collaborators are working from a single shared document and/or it is a document authored by a (signficant) outside authority. These and pehaps other features of the computer setting are not unique to materials designed for that technology: for their structure might be replicated within more traditional media. That is, collaborative resources that are singular and authorised might be designed for students without implicating computers. However, what legitimates characterising this collaborative arrangement in terms of the computer is that if the arrangement is to be found (or generated) in a university setting at all, then the networked computer is the most likely technology to catalyse it.

Partly because the computer-mediated collaboration is likely to be relatively unfamiliar, some care is needed in configuring it for study. In particular, observations should not be made at a point where the experience is still novel. Accordingly, it should be arranged that any analysis refers only to collaborations conducted after a period of participants becoming familiar with the materials and the conditions of observation. Thus, the analysis must follow a period where the computer-based materials have been freely available for general exploration. In addition, the field observations should be in familiar and relaxed surroundings. Appropriate motivation should also be assured. In the present case, students were reviewing a degree course that was shortly to be examined and they were doing so during a period of their course reserved for revision purposes.

Analysis of collaborations organised in these ways will most likely dwell upon talk. However, making sense of that talk will depend upon other considerations. In particular, acts of manipulating and attending to the materials themselves should be captured. This implies video recording. The dense and technical nature of the material these students are reviewing makes a rich interpretation of what they say potentially difficult for the analyst. In the present case, it was arranged that the analyst was also the tutor who delivered the course: thus ensuring a particularly intimate familiarity with the issues that were being discussed and the history of shared knowledge that the collaborators might invoke.

While this study aims to explore isssues to do with the quality of learning, it could not pursue a traditional experimental design to tackle such a question. Thus, it is hardly possible to take an authentic case of to-be-examined study and attempt to restrict the form of student revision allowed. In short, it is not possible to conduct certain sorts of experiment. Accordingly, outcomes of collaborative revisions - in terms, say, of examination results – were not developed as dependent variables. Yet how collaborative talk varies across different conditions of studying may still be useful to our thinking about outcome issues. It may allow extrapolations to likely influences on exam performance, if that seems an important goal. Here, analysis is firmly focussed on the form of study that is preciptitated by a certain kind of resource. This is the "outcome" at issue. Perhaps whether or not it correlates with examination performance depends a lot on how we then chose to design the associated exam. The present observations might even help with such decisions.

Participants and context. The study concerns a class of 49 students who had recently taken a lecture course on Social Development and were preparing for a finals examination. They were all at the end of their second year studying for a Psychology degree. Thirty-three of them completed a sociometric procedure in which they identified class peers they knew socially as well as those they had studied with privately (this is termed a "study sociogram"). Sixteen students declined to take part in this exercise (apparently for reasons of needing promptly to leave the session in which it was suggested). Separately, forty-two volunteers (8 male and 34 female) were recruited from this class to work in pairs on collaboratively revising this course early within a forthcoming timetabled revision period of serveral weeks. Typical of Psychology classes the gender balance was uneven: 31% were male. It was explained that these sessions were to be video-recorded but that the tapes would not be viewed until after graduation and they had the right to withdraw participation at any time. It was also explained that the computer-based revision resources to be studied by some volunteers would be fully available to the whole class whether they participated or not. There was some enthusiasm for what was viewed as a welcome discipline for necessary studying.

Study sociogram: materials and procedure. An exercise to determine social and collaborative practices in this group was conducted using the class photograph. This comprised 49 thumbnail head-and-shoulder pictures. Each picture was numbered for reference. At an end-of-course meeting, students were invited to contribute to a study of learning practices based on sociometric measures. Those who were willing then used the thumbnail numbers to indicate on a response sheet with whom in the class they had had various forms of social contact. In particular they were asked to questions. (a) Who they would nominate as students who were social friends. This was explained to mean those people they knew well enough to arrange social meetings outside of class. (b) Who they could nominate as students whom, at least once during the year, they had arranged to meet with the specific purpose of pursuing collaborative out-of-class study related to any part of their course.

Collaborative revision: materials and procedure. The 42 volunteers were recruited as self-nominated friendship pairs. All pairings were of the same gender. They were randomly assigned to one of three study conditions. The first involved collaboratively revising the course around their own notes. The second did so around a hypertext version of the lecturer’s notes; the third around a screen-based version of those notes, whereby each lecture was presented as a single linear document. Each format of computer-based notes comprised a set of text files readable through a web browser, with an index of the nine lectures available as a browser home page. In the case of the linear condition, selecting any lecture on the home page led to a single file in which the aims, summary, content and references for that lecture were presented as a single document to scrolled up and down. In the case of the hypertext condition, selecting a lecture led to a hierarchy of files in which aims, summary, content headings and references were available as a second level index. Finer detail of content was organised into further hypertext links from within the main content-heading sections.

Students took part in two 45-minute sessions a week apart. Small unattended camcorders recorded each session, although only the second recording was used for analysis. On session 1, the computer-based materials were encountered for the first time. Thereafter, these materials were available on a self-service basis (but not for printing) in the department computer room. All pairings worked separately in a large and familiar classroom that included numerous alcoves for computer-based work. A relaxed atmosphere was ensured: refreshments were available; other students might occasionally walk through the room (although they were requested not to distract participants). It was explained that this was a study of interactions arising during collaborative revision of a lecture course. Participants were encouraged to manage this in any way they felt useful - studying small or large sections of the course as appropriate as they themselves decided. It was suggested that the notes (their own or those on the computer) were a potential resource for reference, but it was not necessary to use them in any particular way - or indeed, to use them at all. At the end of the second session, participants completed a brief questionnaire involving 5-point scale questions asking how useful the the materials were, whether it was helpful to study them with a peer, and whether they would be likely to use them again in collaborative discussion. They also took part in a brief semi-structured conversation in which the researcher discussed their attitudes towards collaborative private study.

Sociometric data was entered into a customised computer program to produce a visual representation of social and study relationships. The session 2 video recordings were transcribed and analysed with reference to the video record. Sections of talk were delineated and categorised according to the following distinctions (1) Other: conversation concerned with topics irrelevant to the revision. (2) Task: conversation concerned with the organisation of the revision task itself, often including navigation around the revision resources. (3) Exams: conversation concerned with general strategies or circumstances of revision, often including exam expecations for this particular course. (4) Stories: biographical anecdotes loosely concerned with personal development and experience. (5) Elaboration: exploratory and evaluative conversation concerning the academic themes of the course itself.

The fourth category of "stories" deserves further comment. The course concerned social development. That is, psychological events during childhood that characterise social experience at this time and which may be relevant to the development of later social characteristics. Recounting details of personal experience or family circumstance was a striking conversational theme in some of these pairings. It seemed naturally stimulated by the content of the course domain. Where such talk was used to illustrate a psychological claim or where it was talked about such as to make explicit links to such theory, then this was regarded as "elaborative" talk. However, it was treated as "story" telling if it entailed no explicit reference to psychological content and was exercised as a form of reminiscing or anecdote telling. This is not to dismiss such talk as without potential relevance for marking a particular psychological topic. However, if placed in this category it would be merely marking (but not elaborating) the nature of a possible experience-theory relation. In short, to the listener such episodes seemed agreeable conversational distractions loosely prompted by the core material.

Sudy Sociogram. The friendship and study networks arising from the sociometric instrument delivered to the whole class are summarised in Figure 1. Squared numbers represent female students and circled numbers males. Arrows indicate the existence of either social links (left) or links involving the arranging of collaborative study (right). The diagram is presented to make two qualitative points. First, a class of this kind becomes tightly-knit as a social community. Most individuals are known to each other and, as the diagram shows, many of them are involved in each others’ social lives outside of class. Second, the practice of deliberately congregating to pursue private study is quite rare. Informal conversation also suggested that some of these would be inevitable demands of practical work carried out in class in small groups - this explains why some pairings linked as collaborators on the right are not linked as social friends on the left (completing a project write up made an out-of-class meeting between non-friends expedient rather than spontaneously invited).

-- Insert Figure 1 about here --

Content summary of collaborative talk. The total amount of talk in each pairing was partitioned into the five categories described above. The median proportion of investment in each is summarised in Figure 2 along with interquartile ranges. In all study conditions, elaborative talk about the academic topic occupied over half of what was said. However, there was more on-task talk of this kind for the students who were collaborating around their own personal notes (F (2,20) = 4.93, p < 0.05). Tukey post hoc multiple comparison tests between individual means indicated that the own-notes group differed from the linear group (p = 0.041) and from the hypertext group (p = 0.007) but the two computer groups did not differ from each other. Collaborators using computer-based materials distributed their non-elaborative conversation fairly evenly: including talk about the management of the task itself, the telling of personal stories and other off-task talk not relevant to the focus of the collaboration.

-- Insert Figure 2 about here --

Qualitative analysis of the collaborative talk. After a period of being immersed in listening to this talk, two features are strikingly apparent. Together they seem to distinguish the collaborative quality of own-notes revision from that of computer-mediated revision. First, talk around the computer appeared more disjointed in its movement between topics; whereas talk around own-notes had a stronger sense of coherence or narrative. Second, talk around the computer was often about defining and/or evaluating claims that were visible in the document the participants were attending to; talk around personal notes was also about negotiating a perspective but, in this case, one that seemed more external to the documents underpinning the talk. I shall comment a little more fully on each of these themes.

The present lectures, perhaps like most, each afforded analysis into a sequence of sub-topics – hopefully, integrating into those larger psychological themes that were implied in the lecture titles. Studying from computer-based notes, it must have felt natural to parse the record in this way. Accordingly, discussion typically took the form of taking what was said about some sub-topic and agreeing what was meant or implied in that particular section of the notes. Working from own notes may also have allowed a moving between sub-topics but such shifts were less apparent to the listener. Thus, the important point of this comparison does not so much refer to the amount of material covered (the number of lectures addressed by the average pair was similar across conditions). It refers more to rhythm than quantity: to a greater sense of discontinuities existing within the computer-based talk.

This effect that a listener becomes aware of arises from the collaborators’ overt management of topic changes. The possibility of changing topic is more often raised as an explicit proposal in the computer-based talk. At such points, a participant will suggest something like "Right, next shall we go on to deal with…" and then there usually follows a change of focus for the conversation. On average, computer-based pairs have 8.3 conversational turns (of on-task elaborative talk) before such an explicit switch is proposed; while pairs working from own notes typically pursue 33.0 turns between declared topic changes. So, it might be said that the screen notes are more oppressive – in terms of a topic organisation that they impose on collaborators. When talking from own notes, there is a greater freedom to invoke material adjacent (in the lecture story) to what is currently being discussed – without conversationally declaring the shift. These progressions are relatively seamless to the listener. Talking from physically separate and personal sets of notes seems to legitimate such moves without requiring the perpertrator to mark them as topic changes.

Put another way, it might be said that the lecturer’s notes invite a narrow conversational route through a topic – one closely-tied to the succession of points made explicit on the shared screen. Once reflection about a given point in this succession is exhausted, it is common to declare "what’s next?" and move onward accordingly. However, these are also points were it may be tempting to drift off task; the relatively larreaders of inexact textbooks or �ger amount of off-task talk among the computer-based collaborators may reflect this temptation.

Working from computer-based notes, collaborative conversation is largely unfolded through use of the scroll bar. However, what sustains activity in the case of working from own notes? And what gives that activity a coherence and direction? The answers involve three sorts of opportunity that seem to be more readily furnished within this condition for collaborating. First, collaborators own notes had a concealed quality that perhaps invited a (useful) effort of discursive elaboration. Thus, these collaborators rarely displayed their own notes to their partner in a manner that might momentarily make them a strong point of mutual attention. There may be a coyness about revealing the limitations of what one has recorded. For whatever reason, this peripheral status of the notes gave each member of a pair openings to invoke particulars from their own notes (including ambiguities and shortcomings) and this animated the conversation. At its simplest, this was provided in openings (or links) of the kind: "Hang on, what I have got here is.." However, it might also involve specific questioning aimed at clarifying something about ones own written record: "I haven’t got the difference between those two, do you know what it means?" The inherent incompleteness of the personal record and the somewhat invisible status of each participant’s document at the time of conversation, provides a useful opportunity to question and challenge.

Secondly, in part because personal notes are felt to be inherently incomplete records, these own-note collaborators were more likely to invoke the object that sat more prominently behind their activity: namely, the event of the lecture itself. Only two of the computer-based pairs made any reference to what was said or done in whichever lecture generated the notes they were looking at. Whereas all of the own-notes pairs made at least one such link, even if only to authenticate an observation: "..’cos that’s what he was saying in the lecture, wasn’t it?" Or "He dismissed that in the lecture, I’m sure he did". The point is not that these partners believe that their main purpose is to reconstruct some event that they both witnessed. Rather, they can exploit the awareness of having shared this event in order to move forward in the discussion: say, in terms of justifying a claim, or opening up a referential device to work on uncertainty, or even integrating a new coherence: "The theme of that lecture seemed to be finding out why siblings are different". This reference seemed less natural to the users of computer-based notes: perhaps it seemed less appropriate to work at collectively remembering an event that seemed to be already represented before them in comprehensive and official form.

Thirdly, the presence of personal notes seemed to afford another conversational move that rendered the collaborators more versatile. That move entailed invoking personal study experience outside of either the lecture or the present collaboration. Thus, such references were made on average only 2.3 times per pairing using computer-mediated notes. While it was made on average 10.3 times by the own-notes pairings. Several individuals using computer-based notes made reference to their partner knowing more than them. These comparative concerns may be a hint that citing knowledge from private study might be inappropriate or intimidating in a conversation that was more focussed on a self-contained learning document (the authorised notes). However, when talking from one’s own notes, this may seem a more legitimate move to make, and one that effectively sustains and develops the conversation. Opening up this private knowledge might occur through a variety of discursive moves. For example, such knowledge might be actively elicted from a partner: "Have you read about joint attention?" Or it might allow probing the authority of a summary opinion: "Hang on, where did you get that idea from?" Or it might be in the interests of defending an opposing view: "Oh no, I remember reading something different about that".

These observations suggest that jointly revising some academic course taken by collaborators may procede differently depending on the material resources that are used to facilitate the discussion. Suchman (1998) has termed such resources the "center of coordination" in joint activity. A lecturer’s own notes viewed in a web browser may be a seductive basis for coordinating collaborative revision, because of the authority such notes enjoy in relation to the perceived task. Indeed, sometimes such notes might be exactly what students do need. This may particularly be the case where their own notes are poor and they have invested little time in private study. Then the detail and structure of official notes may be very valuable. However, this better describes the resourcing of students’ first encounters with ideas – rather than their constructive revisiting of them. For many serious collaborations, the lecturer’s packaged notes may restrict opportunities for creative talk. Certainly these pairs, when using such material on a computer, were more likely to be distracted from the principle task of revising. Moreover, the style of analysis encouraged was very much one of "scrolling through the points". This generated short discussions about such points, in which their focus was defined and/or only briefly challenged. Such a topic rhythm seemed to encourage moving off task; it also afforded the learner a less active position in navigating the conceptual space of the lecture topics.

A hypertext format was constrasted with traditional linear presentation, on the expectation that where a route through the material was more under the reader’s control (hypertext), these documents might become regarded as more personalised – and thus treated more like one’s own notes. In fact, there were no evident differences in the activity supported by the two computer-based conditions. This is unsurprising, as students seemed to approach the hypertext in a very linear manner: mechanically following each link that arose and thereby creating a trail that made the experience very similar to that maintained by a the linearly designed documents. However, interpretation of this and other observations made here needs to be cautious. Claims could be restricted to the constituency studied: namely, psychology students. Yet, while research with greater scope would be welcome, the discipline of psychology does straddle science and the humanities in a way that is likely to make the present students a sound form of sample.

What do we expect from collaborative resources? They should stimulate those ideal learning conditions often described by cognitive psychologists; conditions that emphasise constructive engagement by the learner. "In general, researchers have found that people remember information that they have actively generated better than presented information and that they are better able to put such knowledge to use in novel situations" (McNamara, Kintsch, Songer & Kintsch, 1996, p. 3). This particular observation was prompted by a study in which students who were learning from totally comprehensive and coherent textbooks learned less effectively than those studying texts that were partially incomplete. A comparable point about the virtues of incompletness is made by Bjork (1994) in summarising studies concerning a useful but curious consequence of variability in the design of training situations. While such inexactness can slow down the pace of learning, it can extend the scope of what finally gets learned. Such findings seem to echo earlier observations by Lepper and Chabby (1988) who studied how school teachers managed the teaching discourse of classrooms. They compared this classroom teaching talk to the forms of dialogue typically constructed by designers of educational software. The human teachers seemed actively to avoid the precision and explicitness of the software.

Arguably, in the present study, the collaborators using their own notes were challenged in ways similar to readers of inexact textbooks or pupils listening to the oblique talk of teachers. They were challenged into greater participation. The students here had records that were incomplete: they needed to be updated and polished; they needed to be brought into line with each other. The shortcomings (not necessarily inaccuracies) in these resources effectively motivated whatever discussion was needed to evolve a shared perspective: namely, some sort of jointly-endorsed summary position regarding a taught episode.

Certain detailed features of one’s own course notes have been documented here as possibly potent for collaborators – for sustaining the flow of their exploratory talk. The improvised and (it seemed) concealed-from-each-other status of personal notes readily invited a form of revising that was catalysed by the motive of "endorsing this particular record" – seeing how far partners were working from the same platform. In addition, these notes appeared to encourage an active effort to invoke the occasion of the lecture itself. So, some of what became legitimate to debate and collectively reconstruct was "what got said". Finally, working together around such resources seemed more naturally to allow reference to the private study knowledge of the individual collaborators: that is "what I have found out". It has been proposed here that such discursive openings reflect reaction to the affordances of the shared resources. Personal notes were particularly effective in sustaining conversation and providing it with openings for acts of reconstruction and development. This allowed the realisation of that difficult form of monitoring demanded of any collaborators: "Participants must keep track of what has been established and what has been revised. The relative fluidity and fragility of common ground demands ongoing attention to the ideas and partial understandings of participants" (Barron, 2000, p. 404).

The points made above regarding the status of computer-based resources for collaborators are not points that uniquely apply to computers as a presentational medium. As acknowledged earlier, this study concerns the teaching and learning practices directing use of a medium. Indeed, the empirical observations made above might have been made about joint activity that was mediated by a single set of paper-based notes – particularly if these were also authorised by the relevant lecturer. However, there is a reason that the present results are best discussed in relation to learning that is computer-based. The reason is that this form of presentation is likely to become very visible within university communities. Elsewhere I have reported that when lecturers are given unlimited intranet web space and are encouraged to use it, then they are quick to make machine copies of their lecture notes available (Crook, in press) Moreover, interviews with students indicated that this was a very popular form of resource: something they hoped to find on their course websites. One recent research report of such a website reproduced a menu that included the item "Lecture notes (learning support material)" (Herson, Sosabowski & Lloyd, 1999). Evidently, there is terminology to hand that allows us to make these simple strategies sound sophisticated. Yet there is much to be understood about just what kind of "support" such notes can and do furnish.

All such observations alert us to the general idea that course organisers are increasingly likely to distribute lecture notes in the manner studied here. If such forms of resource become widely available for university taught courses, the question still remains as to whether students would use them in the way they were encouraged to in the present study. Would such resources become a serious opportunity for collaborative revision? It is agreed there are many good practical and theoretical reasons to cultivate more computer-mediated collaboration among students (Crook, 1994; McConnel, 1994). On the other hand, there are indications that established study practices may be resistant to such developments (Crook, 2000). Moreover, the study sociograms reported here indicate that students in this particular class did not incorporate study episodes into their social networks. Against this we must contemplate prospects that may be opened up by new technology. So, in the present study, when asked in a questionnaire whether "it was helpful to study these materials with a partner", 71% said it was helpful or very helpful. A third of them claimed that they planned to study them this way again before the examination. If collaborative study does become more commonplace, then the present findings are relevant to how we think about the necessary resources to support such joint activity.

If such studying does not become commonplace, there is still something to be learned from the findings reported here. While it may be unwise to make strong claims about individual cognition from observations of social discussion (cf. Potter & Wetherall, 1987), the present position on resourcing communual revision also may guide our policy about cultivating private revision. The non-directive and inherently incomplete nature of personal notes seem to furnish a far more provocative basis for course reflection than is provided by the more didactic genre of well-packaged lecturer’s notes. Perhaps both collaborative and private study is best stimulated and challenged by material with a greater quality of indirectness. This may need to be the case, if students are to avoid too passive a relationship with their study material – or what might be termed too great a "deference to resources".

Anderson, A., McAteer, E., Tolmie, A., and Demissie, A. (1998) The effect of software type on the quality of talk. Journal of Computer Assisted Learning, 14, 28-40.

Barron, B. (2000) Achieving coordination in collaborative problem-solving groups. The Journal of the Learning Sciences, 9, 403-436.

Bowers, J.M and Benford, S. (1991) Studies in computer-supported cooperative work: Theory, practice and design. Amsterdam: Elsevier.

Bjork, R.A. (1994). Memory and metamemory considerations in the training of human beings. In Metacognition: Knowing about knowing (ed. J. Metcalfe and A. Shimamura) pp. 185-205. Cambridge, MA: MIT Press.

Clark, R. E., (1994) Media will never influence learning, Educational Technology Research and Development, 42, 21-29.

Crook, C.K. (1987) A Social Context for Classroom Computers. In Computers cognition and development (eds. J. Rutkowska and C. Crook), pp. 35-54. Chichester: Wiley.#

Crook, C.K. (1994) Computers and the collaborative experience of learning. London: Routledge.

Crook, C.K. (2000) Motivation and the ecology of collaborative learning. In Rethinking collaborative learning, (eds. R. Joiner, K. Littleton, D. Faulkner, and D. Miell), pp. 161-178. London: Free Association Press.

Crook, (in press) The campus experience of networked learning. In Networked learning, (eds. C. Steeples, and C. Jones). Springer-Verlag.

Entwistle, A. and Entwistle, N. (1992) Experiences of understanding in revising for degree examinations. Learning and Instruction, 2, 1-22.

Geertz, C. (1983) The way we think now: Toward an ethnography of modern thought. In Local Knowledge: Further essays in interpretative anthropology, (ed. C. Geertz), pp.147-163. New York: Basic Books.

Hearson, K., Sosabowski, M.H., and Lloyd, A.W. (1999) Intranet-based learning: A one-year study of student utilisation. Journal of Computer Assisted Learning, 15, 269-278.

Hounsell, D. (1987) Essay writing and the quality of feedback. In Student learning: Research in education and cognitive psychology (eds. J. Richardson, M. Eysenck and D. Piper), pp. 109-119. Milton Keynes: SRHE & Open University Press.

Lepper, M.R. and Chabay, R.W. (1988) Socializing the intelligent tutor: Bringing empathy to computer tutors. In Learning issues for intelligent tutoring systems, (eds. H. Mandl and A. Lesgold), pp. 242-257. New York: Springer-Verlag.

McConnel, D. (1994) Implementing Computer Supported Cooperative Work, Kogan

Page

McNamara, D.S., Kintsch, E., Songer, N.B., and Kintsch, W. (1996) Are good texts always better? Interactions of text coherence, background knowledge, and levels of understanding in learning from text. Cognition and Instruction, 14, 1-43.

Potter J. & Wetherall M. (1987) Discourse and Social Psychology: Beyond Attitudes and Behaviour. London: Sage.

Roschelle, J. & Teasley, S. (1994) The construction of shared knowledge in collaborative problem solving. In Computer-supported collaborative learning (ed. C. O’Malley), pp. 69-97. Heidelberg: Springer-Verlag.

Schrage, M. (1990) Shared minds. New York: Random House

Suchmann, L. (1998) Centers of coordination: A case and some themes. In (Eds.) Discourse, tools and reasoning: Essays on situated congition. In NATO ASI Series F: Computer and Systems Sciences, 160 (eds. L. Resnick, R. Solju, C. Pontecorvo, and B. Burge) pp. 41-62. New York: Springer-Verlag.

Suppes, P. (1966) The uses of computers in education. Scientific American, 215, 207-220

Teasley, S.D. and Roschelle, J. (1993) Constructing a joint problem space: The computer as a tool for sharing knowledge. In Computers as cognitive tools (eds. S. Lajoie and S. Derry), pp. 229-258. Hillsdale, NJ: Lawrence Earlbaum Associates.

Underwood. J.D.M. and Underwood, G. (1990) Computers and Learning. Oxford: Blackwell.

Figure 1: Acquantance networks from the class under study here. The left digram shows network of contacts among people who knew each other socially. The right shows the network of contacts among those students who have arranged to meet at least once during the year for collaborative study.

Figure 2: Distribution of different types of talk under three conditions of collaborative resourcing