You are on page 1of 4

Computers in Human Behavior 27 (2011) 11141117

Contents lists available at ScienceDirect

Computers in Human Behavior


journal homepage: www.elsevier.com/locate/comphumbeh

Group awareness tools for learning: Current and future directions


Jrgen Buder
Knowledge Media Research Center, Tbingen, Germany

a r t i c l e

i n f o

Article history:
Available online 5 August 2010
Keywords:
Computer-supported collaborative learning
Group awareness
Design features

a b s t r a c t
Group awareness has become an important concept since it was introduced into the eld of computersupported collaborative learning. This paper discusses current trends and future directions in this
research eld. It is argued that the development and implementation of tools should be complemented
by systematic explorations into the mechanisms that moderate the relationship between group awareness and learning. It is suggested that variations in tool design features are a starting point for furthering
our understanding of the processes involved in group awareness. Based on the contributions in this special issue, eight areas for future empirical investigations are identied. The paper concludes with some
theoretical considerations on the nature of group awareness.
2010 Elsevier Ltd. All rights reserved.

1. Introduction
When the term awareness was coined some 20 years ago,
many scholars in the eld of computer-supported collaborative
work (CSCW) held that interacting via computers was lacking the
richness of natural, unmediated interaction. In CSCW research,
awareness became an umbrella term to express precisely those
qualities that were lacking in computer-mediated environments;
in other words, awareness was dened ex negativo. Consequently,
early technological solutions to provide awareness were trying to
recreate the gold standard of face-to-face environments, e.g.
through the use of video cameras that captured how work activities unfolded across space. In subsequent years, the notion of
awareness was extended considerably, and while Gutwin and
Greenberg (1995) were among the rst to theorize about social aspects of awareness, CSCW research and development was still
bound to the idea of facilitating the perception of spatially
grounded activities (seeing who is around; seeing who is located
in real or virtual space; seeing what others are doing).
For large parts of the last 20 years, the notion of awareness has
been conned to the area of CSCW. However, about 5 years ago the
concept was begun to be explored by a number of research groups
in the eld of computer-supported collaborative learning (CSCL) as
well. Along with the move from cooperative work to collaborative
learning came a number of different ideas on what awareness is
about. First of all, providing environmental or spatial cues plays a
much smaller role in the relevant CSCL literature. Rather, awareness tools focus on social aspects, i.e. on information that is inex Address: Knowledge Media Research Center, Konrad-Adenauer-Str. 40, 72072
Tbingen, Germany. Tel.: +49 7071 979 326; fax: +49 7071 979 100.
E-mail address: j.buder@iwm-kmrc.de
0747-5632/$ - see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chb.2010.07.012

tricably bound to other persons (hence the label of group


awareness). Secondly, while classical awareness is often limited
to observable phenomena (presence, activities), group awareness
in CSCL has a much stronger emphasis on cognitive or social categories that are not directly observable (e.g. knowledge or attitudes;
Buder & Bodemer, 2008). Thirdly, along with the shift from observable to non-observable categories, face-to-face interaction is no
longer the gold standard to be achieved. In fact, most group awareness solutions in CSCL focus on qualities of interaction that are difcult or even impossible to be achieved in face-to-face contexts,
thereby providing an added value to computer-mediated collaboration (Buder, 2007). This special issue bears witness to these recent developments in CSCL, and it provides many interesting
insights into the question of whether supporting group awareness
gives rise to better collaborative learning in terms of processes and
outcomes.
The contributions in this issue all focus on awareness about
cognitive and/or social variables, and they all explore the relationship between awareness and learning. Nonetheless, they represent
a wide array of different scenarios, thereby exemplifying the broad
scope of group awareness applications. For instance, the learning
outcomes that were addressed range from individual learning performance (knowledge tests) to collaboratively constructed products (essays). The processes investigated in these studies cover
very different learner activities, among them patterns of verbal
communication, manipulations of graphical elements in interaction, and frequencies of awareness tool use. Finally, each of the
tools described in this special issue differs in what precisely is
made aware. In the study of Bodemer (2011), participants were
made aware of their collaborators situated use of knowledge
pieces. Subjects in the study of Dehler, Bodemer, Buder, and Hesse
(2011) were made aware of self-assessments of their learning

J. Buder / Computers in Human Behavior 27 (2011) 11141117

partners in terms of the degree of understanding of learning materials. Sangin, Molinari, Nssli, and Dillenbourg (2011) explored
awareness about objective levels of partner performance from a
knowledge test. The tool developed by Janssen, Erkens, and Kirschner (2011) provides awareness about the overall writing activity
levels of participants. And the study by Phielix, Prins, Kirschner, Erkens, and Jaspers (2011) addresses awareness about social and
cognitive categories like the friendliness or the productivity of collaborators. Despite all the differences in the scenarios involved,
these studies have all shown a relationship between group awareness and indicators of learning, and they add to an impressive list
of other studies that have found such a relationship in the past.
While the diversity of settings and tools makes it exceedingly difcult to provide a comprehensive and clear-cut denition of group
awareness, it also underlines the enormous potential of group
awareness support systems for computer-supported collaborative
learning.
As the eld matures, it can be expected that more and more
studies will show the benets of group awareness technologies
for collaborative learning, both in laboratory settings and in educational practice. However, apart from developing ever new tools to
support awareness, we should also begin to systematically explore
the underlying mechanisms that impact the relationship between
awareness and learning outcomes. The paper by Fransen, Kirschner, and Erkens (2011) provides many important insights into
the psychological variables that are related to group awareness,
but it does not directly address the role of technology in this process. What we need, then, is an understanding of the potentially
complex interactions between group features, design features, task
features, learning processes, and learning outcomes. The present
contribution is an attempt to integrate some of the ndings from
this special issue and from other sources in order to provide building blocks towards a deeper understanding of why and how group
awareness can foster learning. With a particular emphasis on design features, it tries to identify relevant research questions that
deserve to be tackled in dedicated empirical investigations.
The next sections of this paper are organized around a distinction made by Schmidt (2002) in his review on CSCW awareness research. He identied two observable activities that can be found in
virtually all settings where (group) awareness plays a prominent
role. The rst of these activities is displaying which can loosely
be described as the process of making something aware. The second activity is monitoring, and it refers to the process of actually
becoming aware of information that was displayed by others before. Coordination between collaborators can be regarded as an
ongoing cycle between displaying and monitoring activities. The
next two sections discuss some empirically open questions that
can be associated with displaying and monitoring activities.

2. Displaying
Displaying refers to the processes by which the things to be
made aware of are generated. There are several methods of how
to design and support displaying activities, but we only have a very
rough understanding of what method is appropriate in a given context. This section describes four different issues that are associated
with different design options.
The rst issue to be discussed refers to two alternative principles that can lead to the display of awareness information which
are commonly referred to as explicit feedback vs. implicit feedback
in the literature on information retrieval systems. Explicit feedback
involves a deliberate, intentional and conscious displaying activity
by learners. For instance, in Bodemers study (2010), participants
intentionally assigned graphical elements in the collaborative integration task to express their current understanding of statistics

1115

concepts. Another type of explicit feedback is through user/learner


ratings, and this method was employed in the studies by Dehler
et al. (self-assessments), as well as Phielix et al. (assessments of
self and others). In contrast, in implicit feedback systems tools
automatically generate awareness information without requiring
dedicated learner activities. The studies by Sangin et al. (provision
of information about knowledge tests) and Janssen et al. (provision
of information about the amount of participation) are examples of
this form of display. The question, then, is if one type of displaying
is advantageous with regard to collaborative learning. Not surprisingly, many computer scientists and engineers prefer implicit feedback systems, mainly because awareness information is gained
rather elegantly and unobtrusively. Moreover, as pointed out by
Sangin et al. (2011), tools that do not require explicit ratings provide objective rather than subjective knowledge awareness. However, particularly in the eld of collaborative learning one should
consider potential benets of explicit feedback and display. Ratings
and explicit displaying activities might involve additional workload and potential distraction from the learning tasks, but they cater quite well to the constructivist nature of many learning tasks.
For instance, requiring learners to rate aspects of their collaboration with regard to cognitive or social categories might serve as a
meta-cognitive prompt that helps to reect on a task. The Radar
and Reector tools by Phielix et al. (2010) are based on this potential. Furthermore, explicit activities like assigning graphical elements (Bodemer), rating oneself (Dehler et al.), or rating others
(Phielix et al.) can be regarded as a low-level form of active participation. Participation, in turn, is one of the most important antecedents of success in collaborative learning. Finally, implicit
feedback is probably at an advantage for some types of group
awareness. For instance, it would be quite difcult to extract
friendliness or reliability of co-learners without recourse to explicit
ratings. There might be ways to get such information through computerized means (e.g. latent semantic analysis of contents), but
these solutions are associated with relatively high computational
costs and potentially low validity. Moreover, being told by an algorithm how a group feels, might lead to reactance. Nonetheless, it
would be very interesting to see studies that directly compare explicit and implicit feedback or display. For instance, one could use
the tool from Dehler et al., and compare self-assessments of knowledge with ne-grained results of a knowledge pretest. Or one could
ask learners to rate the participation levels of their collaborators,
and compare this to the tool employed by Janssen et al. (2011).
A second issue with regard to displaying refers to the question
of using dynamic vs. static displays of awareness information. For
instance, the collaborative integration tool employed by Bodemer
(2010) provides learners with awareness information that is constantly updated through ongoing activities in real time. The participation tool of Janssen et al. is similarly dynamic. In contrast, the
remaining tools described in this issue rely on static awareness
information that was gained before group interaction (Dehler
et al., Sangin et al.), or in repeated display cycles during collaboration (Phielix et al.). The advantage of dynamic displaying is that it
provides learners with up-to-the minute information about the
collaborative process that can lead to immediate ne-tuning of
activities. However, at least for explicit feedback methods there
is a trade-off between immediacy and additional workload associated with repeated ratings.
Thirdly, there is the issue of encouraging or even forcing learners to display. This aspect was discussed by Janssen et al., and it
was addressed through the design of the Phielix et al. tool where
participants could only gain access to awareness information when
they had completed their own ratings. Enforcing or scripting the
display of group awareness might be burdensome and lead to lower tool acceptance. On the other hand, gaining a complete picture
about cognitive and social variables in a group is one of the fea-

1116

J. Buder / Computers in Human Behavior 27 (2011) 11141117

tures that provide an advantage over face-to-face environments


where displaying of information is usually optional. Future studies
might try to identify the perfect balance between voluntary participation and comprehensiveness of data.
Fourthly, there is the issue of the display format. This can best
be described by comparing the Radar tool and the Reector tool,
both from the Phielix et al. study. The Radar tool uses a closed format with a rating on a pre-dened scale whereas the Reector tool
uses an open display format in which learners express awareness
information through a text eld. While an open format might have
the advantage of enabling a more nuanced display of group awareness, an earlier study by Phielix, Prins, and Kirschner (2010) has
shown that students regarded the open format of the Reector tool
as less useful. However, such a result needs to be conrmed by further studies.
In sum, awareness tool designers for learning have to decide
whether to use explicit or implicit feedback; they should implement measures in order to ensure that awareness information does
not become outdated through ongoing interaction dynamics; they
might consider introducing display activities as a requirement for
learners; and they should choose a display format that learners
are most likely to use. Each of these design options could or should
be informed by the results of scientic investigations, but these
empirical questions havent been tackled so far.

3. Monitoring
Monitoring describes the process of becoming aware of information that was displayed by other group members. While Fransen
et al. (2011) have investigated monitoring behavior by means of a
post hoc questionnaire, the process itself is an integral part of collaboration as it unfolds. This special issue provides some interesting insights into the monitoring process, e.g. by tracking at what
times and for which duration awareness information is used (Sangin et al., Janssen et al.). This section describes four issues that deserve further empirical investigation with regard to monitoring.
The rst issue relates to the fact that similar to displaying, monitoring can be more or less obtrusive. Regulating awareness is generally held to be a secondary task to the main task that a group has
to accomplish (Gutwin & Greenberg, 2002). Bodemers collaborative integration tool is an interesting departure from this idea, as
no distinction between primary (learning activity) and secondary
task (regulating awareness) can be drawn. However, in cases
where displaying and monitoring become extra activities, there is
the open question of how much monitoring diverts attention from
the main task. Most monitoring tools described in this special issue
were contained in a dedicated window of the main workspace
(Dehler et al., Sangin et al., Phielix et al.). However, the participation tool by Janssen et al. required participants to open an extra
window for monitoring the activity levels of their collaborators.
Further investigations might address the levels of extraneous cognitive load that is associated with splitting attention between different workspace windows vs. opening an additional window
when the situational affordances call for it.
A second issue with regard to monitoring refers to the ability
and ease with which to compare pieces of information that were
displayed. For instance, the participation tool (Janssen et al.) and
the Radar tool (Phielix et al.) enable learners to compare assessments of their own performance with the performance of others,
albeit with the use of visualizations that might be difcult to
understand for novices. In contrast, the collaborative integration
tool (Bodemer) and the GKA tool (Dehler et al.) are maximized
for salience and comparability between own display and partner
display by using adjacent boxes. Consequently, both Bodemer
and Dehler et al. argue that comparability should be a main feature

of group awareness tools. It is interesting to compare these solutions to the tool of Sangin et al., as it is the only example in this
special issue where only partner performance can be monitored,
thus providing no support for comparability. Sangin et al. have
found in their study that learning performance was related to the
amount of uncertainty markers in the linguistic utterances of collaborators, and it might be that not knowing about ones own performance introduces exactly this element of uncertainty. As it is,
comparability is a double-edged sword: in the best case, it triggers
help-giving behavior, as was found in the study by Dehler et al. in
the worst case, it can be associated with social psychological phenomena like downward comparison or sucker effects, and lead to
withholding of information. It should be interesting to explore
the conditions that lead to positive or detrimental outcomes of
comparability.
The third issue deserving further investigation is related to the
notion of comparability, viz. the general normative function of
group awareness tools. It can be argued that many group awareness tools make group norms visible, and this in turn will regulate
the ow of collaboration. In the papers of this special issue, this is
probably most obvious in the case of the participation tool (Janssen
et al.). Perceiving and knowing that others have contributed more
than oneself is likely to activate some degree of normative pressure, thereby preventing free-riding from taking place. On the
other hand, knowing that performance will be monitored by others
can lead to evaluation apprehension (Cottrell, 1972). For instance,
it could be that the GKA tool from Dehler et al. can backre in those
cases where learners of clearly different levels of knowledge are
paired, thereby impeding the acceptance of the tool. How to deal
with evaluation apprehension is a largely unresolved issue. It
might be that learners counteract the dangers of evaluation apprehension through an ination of positive ratings. This conclusion
could be drawn from the unexpected nding of Phielix et al.
(2010) showing that partner ratings increased the longer the collaboration lasted.
The fourth issue that is related to monitoring refers to the directivity and/or guidance of group awareness tools. It can be speculated that the effectiveness of tools for learning is positively
correlated to the degree of behavioral adaptation that they bring
about: a tool that does not change collaboration is unlikely to have
a positive effect. The tools described in this special issue certainly
all adhere to the idea of behavioral adaptation, but it can be argued
that they support different degrees of coupling monitored information with immediate action. For instance, the knowledge awareness tool by Sangin et al. helps learners to mentally contextualize
the utterances of their partners, but it does not trigger immediate
action on the part of the learner who monitors the partner performance. In contrast, the participation tool (Janssen et al.) and the
Radar tool (Phielix et al.) certainly provide cues for behavioral
adaptation, but it might still be difcult for learners to implement
this adaptation. It is one thing to see that one should be friendlier
or more productive, but the tools themselves do not foster the
skills that are necessary to actually become friendlier or more productive. Guidance and directivity are probably more pronounced in
the GKA tool (Dehler et al.) and the collaborative integration tool
(Bodemer). In these cases, the tools provide strong affordances
for immediate action by not only making salient interpersonal
commonalities and differences, but also by triggering adaptive
behavior (asking questions vs. giving explanations in the GKA tool;
adding elements and focusing on conicting assignments in the
collaborative integration tool). However, it is conceivable to
strengthen directivity and guidance even more, e.g. through the
use of explicit recommendations generated by the tool, or through
scripting mechanisms. Dehler, Bodemer, Buder, and Hesse (2009)
have already discussed the need to translate awareness information in a way that coordinated activities are facilitated.

J. Buder / Computers in Human Behavior 27 (2011) 11141117


Table 1
Empirical issues in research on group awareness tools and learning.
Process

Issue

Displaying
Displaying
Displaying
Displaying
Monitoring
Monitoring
Monitoring
Monitoring

Explicit (ratings) vs. implicit feedback


Dynamic display vs. repeated display vs. static display
Voluntary vs. enforced display activities
Closed vs. open display format
Obtrusive vs. non-obtrusive monitoring
Interpersonal comparability of performances
Normative pressure vs. evaluation apprehension
Guidance and directivity of tools

In sum, designers of group awareness tools for learning must try


to nd a balance between primary tasks (collaboration) and secondary tasks (monitoring); they should consider potentials and
drawbacks of providing learners with information that will foster
social comparison; they should take issues like evaluation apprehension into account; and they should nd ways to provide guidance without diminishing learner autonomy. Once again,
empirical investigations can help to elucidate these design issues.
4. Conclusions
Research on group awareness has made considerable progress,
particularly since the concept was applied to the eld of computer-supported collaborative learning. The contributions in this
special issue attest to the power of group awareness in the support
of collaboration. However, this paper argues that we should not
content ourselves with developing more and more working examples of group awareness tools, but should also try to systematically
explore the mechanisms and boundary conditions under which
group awareness tools are particularly useful. It raises a number
of open research questions (summarized in Table 1) that will help
us understand the dynamics of tool design, group awareness, and
learning.
In essence, this paper suggests moving from proof of concepts
(studies comparing tool vs. no tool conditions) to a systematic variation of design features (studies comparing different awareness
tools). The potential empirical advances that could be made
through a systematic exploration of awareness tools might also
spur much-needed theoretical advances. As it is, the notion of
group awareness is almost exclusively used in the context of technology. However, most scholars would agree that group awareness
is a powerful mechanism of group interaction even in the absence
of technology support. The question, then, is how group awareness
relates to well-established theoretical constructs. The paper by
Fransen et al. (2011) makes an important step in this direction
by exploring the relationship between group awareness (mutual
performance monitoring), mutual trust, and shared mental models.

1117

The analysis by Fransen et al. can be interpreted in a way that relatively stable, long-term concepts like shared mental models are a
better determinant of learning success than situational, short-term
group awareness and performance monitoring. But it leaves open
the question of whether group awareness is an important antecedent of stable cognitive structures like shared mental models or
transactive memory systems, a possibility that was raised in an
overview by Engelmann, Dehler, Bodemer, and Buder (2009).
In sum, research on group awareness still has many empirical
and theoretical areas that are left unexplored. However, this should
not be regarded as a drawback, but rather as an encouraging sign
that group awareness is likely to stay an active eld of research
in the area of computer-supported collaborative learning.
References
Bodemer, D. (2011). Tacit guidance for collaborative multimedia learning.
Computers in Human Behavior, 27(3), 10791086.
Buder, J. (2007). Net-based knowledge-communication in groups: Searching for
added value. Zeitschrift fr Psychologie/Journal of Psychology, 215(4), 209217.
Buder, J., & Bodemer, D. (2008). Supporting controversial CSCL discussions with
augmented group awareness tools. International Journal of Computer-Supported
Collaborative Learning, 3(2), 123139.
Cottrell, N. B. (1972). Social facilitation. In C. McClintock (Ed.), Experimental social
psychology (pp. 185236). New York: Holt, Rinehart & Winston.
Dehler, J., Bodemer, D., Buder, J., & Hesse, F. W. (2009). Providing group knowledge
awareness in computer-supported collaborative learning: Insights into learning
mechanisms. Research and Practice in Technology Enhanced Learning, 4(2),
111132.
Dehler, J., Bodemer, D., Buder, J., & Hesse, F. W. (2011). Guiding knowledge
communication in CSCL via group knowledge awareness. Computers in Human
Behavior, 27(3), 10681078.
Engelmann, T., Dehler, J., Bodemer, D., & Buder, J. (2009). Knowledge awareness in
CSCL: A psychological perspective. Computers in Human Behavior, 25(4),
949960.
Fransen, J., Kirschner, P. A., & Erkens, G. (2011). Mediating team effectiveness in
collaborative learning: The importance of team and task awareness. Computers
in Human Behavior., 27(3), 11031113.
Gutwin, C., & Greenberg, S. (1995). Support for group awareness in real-time
desktop conferences. In: Proceedings of the second New Zealand computer
science research students conference. University of Waikato, Hamilton, New
Zealand, April 1821, 1995.
Gutwin, C., & Greenberg, S. (2002). A descriptive framework of workspace
awareness for real-time groupware. Computer Supported Cooperative Work, 11,
411446.
Janssen, J., Erkens, G., & Kirschner, P. A. (2011). Group awareness tools: Its what you
do with it that matters. Computers in Human Behavior, 27(3), 10461058.
Phielix, C., Prins, F. J., Kirschner, P. A., Erkens, G., & Jaspers, J. (2011). Group
awareness of social and cognitive performance in a CSCL environment: Effects
of a peer feedback and reection tool. Computers in Human Behavior., 27(3),
10871102.
Phielix, C., Prins, F. J., & Kirschner, P. A. (2010). Awareness of group performance in a
CSCL environment: Effects of peer feedback and reection. Computers in Human
Behavior, 26, 151161.
Sangin, M., Molinari, G., Nssli, M.-A., & Dillenbourg, P. (2011). Facilitating peer
knowledge modeling: Effects of a knowledge awareness tool on collaborative
learning outcomes and processes. Computers in Human Behavior, 27(3),
10591067.
Schmidt, K. (2002). The problem with awareness. Computer Supported Cooperative
Work, 11, 285298.