Dirkx & SmithCase study, group work online

Dirkx, J.M. & Smith, R.O. (2004) p132-159

Thinking out of a bowl of spaghetti: Learning to Learn in online collaborative groups

Chapter 6 in Roberts, T.S. Online collaborative learning: theory and Practice

Students need orientation in how to do collaborative learning, and the value of it, and how to do it online.

Cooperative v collaborative learning

Coop, p137 – ‘learning is regarded as an individual effort that is facilitated by participation in the group’

Collaborative, p 137 ‘In collaborative learning, students call into question, through self-reflective & self critique, assumptions, values, beliefs, symbols and rules of conduct that characterise their existing way of meaning making’  teachers de-authorize themselves p142 ‘ group members are encouraged to assume responsibility for their own structure and direction’

Methods

Case study approach

26 students assigned to small groups. Well defined task.  Allowed to meet f-f only one group did so.  Orientation for 2 weeks, group work for 16 weeks.  Experienced tutor available. 

Source of data.  Interviews, student journals, discussion and chat transcripts.

Findings

‘the students in our case described powerful emotional responses to what was occurruing with their small groups’

p144

two orientations to authrority emerge, one remains with the tutor the other to the group

p145

‘the gradual development of interdependence and intersubjectivity requires the cultivation of a culture of intimacy within the group. Intimacy involves an ability to open oneself up to another, as well as being receptive to otherness’

p148

quote from a student about feeling that you are meeting all the time ‘ you are never not meeting with your group. You turn on the computer and the group is there. In the form of email messages and postings to the discussion board’

Conclusions about CSCL

Subjectivity and intersubjectivity p151 ‘ need for group members to reconstruct a sense of identity that is grounded in the intersubjective realities of their group contexts’

Associated emotional dynamics

Have to reach concensus and this is not always easy

P155 ‘ Implementation of collaborative learning ….., will not be successful unless the broader contextual and systematic issues are also attended to’

p155  there is 'a psychological resistance to the need to move from a subjective to a more intersubjective form of identity'

Groups Organisational and learning perspectives

Piet Van den Bossche, Wim H. Gijselaers, Mien Segers, Paul A. Kirschner (2006)

Social and Cognitive Factors driving teamwork in collaborative Learning environments: Team learning Beliefs and Behaviours

Small Group Research (2006) 37:490-521

P492 ‘In the past, research on collaborative learning has particularly focused on the structural conditions leading to better outcomes eg group size, nature of the task ( Dillenbouurg, Baker, Blaye & O’Malley (1996), Webb and Palincsar, 1996)’

This research P490 ‘ looks for the discourse practices managing the co-construction of mutually shared cognition and reveals conditions in the interpersonal context that contribute to engagement in these knowledge- building practices’

This research argues that the Co-construction requires integration of different perspectives ( cognition) and practice ( group working), on social as well as cognitive P491 ‘ , although these two perspectives on collaboration are profoundly intertwined, most research focuses on one of them ( Krejins, Kirschner and Jochems, 2003). In order to integrate these two perspectives P492 ‘Barron(2003) points out that this entails an articulation of how characteristics of the interaction ( discourse practices ) interact with knowledge –building processes that lead to mutually shared cognition’ P493 ‘groups have to deal with what Barron calls both a relational space and a content space, which compete for limited attention’ that ‘ competiveness and friendships can hinder or stimulate groups , respectively, ‘

The model that will be built p494 ‘ presents an integrative perspective, building on the strength of different research strands. It includes both discourse practices that manage the co-construction of mutually shared cognition and conditions in the interpersonal context that contribute to engagement in these knowledge-building practices’ It also intergrates ideas about team working from organisation with collaborative learning research.

P494-495 “collaboration is defined as the process of building and maintaining a shared conception of a problem or task, distributing responsibility across members of the group, sharing expertise, and mutually constructing and negotiating cognition ( Roschelle, 1992).’ ‘ the interaction between members of the group and the characteristics of their discourse is considered the process’ by which this is achieved. Negotiation ( Dillenbourgh) , the process of achieving agreement, is a key element in collaborative learning with ‘real’ agreement depends on joint understanding.

The cognitive view

In a social learning situation on member will offer up their understanding ( Stahl) which can act as a trigger for others to do likewise. In this way an individual will develop ( construct) their own meaning by adapting and accommodating the views of others. (KRO Triggers and trigger clues can be built into the course design). This can evolve into co-construction by refining, building on and modifying the original offer ( trigger) in some way.

The idea of constructive conflict as a vehicle to enhance (?stimulate) co-construction

The authors argue that mutual understanding is not enough that the understanding p496 ‘ must also be accepted before they form the basis for action’ in fact coordinated action. In this way it becomes an agreed upon interpretation. ‘ So disagreement or divergence in itself seems to be less important than the fact that it generates communication among peer members ( Dillenbourgh et al, 1996). The team will only benefit if the divergence in meaning leads to further negotiation.

However, there can be times when a negotiation is seen as ‘a personal, emotional rejection and as such it can interfere with productive team behaviour ( De Dreu & Weingart, 2003)’

For the model the cognitive is identified as

Team learning behaviour

1. Construction

2. Constructive criticism

3. Co-construction

This aspect of the model embraces the ides of p504 ‘conversational actions enabling team members to become partners in the construction of shared knowledge ( Roschelle, 1992).

The social view

P497 ‘ the identification of the social conditions under which teams make this effort to reach shared knowledge is an essential prerequisite for developing enhanced understanding of successful collaboration’

The group needs to evolve as a social system. It is ‘believed’ that the interpersonal context , the social system , form the context that stimulates or inhibits learning behaviour.

Cohen & Bailey (1997) identified 4 group –level beliefs p498 ‘ that potentially affect the learning behaviour in teams’

For the model the social is identified as

Beliefs about interpersonal context

1. Psychological safety: p499 ‘ it alleviates excessive concern about others’ reaction to actions that have the potential for embarrassment or threat, which learning behaviours often have’ ( Edmonson, 1999)

2. Cohesion: Mullen & Copper (1994) distinguished between task and social cohesion. P500 ‘Teams that perform well are committed to successful task performance and regulate their behaviour to that end’ ‘high task motivation shows the existence of shared goals and the motivation to strive for it. It regulates the (learning) behaviour that fosters the achievement of these goals’ Chang and Bordia, 2001, ‘ social cohesion is a predictor of team viability’ ‘The relation of social cohesion with learning behaviour is more complex. On the one hand, it promotes learning behaviour because of the willingness to help each other, whereas on the other hand, high social cohesion could lead to an uncritical acceptance of solutions’

3. Task Interdependence: P500 ‘ Studies have shown that task interdependence leads to more communication, helping, and information sharing than individualistic tasks

4. Group Potency: ‘ the collective belief of group members that the group can be effective’ ( Shea & Guzzo, 1987a, p.26) p501 ‘ that positive evaluations of the team’s potency are expected to have positive effects on collective motivation and performance’. ‘The sense of confidence generated by high levels of potency is believed to help teams to persevere in the face of adversity ( Gully et al, 2002).

The model: p502 ‘ Beliefs about interpersonal context shape the willingness to engage in learning behaviour’

Method

2 first year undergraduate cohorts. 7 weeks f-f. One cohort self selected group members, other allocated . Questionnaire ( Team Learning beliefs and behaviors questionnaire) administered during the last week of the course. ( ‘Please indicate to what extent you agree with the following statements concerning the team in which you are working and the task with which you are dealing’ 75 teams. Assessment of psychometric properties by PCA ( varimax rotation with eigenvalues of 1.0 or above; reliability using Cronbach’s alpha)

Measuring instrument

Items adapted from research that matched the conceptual framework of the current research. Used PCA with optimal rotation and Cronbach’s alpha to test the psychometrics alongside an independently conducted theoretical critique of the individual questionnaire items. The questionnaire items can be grouped according to the components of the model under test.

1. Beliefs about learning context - psychological safety, task interdependence, cohesion ( group and task) and group potency

2. Learning behavior, items about construction, co-construction and constructive conflict)

3. Student perception of mutual understanding

4. Team effectiveness ( Hackman, 1989) 3 dimensions, performance, viability, learning

Individual measures from team members was aggregated!!!!!! P507 ‘ Within –group agreement was assessed using a multiple – item estimator’ Authors argue that individual responses within the team are examples of repeated measures.

1. The social side of collaborative learning – Beliefs about the interpersonal context

As with the cognitive dimension, survey items were selected as appropriate from a range of other research work. Initial FA of the entire set led to a number of items being dropped from the survey .

2. The Cognitive side of collaborative learning – Team learning behaviours in the model

9 questionnaire items selected from other questionnaires as best suited to obtaining student responses about construction, constructive conflict and co-construction.

FA led to one factor ie these items tap into one factor that could be termed team learning behaviour.

3. Student perception of mutually shared cognition

2 items on the questionnaire ( once again based on research)

one addresses outcome ‘ At this moment the team has a common understanding of the task we have to handle’

the other addresses process ‘ at this moment the team has a common understanding of how to deal with the task’

FA gave 1 factor loading with good results for Cronbach’s alpha.

4. Student beliefs about team effectiveness

Hackman (1989) defined 3 dimensions of team effectiveness;

Performance ( 1 item measured the product another measured the process)

Viability ( ‘ I would want to work with this team in the future’,

Learning (‘As a team we have learned a lot’ )

FA gave 1 factor loading with good results for Cronbach’s alpha.

Analysis of the model : using multiple regression and path analysis

1. Does collaborative learning build mutually shared cognition? (2-3) multiple regression analysis

2. Do identified beliefs about interpersonal context influence team learning behaviour ( 1-2) multiple regression analysis

3. Does mutually shared cognition predict team effectiveness? (3-4) stages 1-3 tested as a model using path analysis(LISREL)

4. Does mutually shared cognition mediate the relation between team learning behaviour and team effectiveness? Stages 1-4 tested as a model using path analysis(LISREL)

Acceptable goodness of fit on page 508

3 regression analyses (table 2)

Does team learning behaviour significantly predict the mutually shared cognition as reported by the team (2-3)(YES)

Does the mutually shared cognition predict team effectiveness? (3-4)(YES)

Does mutually shared cognition mediate the relationship between team learning behavior and team effectivesness (2-3-4) (PARTIALLY)

P511 ‘The originally hypothesized model is composed of paths leading from the four constructs measuring beliefs toward team learning behaviour and a path from learning behaviour toward mutually shared cognition. ‘

‘Inspection of the modification indices ( Sorbom, 1989) suggests one additional path between the task cohesion and mutually shared cognition.’

The relation between the team learning behaviours (2) and team effectiveness(4) in the first analysis is only partially mediated by mutually shared cognition.(3)

‘This model is most important and interesting if one looks from an educational point of view because in this model one can see the factors that are influencing the cognitive outcomes of team learning; the mutually shared cognition that is built through participating in the team learning activities. This is, in other words, the cognitive , the cognitive residue or the conceptual development resulting from team learning.’

A fourth multiple regression then tested i.e. bring in the social

Do the identified team beliefs (psych safety, group cohesion , group potency , interdependence) of the interpersonal context predict the occurrence of team learning behaviour? (1-2) note cohesion was split to two variables ( task & social)

Based on a multiple regression analysis, four of the five identified team beliefs, interdependence, task cohesion, group potency and psychological safety significantly and substantially predict team learning behaviour – Social cohesion was the exception and was therefore excluded from further analysis.

Further analysis supported the view that these four beliefs about the interpersonal form the context in which teams are motivated to display the crucial learning behaviour.

Testing the full model – 2 steps, step 1 does not include team effectiveness

To recap of stage 1 – testing the model stages 1-3

Taking the original model (1-2-3) and excluding team effectiveness (4) and looking at the goodness of fit indicators there is room for improvement . p511 ‘Inspection of the modification indices (Sorbom, 1989) suggests one additional path between task cohesion and mutually shared cognition’ as already described above, ( ie a path from task cohesion that bypasses group learning behaviours). Therefore the model, stages, 1-3 , is confirmed and also an additional element is indicated, p512 ‘ the shared commitment toward the task seems to have effects on mutually shared cognition (3) that are not grasped by the learning behaviours alone’

The next step included team effectiveness(4). P512 ‘ However the fit indices show that this model is not probable ………. Inspection of the modification indices ( Sorbom, 1989) showed that two additional paths are necessary; one path from task cohesion towards team effectiveness and one path from group potency to team effectiveness. P513 ‘ Both these adjustments seem theoretically acceptable. …… a high shared commitment to the task and a high group potency of the team will probably show itself in other team behaviour leading to effectiveness that is not fully grasped by the identified team learning behaviours’

	1

Note coeficients not shown – nor are the interrelationships between the 4 team learning beliefs

Conclusion

P514 ‘ team learning does not occur by just putting people together’ .

p515 ‘The identified aspects such as interdependence, task cohesion, psychological safety, and group potency , turned out to be crucial for the engagement in team learning behaviour in teams, which in turn gives rise to mutually shared cognition, in turn leading to higher perceived team effectiveness. The results of this research show that constructs and insights from organizational science concerning beliefs about the interpersonal context in teams are transferable to collaborative learning in educational settings’

‘task commitment supplemented with shared responsibility, drives people to collective learning processes. Furthermore, a high group potency belief strenghthens the idea that the investment will pay off and so encourages the process of learning’

‘students and professionals need (to learn how) to cope with these beliefs about the interpersonal context and processes’

‘This research sheds light on the cognitive demands of teams in dealing with framing of the task or problem at hand; …….. that room for construction, co-construction , and constructive conflict need to be made in the process of reaching mutual shared cognition. This can involve slowing down the interaction’

‘conflicts need to be seen as windows of opportunity instead of threats to progress. By taking them as conflicts around the interpretation of a problem, they can be the motor of further communication ( Dillenbourg et al., 1996) ‘

de Laat & Lally (2003)

Maarten de Laat & Vic Lally (2003)

Complexity, theory and praxis: Researching collaborative learning and tutoring processes in a networked learning community

Instructional Science 31: 7-39,

P29 ‘ a kind of exploratory coversation between theory and praxis that may be mediated by methodology’

Theory

P8 Stenhouse 1983 ‘ argued that the development of theoretical understanding of educational action and doing educational research into the practical problems of education are inseparable.’

On complexity the authors argue that ‘no single theoretical framework is yet capable of offering us a sufficiently powerful articulation of description, rhetoric, inference or application’ ( the 4 ways in which theory can contribute). Theories differ in focus and some are, to an extent, complimentary.

Praxis ( teaching and learning - of both the teacher and the learner is also complex)

Methodology

Then there is the question of methodology. Theory influences what we focus on but does not tell us how. P10 ‘ Methodology assists the conversation between theory and praxis by providing rules for their interaction. However, as a community of researchers, we are still confronted with the methodological challenge of agreeing the rules’

Comment on Content analysis

· Can be cumbersome and time consuming

· Choice of coding strategy

· Can be issues of reliability and validity due to subjectivity when interpreting codes

· Multiplicity of meaning in one unit of analysis

· P11 ‘ more fundamentally, the available theoretical framework may not be sufficiently robust to enable valid inferences to be made about any of these processes from the textual traces’

· ‘Furthermore, what does one do about those aspects of learning that are

not expressed in, and therefore not amenable to, content analysis’

get Hara, Bonk and Angeli (2000) Content analysis of online discussion in an applied educational psychology course, Instructional Science 28(2) 115-152

p18 there is an important question when deciding on methodology p18 ‘ it could be argued that a more grounded approach, ( compared to content analysis) using categories that emerge from a reading of the messages, would provide a more ‘authentic’ summary of the intentions of the participants. In outr view this is a valid and important way of approaching the analysis. However we wanted to connect with some of the conceptual and theoretical ideas about learning and tutoring in the literature using schemas that were already in use’ It also means that results can be more easily shared and compared with other researchers.

Theoretical frameworks (of the study)

1. Social constructivist

Therefore search for evidence that learners are linking new knowledge to their prior knowledge to actively construct new internal representations.

2. the situativity of learning and meaningfulness

p13 ‘ learning as a set of processes by which a learner personalizes new ideas by giving meaning to them, based upon earlier experiences. However meaning is also rooted in, and indexed by experience ( Brown, Collins & Duguid, 1989). Therefore, each experience with an idea, and the environment of which that idea is part, becomes part of the meaning of that idea ( Duffy and Jonassen, 1992). Learning is therefore viewed and considered by us as situated by the activity in which it takes place ( Brown et al., 1989; Lave and Wenger, 1991). This view has led us to also seek evidence , in the online texts, for the cognitive, social, and affective processes in which participants engage in trying to make meaning of the ideas presented to them by the tasks they are undertaking’ We have also used Critical Event Recall (CER) to try to access the meaning making, and awareness of context, that participants use to make judgements and engage in activities in their course of study’

3. group dimensions of learning

( Lave, 1996, Levine, Resnick & Higgins, 1996, Wegerith, Mercer & Dawes, 1999) – the role of the group in shaping and driving individual cognition. There are three aspects

· the group situation p13 the situation itself may exert a strong mediating effect on individual cognitive and conceptual processes; the thinking of individuals is influenced by the group in which they are working’

· merger of intellectual and social processes

· ‘tension between the conceptual structure or understanding ( of the problem or ideas under discussion) of the group of individuals within it. These individual understanding may vary from each other as well as from the group. This tension may be the driving force for the collective processing of the group. So, for example when an individual member of the group expresses his or her opinion in relation to the shared public understanding of the group, this may be based on an attempt to synthesise the understanding with the public ( that is group or shared) one. The other members of the group might compare this new synthesis with their own understandings of the group-accepted version and their own disagreements with it. Depending on the outcome of this process there may be further interaction and negotiation until a new meaning or understanding is accepted by the group. ’

p14 ‘ In this way interaction between individuals, as well as their shared and individual cognitions, can be viewed as key aspects of the co-construction of knowledge, meaning and understanding

4. socio-cultural

p14 ‘ whereas the social-constructivist perspectives makes a distinction between the individual cognitive activities and the environment in which the individual is present, the socio-cultural perspective regards the individual as part of that environment’ According to this view , learning cannot be understood as a process that is solely in the mind of the learner (Van der Linden & Kanselaar, 2000). Learning is a process of participating in cultural practices, a process that structures and shapes cognitive activity ( Lave & Wenger, 1991)’ Therefore negotiation is an important aspect p14 ‘ additionally , they can motivate each other and help each other to finish the task’

Research approach

‘attempt to reveal and understand the richness of processes beyond the capability of any one of the methods, when used by itself’

p12 ‘ it is the interaction between teaching and learning that is of central importance in all educational endeavours. Therefore one of our key aims is to enquire systematically into this key educational interaction’ ‘Learning and tutoring, as ongoing sets of processes, happening in time and space, within an individual or a group, do not feature in detail in this general analysis’

‘ a parallel focus on group and individual processes’

Method

Content analysis of individual ( student and tutor) contributions and differences in learning and tutoring processes. (KRO but note that tutor contributions do not include any social or affective)

Integrated with the use of Critical Event Recall (CER) to probe learning and teaching that may not be expressed in the text records used for the data analysis

Participants

Masters in E-Learning based on an action research approach to professional development. A sophisticated group of professionals who communicate outside the formal Web CT environment. Many are mature learners who bring more than one relevant body of expertise to the course’

Focus on one group, 7 students and 1 tutor ( does not say how selected) engaged in a 10 week course.

RQ

What is the relationship between knowledge construction (learning) and tutoring processes as these evolved over time within the workshop?

Content analysis

Used NVivo p16 ‘ to search, by individual participant, for his or her contributions to each category of the learning and tutoring coding schemes’

Approx 1000 messages , sampled 10% from three phases, beginning, middle, and end.

Unit of analysis – p17 ‘ using semantic features such as ideas, argument chains, topics of discussion, or by regulative activities such as making a plan or explaining unclear information. ( Chi, M.T.H. (1997) Quantifying qualitative analyses of verbal data: A practical guide. Journal of the Learning Sciences 6, 271-313.). Thus, the content of the messages had to be read for meaning to determine segment boundaries.’ i.e. used semantic rather than syntactic boundaries.

Learning ( content analysis of student text)

Used code ( 23 items) developed by Veldhuis-Diermanse for a PhD thesis (unpublished)

Cognitive

Debating (5 items)

Using external information and experiences(4 items)

Linking or repeating internal information ( 2 items )

Affective

General: reacting emotionally to notes of peers without recting directly to the content of that note. This reaction can be +,_- or neutral

Asking for (general) feedback, responses or opinions by fellow-students

‘Chatting’ or ‘social talks ‘ contributions that are not relevant to solve the case/task

metacognitive ( some of these include social processes)

planning ( 3 items)

keeping clarity ( 3 items)

monitoring (2 activities)

rest activities ( units that cannot be decoded by using the above categories

Teaching ( 19 indicators) from ( Anderson, T., Rourke, L., Garrison, D.R. & Archer, W. (2000). Assessing Teaching Presence in a Computer Conference Context. http:///www.atl.ualberta.ca/cmc/publications.html)

Instructional design and organization ( 6 indicators)

Facilitating discourse (6 indicators)

Direct instruction (7 indicators)

Critical Event Recall p32 ‘has the potential to access asepects of learning and tutoring that are not directly available in the discussion transcripts’

P18 ‘ the recall enables the articulation of many previously unexpressed aspects of learning’

Trigger for tutor recall was

1. Summary based on the content analysis of the student interactions – ‘ an overview of the patterns of learning’

2. Full summary ( therefore tutor able to select what were the critical events for him)

Results

Content analysis

Learning (students)

Descriptive , presented by items and individuals in tabular form.

	beginning	mid	end
Cognitive	60%	81%	42%
metacognitive	20%	14%	20%
affective	14%	3%	8.9%
miscellaneous		1.9&	28.8%

Tutor/teaching

Instructional design codes at high level initially (51%), then 22%,33%

Facilitation initially (41%) then rises to (75%), 64%.

Some of the students make contributions that are rated as high for tutor processes, including two student who maintains this instructional mode in the mid phase whereas for another this declines during the mid and final stages.

Critical Incident Recall

Tutor agenda

· Concern with the learning process

· Constant awareness of the dynamics of the group

In particular

Containing Charles who is mostly thinking about getting task completed rather than the process of doing the task

Andrea also strong but supportive, also trying to hold Charles, so tutor tried to support her in doing this but staying neutral and keeping a small footprint

‘trying to avoid them making any decisions before the rest of the group appeared’

at one point offered up two ideas as catalysts.

During the middle phase p26 ‘conscious of sitting back’,’but I was tracking it, like a hawk really, but I was in a more relaxed mode’

Then started to ‘focus more on the dynamics within the group’

‘sense of two groups forming’ two, maybe three cracking on whereas ‘ the others were more shadowy figures’

Kate struggled the most and seemed able to find a niche , she articulated her struggle, ‘Andrea responded academically’…..’ whereas Pauline gave a more emotional response’ and ‘seemed to respond to her more as a person’

Articulates some assumptions about members of the group – e.g. Bill new to this form of learning ( note from learning analysis Bill’s participation level increased as the course progressed ? as the purpose of the exercise clarified for him.

Kate – difficult personal circumstances – p28 ‘her model of the experience to come was possibly of a more directive nature, so it was a struggle to accept a new mind set of knowledge being distributed amongst themselves’

Felicity ‘ she felt more comfortable thinking about the task than providing tutoring support for others’ – ‘ she was wuite a strong person but did not dominate’

Conclusions

P29 the tutor was actively engaged in a social process of actively designing his involvement through interaction, with a view to actively enhancing the learning of the group members. He was concerned ……. To maintain balance and integration within the group’

‘the complexity of practice in networked collaborative learning envionments is such that models of social-constructivism, situated learning and socio-cultural theory are not, separately capable of providing an account of the role of learning in meaning making, the function of context or the power of the interaction between tutoring and learning processes’

‘results show …. How professionals collaborate successfully to develop their own practice