Stahl et al, 2006

Stahl, G., Koschmann, T., & Suthers, D. (2006)

Computer- supported. Collaborative learning: An historical perspective

In R.K.Sawyer (Ed) , Cambridge handbook of the learning sciences (pp 409- 426). Cambridge, UK

http://GerryStahl.net/cscl/CSCL_English.pdf

CSCL within education

Computers within education

p 2 ' As CSCL developed, unforeseen barriers to designing, disseminating and effectively taking advantage of innovative educational software became more and more apparent. A transformation of the whole concept of learning was required......'

E-learning at a distance

'CSCL too often conflated with e-learning, the organisation of instruction across computer networks'

Co-operative learning in groups

Dillenbourg 1999  defined the distinction between cooperative and collaborative learning roughly as follows. p3 'In cooperation, partners. Split the work, solve sub-Tasks individually and then assemble the partial results into the final output. In collaboration, partners do the work together'. i e learning cooperatively can be studied with traditional cf .CSCL methods

Collaboration and individual learning

p 3 ' The relationship between viewing collaborative learning as a group process versus an aggregation of individual change is a tension at the heart of CSCL'

' In CSCL, by contrast, learning is analysed as a group process; analysis of learning at both the individual and the group unit of analysis is necessary'

The Historical Evolution of CSCL

The beginnings

Three projects in the 90s. All shared the goal of making instruction more orientated towards meaning making. All turned to computers as resources, all three introduced some organised social activity.

From conferences to global community

Key Books p 5

From AI to collaborative support

p 6 'AI is computer software that closely mimics behaviours that might be considered intelligent if done by a human' 'Intelligent tutoring systems are a prime example of AI, because they replicate the actions of a human tutor' 'This is still an active research area within the learning sciences, but is limited to domains of knowledge where mental models can be algorithmically defined'

The role of the computer has shifted from 'providing instruction - either in the form of facts in computer-aided instruction or in the form if feedback from intelligent tutoring systems - to supporting collaboration by providing media of communication and scaffolding for productive social interaction' - scaffolding  can involve AI techniques eg by offering alternative views.

From individuals to interacting groups

The other group members no longer seen merely as a social background p7 ' the group itself has become the unit of analysis and the focus has shifted to more emergent, socially constructed, properties of the interaction'

Collaboration involves complex interacting factors therefore empirical research that investigated causal factors by controlling for other factors is generally not very successful. p7 for page 189 of Dillenbourgh et al,1996 ' Hence, empirical studies have more recently started to focus less on establishing parameters for effective collaboration and more on trying to understand the role that such variables play in mediating interaction. This shift to a more process-orientated account requires new tools for analysing and modelling interactions'

From mental representations to interactional meaning making

p 8. ' meaning making not assumed to be an expression of mental representation of the individual participants, but is an interactional achievement'

From quantitative comparisons to micro case studies

In collaborative contexts individuals visibly show their learning p 8 'the methods that people use to interact are widely shared ( at least within appropriately defined communities or cultures)' therefore case study method is popular.

p8/9 ' how can the analysis of interactional methods help guide the design of CSCL technologies and pedagogues? This question points to the complex interplay between education and computers in CSCL.

The interplay of learning and technology in CSCL

The traditional conception of learning

Thorndike - learning is measurable

Learning as a psychological phenomena with three essential features, it represents a response and recording of experience, it is change that occurs over time, it is a process that is not available to direct inspection

CSCl embraces a more situated view, p 9 'locates it in meaning negotiation carried out in the social world rather than in people 's heads'

Designing technology to support Grouo meaning making

p 9 ' No form of technology, however, no matter how cleverly designed or sophisticated, has the capacity, in and of itself, to change practice. To create the possibility of an enhanced form of practice requires more multifaceted forms of design (bringing in theories and practices from various disciplines').

p 10 ' An environment for a desired form of practice becomes so through the organised actions of its inhabitants '. i e an environment is constructed by the people who use it' and therefore in order to design we need to understand how people use these environments.

'CSCL research  has both analytic and design components. Analysis of meaning making is inductive and indifferent to reform goals. It seeks only to discover what people are doing in moment to moment interaction without prescription or assessment. Design, on the other hand, is inherently prescriptive ' therefore the relationship between praxis and design needs to be a symbiotic one.

The analysis of collaborative learning

p 11 ' The aspect of collaborative learning that is perhaps hardest to understand in detail is what may be called " practices of meaning - making in the context of joint activity " inter subjectivity learning (Suthers, 2005) or group cognition (Stahl, 2006) 'This is learning that is not merely accomplished interactionally, but is actually constituted of the interactions between participants'

Most research use codes and categories

'the knowledge building that takes place within small groups becomes "internalised by their members as individual learning and externalised in their communities as certifiable knowledge"' Stahl (2006)

The analysis of computer support

p 12 'Computer support for inter subjective meaning making  is what makes CSCL unique' ......'Design should leverage the unique opportunities provided by technology rather than replicate support for learning that could be done through other means' .......'We should explore    the potential of adaptive media as an influence on the course of subjective processes, and take advantage of its ability to prompt, analyse and selectively respond

The multidisciplinarity of CSCL

Hence the different methodologies

Quant

Experimentally based studies based in controlled factors  p13 ' do not directly analyse the accomplishment of inter subjective learning'

QualCrosby

'The grounded approach is data driven, seeking to discover patterns in the data rather than imposing theoretical categories. The analysis is often micro-analytical, examining brief episodes in great detail' '

KRo but then misses the evolving nature

Suggestion

Use data from qual to inform  quant and then quant to focus the detail work of qual ie an iterative approach.   A quisitive approach (Goldman, Crisby, &  Shea, 2004) p 14 ' a purely data driven approach that dives theory, but never applies it, won't be adequate....... 'Having identified where successful methods were not applied , we then examine the situation to determine what contingency was missing or responsible 

Davidson (1992)

Davidson, R. (1992)

Anterior cerebral asymmetry and the nature of Emotion

Brain and Cognition, 20, 125-151

Cf cognition Emotion is observable by others and often accompanied by frank biological changes that are readily perceived by the individual concerned.

Theory that motivates the research - role of cerebral hemispheres in emotional processing

p 126 ' approach and withdrawal are fundamental motivational dimensions which may be found at any level of phylogeny where behaviour itself is present.

Damage to left frontal p127 'results in behaviour an experience which might best be characterised as a deficit in approach. (KRO ? Right hemisphere modulation of emotional experience predominates)

'during the experimental arousal of withdrawal -reared emotional sates ( e.g. Fear and disgust) the right frontal and anterior temporal regions are selectively activated. In addition, subjects with baseline tonic activation in these regions show a propensity to respond with accentuated withdrawal related negative affect to appropriate emotional elicitors'

Anterior activation asymmetry (?kro as a trait) functions as a diathesis which predisposes an individual to respond with predominately positive or negative affect, given an appropriate emotion elicitor.

Methods description

p 130 time resolution

sub zero to several minutes ' many of the core phenomena of emotion are brief  and therefore require measures that have a very fast time resolution'. 'periods of peak emotional intensity are unpredictable' 'It is also important to be able to record physiological activity over much long time intervals. This is needed when subjects are presented with affect elicitors that last several minutes, requiring physiology to be integrated over the entire period of the eliciting stimulus'

Non invasive less like that PET, fMRI, to cause emotion related states such as anxiety

Can synchronise with behavioural signs

P 152 ' The principal measure extracted from the EEG in the studies I present in this article is power in the alpha band....... A wealth of evidence indicates that power in this frequency band is inversely related to activation in adults (e.g.  Shagass, 1972)....... Our measures. Of band power are computed from the output of the FFT which decomposes the brain activity into its underlying sine wave components ( KRO why not cosine as well).

From another Davidson article and based on   Schaffer, C.E., Davidson, R.J., and Saron, C. (1983) Frontal and parietal asymmetries in depressed and non-depressed subjects, Biological Psychiatry, 18, 753-762

Resting EEG was recorded for both eyes open and eyes closed for 30 sec periods prior to the administration of various cognitive and affective tasks. Alpha activity from left and right frontal and left and right parietal regions (F3, F4,P3, P4) referenced to a common vertex (Cz) was extracted from all artefact-free epochs. The results revealed no group differences. For either frontal or parietal EEG during the eyes open base-line period. However for the eyes-closed resting period, a significant (0.5) Group effect was obtained on a measure of frontal alpha asymmetry (R-L/R+L. Alpha power; higher numbers on this index reflect greater relative left-sided activation). Depressed subjects show greater relative right-frontal activation than do non-depressed subjects

Research on anterior asymmetries associated with the phasic arousal of emotion

IDs in baseline asymmetry and affective reactivity

Anterior asymmetries during the experimental arousal of approach and withdrawal related emotion

p 132 ' adult subjects were exposed to short film clips designed to induce approach-related positive emotion and withdrawal-related negative emotion'

' when two or more emotions are compared for their effects on either physiological or behavioural dependent variables the intensity of the elicited emotion must be comparable,. If the emotions differ in intensity, then  any differences found between them could be attributed to intensity per se , rather than to the qualitative nature of the emotions aroused.'

-matched intensity determined by self report!

FACs scoring of concurrent video recording used to termite onset of happy and disgust facial expressions

EEG during the periods was Fourier transformed and power in different frequency bands was calculated. p 133 'Based upon on the theory evidence reviewed above, we hypothesised that the disgust periods would be associated with greater right sided anterior activation that would happy periods and that the happy periods would be associated with greater left- sided activation than would disgust periods.

Davidson takes the mean log-transformed alpha power and with this transformation ' more negative numbers indicate less alpha power. The negative numbers are a function of the log transformation.  Lower  numbers  ( i.e. more negative ) are associated with increased activation' (KRO ???? implies that less alpha power is associated with increased activation)  The results p 133-134 ' disgust is associated with less alpha power I.e. more activation, in the right frontal lead than is happiness, while happiness is associated with less alpha power in the left frontal lead than is disgust. '  Also found at temporal electrodes but not at the parietal.

For asymmetry index p 134 ' log right minus log left alpha power. Higher numbers on this scale denote greater relative left sided activation' all subjects showed a lowers ore on is index during the disgust rather than happy phase. 'We also found at there were large IDs among individuals in their average asymmetry scores. ( across emotion conditions ). The difference between happy and disgust conditions appears to be superimposed upon subjects' basal levels of asymmetry.'  When comparisons were made without preselects on of EEG based on facial expression the means were in the expected direction but   a statistically significant difference could not be  confirmed for the asymmetry measures.

Studies in infants: Important because it compares spontaneous  and voluntary smiling

Used taste to investigate whether asymmetries are present at birth - yes.

Spontaneous - both zygomatic and orbicularis oculi - felt happiness (Duchenne)

Voluntary zygomatic

Darwin (1872) suggested that the duration of each smile type differed. Ie spontaneous longer.  Confirmed by Ekman, Davidson & Frieson, 1990 self report of emotion and relative activation of the appropriate facial muscles was correlated.

Facial behaviour coded by EM FACs system (streamlined version of the FACs)

Stimulus - approach of either a stranger or the mother. Results p 138  'We specifically predicted that smiles with orbicularis activity would be accompanied by more relative left - sided frontal activation than the other smile.  ..... Data strongly supported this hypothesis.'  more left frontal activation with spontaneous smiling  has been confirmed in adults. (Ekman, Davidson & Friesen, 1990)

Individual differences in anterior symmetry: A neural substrate for affective style.

Davidson sees any response to emotional stimuli as  p 139 ' superimposed upon large individual differences in the overall magnitude and direction of asymmetry. In other research, we have found that a subject's overall (across task) EEG asymmetry during task performance  is highly correlated with their asymmetry during a resting baseline. ( e.g. Davidson, Taylor and Sharon, 1979)  and that anterior asymmetries during resting baselines are stable over time ( Tomarken et al, 1990). With test- retreat correlations ranging between 0.66 and 0.73 for different measures of anterior activation asymmetry

Scaffer, Davidson & Saron, 1983 -depressed subjects had less left frontal activation than the non depressed  subjects. p 140 ' An important question concerning these findings with depressives is the degree to which they are state - dependent. Is the decrease in left frontal activation a marker of the state of depression or  is a more trait like characteristic which marks an individual' s vulnerability to depression'

Results from an experiment to investigate that question  ' The results indicate that the decreased left anterior activation which is characteristic of depression remains. Even when the depression is remitted. In turn, these findings suggest that 'depressogenic' asymmetry patterns may be a  state -independent marker that indexes risk for depression.'

In a further study p 141 ' We hypothesised that subjects with greater right sided frontal activation at rest would report more intense negative affect in spouse to film clips signed to elicit fear and disgust. ...... W found that measures of frontal activations asymmetry recorded prior to the presentation of the film clips accounted fro a significant variance in subjects' self-reports of  negative affect in response to the clips. Subjects with greater relative right-sided frontal activation reported. More intense negative affect to the clips.'

Balaam et al (2011)

Balaam, M., Fitzpatrick, G., Good, J., &Harris, E. (2011)

Enhancing Interactional Synchrony with a ambient display

CHI 2011, May7-12, Vancouver, BC, Canada

Paper brings together 'research from social psychology with advances in HCI to explore a new role for pervasive and ambient technologies'  p 1

Pervasive and ambient technologies could provide  support both for encoding and decoding in authentic environments

Most of the technologies that have been developed to support social interaction (face-to-face) are informational  or support the verbal aspects e.g. RFID badges, iBand, GroupWear with mixed reactions from those involved.

Other approaches use para linguistics information to provide a view of an  interactional context.

Non verbal communication ' Research suggests that it communicates information about emotion, interpersonal relationships, and is an essential aspect of social competence' p 2 (2,7,20,22,37)

Rapport

'Rapport is one element of social interaction that is strongly encoded through nonverbal communication, and additionally theorised as key to the development of interpersonal relationships, and the success of an interaction. Rapport is believed to be crucial to develop trust, to create closeness, and be included in a  group(22) as well as help interactions feel smooth (36) .' P 2

Has three components ( see ref 37 )

1.    Mutual attentiveness

2.    Positivity

3.    Coordination ( sometimes referred to as interactional synchrony) - postural and movement. Much of this work done by Bernieri who foregrounds the importance of interactant proximity. Also that physical aspects (physical artefacts) of the task are important.

Each of which is encoded through non verbal communication with p7 ' with positivity most important at the outset  whilst coordination becomes more important later.

'there can be a disconnect between how rapport is encode and how observers decode rapport from NV signals' p 2. That  it is ' expressivity that drives social judgements' p 3

Ethics of using persuasive ambient technology is discussed on p 8.

the implication he is that more subtle NV communication can be missed  and led to the research aim - to use ambient technology to amplify (make more salient) the the more subtle NV that is sometimes missed by the observer.

'There are very few guidelines available' p 3 technologically speaking that is.

Scenario

Holiday planning in duos . Map display surrounded by representation of water that could be animated according to a wizard of Oz observer who took the role of the ambient technology of the future.

Ripples activated by

•   Simultaneous movement

•   Tempo similarity

•   Coordination

•   Posture matching

•   Back channel responses ( e.g. Head nodes, utterances showing agreement,/listening such as 'uh ha'

•   Eye contact

Explored ' the relationship between self reports of rapport, observer reports of rapport, and the body language of the participants in the holiday planning.' P 3 did the task twice (for personal use, for an older pair), once with the animation and once without. Untrained observers rated a video of the exchange , the video record was subject to a mosaic procedure that removed some of the expressive codes ( the  ones that are attended too at the expense of more subtle NVs)

Rapport scale by Bernieri (

ref 7) based on theory of 37.

'The data presented in the Results section indicates that our amplification of rapport through an ambient display increased the amount of interactional synchrony, and more specifically, the amount of coordination and posture similarity seen within an interaction by untrained observers.  At the same time, interact ants themselves were seemingly unaware of these increased behavioural manifestations of rapport, with no significant differences in their self reports of rapport between conditions'

Barron and mutual attention p 9 ' To aid this mutual attention, groups coordinated themselves and their work through the use of physical artefacts in the environment, conversational devices, and coordinated physical movement'   Overlaps with Salt and Bernieri

Discussion section contains more detailed account of HCI considerations

Refs

23, 34 spontaneous versus impression management

28 empathy

37, 5, 22, 7rapport

12, 18 Pervasive technology to support Group work