Scripting argumentative knowledge construction in
computer-supported learning environments
Armin Weinberger, Karsten Stegmann, Frank Fischer, Heinz Mandl
To cite this version:
Armin Weinberger, Karsten Stegmann, Frank Fischer, Heinz Mandl. Scripting argumentative
knowledge construction in computer-supported learning environments. F. Fischer, I. Kollar, H.
Mandl, J.M. Haake. Scripting Computer-Supported Collaborative Learning Cognitive, Com-
putational and Educational Perspectives, Springer, pp.191-209, 2006. <hal-00190423>
HAL Id: hal-00190423
Submitted on 23 Nov 2007
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entiﬁc research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destin´ee au d´epˆot et `a la diﬀusion de documents
scientiﬁques de niveau recherche, publi´es ou non,
´emanant des ´etablissements d’enseignement et de
recherche fran¸cais ou ´etrangers, des laboratoires
publics ou priv´es.
SCRIPTING ARGUMENTATIVE KNOWLEDGE
CONSTRUCTION IN COMPUTER-SUPPORTED
, Karsten Stegmann
, Frank Fischer
, and Heinz Mandl
Knowledge Media Research Center, Tübingen;
: Computer-supported collaborative learning (CSCL) environments may encour-
age learners to engage in argumentative knowledge construction. Argumenta-
tive knowledge construction means that learners work together to elaborate on
concepts by constructing arguments and counterarguments. This is achieved
through discourse with the goal of acquiring knowledge within a specific do-
main. However, learners may encounter problems relating to one of three di-
mensions of argumentative knowledge construction. First, learners seem to
have difficulties in constructing arguments that contribute to solving the task.
Second, learners’ arguments may lack important components such as data and
warrants. Third, learners rarely build upon the arguments of their learning
partners. Structuring argumentative knowledge construction with collaboration
scripts is a promising instructional approach for facilitating specific process di-
mensions of argumentative knowledge construction. Little is known, however,
about how to most effectively facilitate the acquisition of knowledge by di-
recting collaboration scripts at specific dimensions of argumentative knowl-
edge construction. This chapter will outline the theoretical background of ar-
gumentative knowledge construction and will then describe script components
that target different dimensions of argumentative knowledge construction. The
chapter will then discuss the empirical findings of two studies regarding the ef-
fects of these script components.
Collaborative learners are sometimes meant to construct and exchange
arguments by collecting and balancing evidence and counterevidence
through discourse with the goal of acquiring individual knowledge within a
domain. Typically, however, learners do not work well on collaborative
learning tasks in terms of constructing adequate arguments and interacting
productively (Kuhn, Shaw, & Felton, 1997; Mandl, Gruber, & Renkl, 1996).
Kuhn’s work points out that even adult discussants rarely warrant or qualify
192 Chapter 12
their claims and thus rarely construct complete arguments. Furthermore, dis-
cussants are often unable to balance and integrate arguments and counter-
arguments critically. It has become clear that simply asking learners to col-
laborate is not sufficient for fostering argumentative knowledge construction
(Mandl et al., 1996).
Asynchronous computer-supported collaborative learning has been re-
garded as a suitable context for facilitating argumentative knowledge con-
struction (Andriessen, Baker, & Suthers, 2003; Marttunen & Laurinen,
2001). Learners communicate simultaneously or in an unspecified sequence
via text-based interfaces and are thus able to type and read messages at their
own individual pace. In this way, learners have more time than face-to-face
learners to both compose their own messages and understand the messages
of their learning partners. This time advantage may encourage learners with
heterogeneous argumentation skills to take part in an argumentative debate.
Besides the time advantage, learners who communicate asynchronously via
computer can repeatedly access the arguments that have been already
contributed and can easily revise the wording of their own arguments (see
Pea, 1994). In text-based asynchronous communication, learners may
compensate for individual deficits in learning prerequisites by investing
more time in the reception and production of individual contributions.
Learners may also take advantage of individual adaptive instructional sup-
port that is provided as part of the communication interface, such as com-
puter-supported scripts. The question is, to what extent single script compo-
nents could be directed at specific process dimensions of argumentative
knowledge construction to improve individual knowledge acquisition. These
could potentially be used to varying degrees, depending on how well the
script components are able to compensate for the deficits of the student
In this chapter, we first describe our approach to argumentative knowl-
edge construction according to three process dimensions and their concep-
tual relationship to individual knowledge acquisition. Furthermore, we ana-
lyze how computer-supported script components may facilitate argumenta-
tive knowledge construction within these process dimensions. Finally, we
summarize and discuss the results of two empirical studies that investigate
the effects of computer-supported script components (with specific goal di-
mensions) on the processes and outcomes of argumentative knowledge con-
struction. These studies have also been published in greater detail (see
Weinberger, Ertl, Fischer, & Mandl, 2005 and Weinberger, Stegmann, &
12. Scripting argumentative knowledge construction 193
1. ARGUMENTATIVE KNOWLEDGE CONSTRUC-
In argumentative knowledge construction, learners acquire knowledge
through the elaboration of learning material by constructing arguments.
Typically, argumentative knowledge construction scenarios are based on
collaborative learning tasks (Leitão, 2000). For instance, learners may be
provided with contrasting hypotheses about a problem and then argumenta-
tively discuss, which hypothesis applies (Kollar, Fischer, & Slotta, 2005).
Learners also engage in argumentative knowledge construction when work-
ing on open-ended and complex problem cases for which they have to create
and balance multiple hypotheses (Means & Voss, 1996). Our approach to
argumentative knowledge construction in collaborative, problem-based sce-
narios differentiates between three process dimensions, namely an epistemic
dimension that describes arguments as steps towards solving the learning
task, an argument dimension in which formal criteria for the composition of
arguments are represented, and a dimension of social modes of co-construc-
tion that represents how learners interact with one another. This interaction
is described in terms of how the learners relate their own arguments to the
arguments of their learning partners (see Weinberger & Fischer, 2006).
1) The epistemic dimension refers to the question of how learners work
on the tasks they are confronted with, for example, by constructing relation-
ships between the conceptual space and the problem space (Fischer, Bruhn,
Gräsel, & Mandl, 2002). Arguments may provide hypotheses on how to
solve complex tasks. Learners relate the theoretical concepts of a given the-
ory to the information in a problem. The epistemic dimension of arguments
can indicate which concepts learners refer to and how learners connect con-
cepts to solve a problem. This dimension can also show the extent to which
learners are able to adequately apply knowledge. Collaborative learners do
not always apply knowledge appropriately. When learners verbalize inap-
propriate applications of knowledge, there is a chance that their learning
partners may adopt these misconceptions (e.g., Jeong & Chi, 1999). In this
respect, the epistemic dimension of arguments is an important component of
argumentative knowledge construction. Depending on how well learners are
able to construct arguments relating to the epistemic dimension, they may
acquire adequate knowledge or misconceptions (Weinberger, 2003).
2) The argument dimension comprises how learners construct arguments
with regard to defining formal relationships between specific components of
arguments, such as claims, data, and warrants. Toulmin (1958) has put for-
ward a structural model of single arguments that is made up of several com-
ponents. In this model, a single argument consists of a claim, which is a con-
clusion that is being presented and justified in the argument. The claim is
194 Chapter 12
based on a datum, which is a fact that is supposed to support the claim. A
warrant specifies the principle of how the datum supports the claim. Some-
times a warrant itself needs support, which is called backing. Thus, the
backing indicates the principle upon which the warrant is based. Arguments
may optionally also provide components that limit the validity of the claim
and anticipate counterarguments. The qualifier indicates the extent to which
the datum warrants the claim or may limit the validity of a claim. A rebuttal
serves to anticipate parts of a counterargument that attack the data, the war-
rant, or the backing.
Toulmin’s model poses an alternative to formal logic, which is closer to
everyday reasoning in uncertain situations based on probabilities. However,
Toulmin’s model has been criticized for difficulties in distinguishing be-
tween the single components of the model in everyday argumentation, for
example, distinguishing backing from data or differentiating between a
qualifier and a rebuttal (Voss & van Dyke, 2001). We will therefore apply a
condensed argument model using the components of claim, datum with war-
rant and qualifier.
How is formal argumentation structure related to individual knowledge
construction? When learners construct arguments they elaborate and self-
explain the learning material (Baker, 2003). These self-explanations help
learners integrate new information into existing cognitive structures (Chi,
Bassok, Lewis, Reimann, & Glaser, 1989). In terms of Toulmin’s model
(1958), self-explanation could be described as a process of composing an
argument from several components. From this perspective, learners are sup-
posed to seek data that supports or opposes a claim, make an inference
through a warrant that indicates how the data supports a claim, and limit the
validity of a claim by constructing qualifiers.
3) The dimension of social modes of co-construction indicates how learn-
ers interact with one another. In this dimension, a number of social modes of
co-construction and their relationship to individual knowledge construction
have been identified. These indicate the different degrees to which learners
operate using the reasoning of their peers (Fischer et al., 2002; Teasley,
1997). For instance, when building consensus in a conflict-oriented manner,
learners need to pinpoint specific aspects of their peers’ contributions and
either modify them or present alternatives. In these terms, learners need to
build their reasoning more closely upon the reasoning of their peers when
working to build consensus in a conflict-oriented manner. This is in contrast
to quick consensus building, that is, when learners only appear to accept the
contributions of their learning partners in order to continue with discourse
(Weinberger, 2003). The extent to which learners operate on the reasoning of
what has been said before in discourse has been termed transactivity of dis-
12. Scripting argumentative knowledge construction 195
course, which is known to be positively related to individual knowledge ac-
quisition (Teasley, 1997).
Table 12-1. Process dimensions of argumentative knowledge construction (see Weinberger &
Fischer, in press)
Process dimension answers the question
Epistemic dimension How do learners’ arguments contribute to
solving the task?
Argument dimension How do learners construct arguments for-
Dimension of social modes of co-construction To what extent do learners operate on the
reasoning of their learning partners?
In summary, we propose to include three process dimensions for the
analysis and facilitation of argumentative knowledge construction based on
problem-oriented learning tasks (see Table 12-1). The processes of argu-
mentative knowledge construction can be analyzed on a) an epistemic di-
mension (constructing arguments that contribute to solving a task), b) an
argument dimension (building formally complete arguments), and c), a di-
mension of social modes of co-construction (operating on the reasoning of
2. SCRIPT COMPONENTS FOR ARGUMENTATIVE
Argumentative knowledge construction is based on the assumption that
learners need to construct arguments appropriately in order to benefit from
collaborative learning environments. One approach for facilitating the out-
comes of argumentative knowledge construction is to support learners in
constructing arguments appropriately. Collaboration scripts provide an in-
structional approach that aims to facilitate the processes of argumentative
Collaboration scripts were initially developed to encourage college stu-
dents working in dyads to acquire knowledge from texts on the natural sci-
ences (O’Donnell & Dansereau, 1992). Collaboration scripts provide more
or less explicit and detailed instructions for small groups of learners on what
activities need to be executed, when they need to be executed, and by whom
they need to be executed in order to foster individual knowledge acquisition.
Prototypical scripts are instructed prior to collaborative learning. Partici-
pants are trained to engage in the scripted collaborative activities, which are
in turn supposed to facilitate the individual acquisition of knowledge. In
computer-supported collaborative learning, there is typically no opportunity
for antecedent collaboration training. Thus, computer-supported collabora-
196 Chapter 12
tive learning (CSCL) is often facilitated by the design of the interface (Baker
& Lund, 1997; Scardamalia & Bereiter, 1996). Learners communicating via
these interfaces are, to varying degrees, implicitly guided to engage in ac-
tivities, as the interface suggests or limits specific discourse activities
(Runde, Jucks, & Bromme, this volume). Computer-supported scripts simi-
larly aim to directly influence the interaction patterns of collaborative learn-
ers rather than train learners prior to actual collaboration.
When analyzing collaboration scripts in the context of argumentative
knowledge construction, it can be noted that scripts typically aim to facilitate
different process dimensions simultaneously. For instance, prototypical
scripts may support epistemic activities, for example, summarizing a para-
graph, as well as specific social modes, for example, criticizing the contri-
butions of the learning partner (O’Donnell & Dansereau, 1992). Little is
known about the effects of single script components that target specific
process dimensions on knowledge construction. In research on collaboration
scripts, Larson and colleagues (1985) compared the effects of script compo-
nents with different, specific goal dimensions, namely an elaborative and a
metacognitive script component. The elaborative script component sup-
ported elaborative activities by modeling the role of a recaller, who was
given the task of personalizing information or of using imagery to help re-
member the learning material. The metacognitive script component modeled
the role of a listener who was given metacognitive tasks, such as error de-
tection. The metacognitive script component impeded individual knowledge
construction, whereas the elaborative script component facilitated individual
knowledge construction. This study thus indicates that differentiated effects
can be expected from script components with specific goal dimensions. Lar-
son and colleagues (1985) argued, for instance, that some script components
may also distract learners from learning goals. It is important to expand re-
search on script components with single goal dimensions in order to better
understand why and what kind of script components facilitate argumentative
knowledge construction. The research should also analyze possible side ef-
fects that single script components may have on the processes and outcomes
of argumentative knowledge construction. We may thus accumulate knowl-
edge on how multiple processes of argumentative knowledge construction
interact and affect individual knowledge acquisition.
With reference to the framework of argumentative knowledge construc-
tion, we differentiate between epistemic script components, argumentative
script components, and social script components (Weinberger & Fischer,
Epistemic script components aim to structure the discourse activities of
collaborative learners with respect to the content of the discussion and with
regard to the steps towards solving the task. Epistemic script components
12. Scripting argumentative knowledge construction 197
may support learners in finding adequate task strategies and may ask learners
to elaborate on aspects of the task they would not normally consider (cf.
Reiser, 2002). Approaches that we classify as epistemic script components
may thus map expert-like strategies onto the interaction of learners
(Dufresne, Gerace, Thibodeau Hardiman, & Mestre, 1992; Herrenkohl and
Guerra, 1998). For instance, these components provided collaborative learn-
ers with task strategies such as predicting and theorizing, summarizing re-
sults, and relating predictions and theories to results. As these qualitative
studies indicate, epistemic script components may need to be reinforced by
social script components.
Argumentative script components aim to support the construction of ar-
guments in terms of warranting and qualifying claims based on argument
models such as Toulmin’s (1958). Argumentative script components aim to
help learners construct formally adequate arguments and thus better elabo-
rate the learning material (Andriessen et al., 2003). As learners supplied with
argumentative script components are supposed to formulate better arguments
in discourse, learners may also acquire knowledge on how to argue within a
Social script components specify and sequence the interaction of learners
in order to promote knowledge construction (King, 1999). Social script com-
ponents may thus support learners to engage in adequate interaction strate-
gies that they would not apply spontaneously. For instance, social script
components may facilitate transactivity by asking learners to respond criti-
cally to the contributions of their learning partners. Social script components
typically also have learners rotate to work on different activities (e.g., Her-
renkohl & Guerra, 1998). The reciprocal teaching approach (Brown &
Palincsar, 1989), for instance, assigns the roles of a teacher and a learner for
various text comprehension tasks.
In summary, prior research on different script components found that not
all components are equally effective for promoting knowledge construction.
Some script components appear to distract learners from the actual task or
replace central learning activities rather than support learners in engaging in
the activities themselves. To date, there has been little systematic research on
the effects that various script components have on argumentative knowledge
construction. Furthermore, most research on scripts deals with trained face-
to-face learning environments rather than computer-supported scripts. There-
fore, there is little knowledge about how specific computer-supported script
components with different goal dimensions can facilitate the processes and
outcomes of argumentative knowledge construction.
198 Chapter 12
3. GOALS OF THE EMPIRICAL STUDIES
The goal of the empirical studies is to investigate the effects of computer-
supported epistemic, argumentative and social script components on argu-
mentative knowledge construction. These single script components focus on
different process dimensions of argumentative knowledge construction and
may have differentiated effects on its outcomes. We conducted two studies
in order to investigate how individual computer-supported script components
can facilitate argumentative knowledge construction (for further details see
Weinberger, Ertl, et al., 2005; Weinberger, Stegmann, et al., 2005).
The results are based on a one factorial analysis and single group com-
parisons of the experimental groups with epistemic, social, and argumenta-
tive script components with the control group. Each experimental group in
the two studies consisted of 24 students. Thus, 96 participants in 32 groups
of three entered the statistical analyses. The methods that were used in each
of the studies were identical.
4.1.1 Sample and setting
First-semester educational science students from the University of Mu-
nich participated in the studies. The students took an obligatory introduction
course to educational science. One of the regular face-to-face sessions of the
course was transformed into an online learning session. Participation in this
session was required in order to receive a course credit at the end of the se-
mester. The learning outcomes of the experimental session, however, did not
count towards the students’ overall performance. The introduction course
sessions normally consist of a one hour lecture and a successive two hour
seminar. Similarly, the collaborative online learning session took three
hours. The students were randomly assigned to the experimental conditions
in groups of three. Participants in each group of three were separated from
each other in different laboratory rooms and communicated asynchronously
with the help of web-based discussion boards in a computer-supported learn-
12. Scripting argumentative knowledge construction 199
4.1.2 Learning task
The task of the participants was to apply the attribution theory of Weiner
(1985) to three problem cases (see Table 12-2 for an example of a problem
case) and reach agreement on a final analysis for each case.
Table 12-2. One of the three problem cases, namely the “math case”, learners needed to ana-
lyze and discuss
As a student teacher in a high school, you participate in a school counseling session with
Michael Peters, a pupil in the 10
“Recently I’ve started to realize that math is just not my thing. Last year I almost failed math.
Ms. Weber, my math teacher, told me that I would really have to make an effort if I wanted to
grade. Actually, my parents stayed pretty calm when I told them this. First mom said
that nobody in our family is a math whiz. My father just kept smiling and told the story about
how he cheated on his final math exams by copying from other students and using cheat
sheets. ‘The Peters family,’ he said, ‘has always been a math teacher’s nightmare’. Once
when I was slightly tipsy at a school party, I told this story to Ms. Weber. She said that it was
not a bad excuse, but not a good one either. She said it was just one of a number of excuses
you could come up with to justify being lazy. Last year I barely made it through mathematics,
so I am really nervous about the upcoming school year!”
The descriptions of the problem cases were embedded into the web-based
learning environment, so that the participants could study the problem case
while composing new messages on the web-based discussion boards.
4.1.3 Computer-supported learning environment
All groups collaborated in three web-based discussion boards – one for
each case. The web-based discussion boards provided a main page with an
overview of all message headers. In this overview, answers to original mes-
sages appeared in outline form. The learners could read the full text of all
messages, reply to the messages, or compose and post new messages. In the
replies, the original messages were quoted out with > as in standard news-
readers and e-mail programs.
Prior to collaboration, the randomization of participants was successfully
controlled using individual questionnaires and tests, for example, on prior
knowledge, ambiguity tolerance, and computer experience. Subsequently,
learners were able to study a three page summary of the attribution theory
for 15 minutes. Learners were allowed to make notes and keep the text and
their notes during the collaborative phase. The collaboration time, in which
200 Chapter 12
learners communicated with each other via asynchronous, text- and web-
based discussion boards, was 80 minutes in all experimental groups. All dis-
course activities were recorded within the web-based discussion boards to
collect data on the dimensions of argumentative knowledge construction.
The experimental conditions differed only with respect to the computer-sup-
ported script components that were implemented using the interface of the
computer-supported environment. After collaboration, learners were tested
for individual domain-specific knowledge using individual post-tests similar
to the pre-tests on prior knowledge.
In order to analyze the extent to which the script components influenced
the processes of argumentative knowledge construction, we segmented and
analyzed each of the single arguments the learners put forward in the written
discourse on the epistemic dimension, the argument dimension, and the di-
mension of social modes of co-construction. On the epistemic dimension we
differentiated, for instance, between adequate or inadequate arguments in
terms of the relationships the learners constructed between concepts and case
information. With respect to the argument dimension, we coded the com-
pleteness of arguments according to a model of arguments consisting of
claim, datum with warrant, and qualifier. We also coded the arguments with
regard to their social mode. The interrater reliability was sufficiently high
(for a detailed description of the process analyses see Weinberger & Fischer,
Pre- and post-tests consisted of problem cases comparable to the three
cases learners were asked to analyze during the collaborative phase. The
case analyses the learners needed to produce in the pre- and post-tests were
segmented into units consisting of a theoretical concept applied to problem
case information. In a manner similar to the process analysis, these units
were coded with respect to their adequacy in terms of the relationships
learners constructed between theoretical concepts and case information to
indicate domain-specific knowledge. The adequacy of the individual
learner’s case analyses in the pre- and post-test was determined by their fit to
expert solutions of the problem cases. These expert solutions particularly
stressed the application of multiple perspectives to the cases.
More detailed information on different aspects of the quantitative analy-
ses of the individual empirical studies has been provided in various publica-
tions (e.g., Weinberger, 2003; Weinberger, Ertl, et al., 2005; Weinberger &
12. Scripting argumentative knowledge construction 201
The following experimental groups were examined in Study 1:
The control groups accessed the three distinct web-based discussion
boards of the CSCL environment to read or contribute messages. When
composing a new message, learners were free to choose to start a new dis-
cussion thread or to reply and contribute to an existing discussion thread.
The epistemic script component group could access and contribute to the
web-based discussion boards in a manner identical to the control group, but
whenever a new discussion thread was started, the text window was struc-
tured with prompts of the epistemic script component. These prompts asked
learners to separate relevant from irrelevant case information, structured how
learners applied theoretical concepts to the problem cases, and asked learners
to suggest future developments of the case and pedagogical interventions
(see Figure 12-1).
Case information, which can be explained with the attribution theory
Relevant terms of the attribution theory for this case:
- Does a success or a failure precede this attribution?
- Is the attribution located internally or externally?
- Is the cause for the attribution stable or variable?
- Does the concerned person attribute himself/herself or does another person
Prognosis and consequences from the perspective of the attribution theory:
Case information which cannot be explained with the attribution theory:
Figure 12-1. Prompts of the epistemic script component to apply the concepts of Weiner’s
(1985) attribution theory to problem cases.
The groups with the argumentative script component were provided with
three text windows named claim, datum with warrant and qualifier. Learners
were supposed to collect at least one datum for their claim, explicit the war-
rant for how the datum supports the claim, and provide a qualifier for their
claim by filling out all three text windows (see Figure 12-2). Subsequently,
learners could click an “Add”-Button which displayed the three argument
components in the actual text window of the web-based discussion board.
Learners could add any number of arguments in the main text window.
202 Chapter 12
Figure 12-2. User interface realizing the argumentative script component with four text win-
dows: claim, datum with warrant, qualifier, and message body.
The learners in the social script component groups were assigned two
roles – a) analyst for one of the cases with the task of composing initial and
concluding analyses of the case and responding to critiques, and b) con-
structive critic for the other two cases with the task of repeatedly criticizing
the case analyses. The number of messages was determined by the social
script component (one initial case analysis, two critiques, two replies to the
critiques, two more critiques and one concluding analysis). These messages
given by the social roles were automatically sequenced, that is, learners were
led through each discussion board to submit eight messages in total (see Fig-
ure 12-3). Furthermore, the single messages were supported by prompts
within the text windows such as “These aspects are not clear to me yet”,
“We have not reached consensus concerning these aspects” or “My proposal
for an adjustment of the analysis is”.
12. Scripting argumentative knowledge construction 203
Figure 12-3. Structure of a discourse supported with the social script component with an ini-
tial analysis of the math case, two constructive critiques, two replies of the case analyst to the
critiques, two more critiques and a new analysis of the case by the case analyst.
The epistemic script component reduced the amount of off-topic dis-
course and focused learners’ discourse on just a few new and adequate
knowledge concepts. The learners who were supported by the epistemic
script component contributed more to solving the problem cases than learn-
ers without the script component (see Table 12-3; see also Mäkitalo,
Weinberger, Häkkinen, Järvelä, & Fischer, 2005; Weinberger, 2003;
Weinberger, Ertl, et al., 2005).
Table 12-3. Example of a learner’s message with a case analysis supported with the epistemic
Does success or failure precede this attribution?
Is the attribution located internally or externally?
Is the cause for the attribution stable or variable?
- Michael and his parents: stable causes (talent)
- Teacher: variable causes (effort)
Who is attributing? Self or other?
- His parents
- The teacher
204 Chapter 12
The epistemic script component, however, also affected the argument
dimension and the dimension of social modes of co-construction. Learners
with the epistemic script component constructed less formally complete and
less transactive arguments than learners without the epistemic script compo-
The argumentative script component reduced off-topic discourse and fa-
cilitated the formally adequate construction of single arguments. Learners
with the argumentative script component warranted and qualified their
claims substantially more frequently than learners without this script compo-
nent (see Table 12-4).
Table 12-4. Example of a learner’s message with a case analysis supported with the
argumentative script component.
in this case, it has to be an internal stable attribution
Datum with warrant:
parents say that 1.) the whole family was not “witty” in math and the father adds
that he barely passed his math exam.
internal because talent is to be located within the person and stable because tal-
ent does not change.
i can’t think of any
However, this script component also impeded the content quality of the
single arguments and reduced the adequate application of new knowledge
concepts that were to be learned (see also Weinberger, Stegmann, et al.,
The social script component reduced off-topic discourse and facilitated
the dimension of social modes of co-construction of argumentative knowl-
edge construction. The discourse of groups with the social script component
was more critical and transactive than the discourse of groups of learners
without social script component (see Table 12-5).
Learners supported with this script component operated more on the rea-
soning of their learning partners. Additionally, the social script component
seemed to foster epistemic activities. Learners supported with this script
component engaged more frequently in epistemic activities to solve the
problem case than learners without this script component (see also
Weinberger, 2003; Weinberger, Ertl, et al., 2005 for more detailed process
12. Scripting argumentative knowledge construction 205
Table 12-5. Example of a learner’s critical reply supported by the social script component.
These aspects are not clear to me yet:
What attribution according to attribution theory can be applied? If mother is not
talented - so is the son?
We have not yet reached consensus concerning these aspects:
The teacher does mention that it is only his laziness; she doesn’t explain it to
My proposals for an adjustment of your analysis:
Proposal for a solution: Parents should attend a re-attribution training!
Regarding individual knowledge acquisition, large improvements were
observed between pre- and post-test for learners in all experimental groups,
including the control group. The results further support the notion that the
acquisition of individual domain-specific knowledge can be influenced by
specific script components implemented within CSCL environments. The
epistemic script component impeded the individual acquisition of knowledge
compared to the control group. The argumentative script component did not
facilitate knowledge acquisition beyond the levels of the control group. The
social script component, however, proved to support the individual acquisi-
tion of domain-specific knowledge. After the collaborative learning phase,
learners provided with the social script component were better able to indi-
vidually apply different concepts to problem cases than learners without the
social script component (see Weinberger, Ertl, et al., 2005; Weinberger,
Stegmann, et al., 2005).
The studies investigated the effects of different script components on ar-
gumentative knowledge construction in computer-supported learning envi-
ronments. The learning environments investigated differed only with respect
to the script components, namely (1) the epistemic script component, which
structured how learners handled the task and which concepts they used, (2)
the argumentative script component that asked learners to warrant and qual-
ify their claims, and (3) the social script component that aimed to facilitate
how learners interacted with each other. All computer-supported script com-
ponents substantially reduced off-topic discourse and facilitated the specific
processes of argumentative knowledge construction that they were focusing
on. Based on these findings, all script components seem to have the general
effect of focusing learners on the task. Script components guide and inform
learners of what to do next to solve the task in one way or another. There-
206 Chapter 12
fore, learners seem to have less opportunity to engage in off-topic discourse.
Apart from this general effect, script components can be very specific. Script
components with single goal dimensions can be implemented deliberately
into CSCL environments to address specific shortcomings in the interaction
of groups of learners rather than providing a “one-script-fits-all” model.
The results indicate that epistemic script components help learners to
construct arguments that contribute to solving problem cases, but that learn-
ers do not necessarily benefit from this support with regard to individual
knowledge acquisition. One explanation for this could be that epistemic
script components might not sufficiently support joint elaboration of the
learning material, but rather function as checklists. Thus, epistemic script
components may enable learners to solve the tasks with a limited elaboration
of the learning material. In order to avoid this elaboration-reduction-effect,
epistemic script components may need to be faded out. It may also be neces-
sary in some cases to make collaborative learning tasks harder instead of
simplifying the collaborative learning task in order to facilitate the active
elaboration of the learning material (Palincsar & Herrenkohl, 1999; Reiser,
2002). Furthermore, the degree to which epistemic script components de-
mand the elaboration of learning material or “micromanage” the task may
depend on the prior knowledge of the participants. It may prove unnecessary
to provide epistemic script components to learners with above-average prior
knowledge and skills. Advanced learners may already possess functional
strategies for solving a task and additional epistemic scripting might simply
distract learners from the actual task (see Larson et al., 1985). In order to
avoid this over-scripting effect (Dillenbourg, 2002), epistemic script compo-
nents may need to be carefully matched with the individual prior knowledge
of the participants. Too much or too detailed epistemic scripting may impede
the elaboration of the learning material and the interaction of learners; par-
ticularly when the script oversimplifies the task and divides it into subtasks
that can be worked on by each learner individually (Cohen, 1994).
The argumentative script component, like the other script components,
facilitated the process dimension that it targeted. This study showed that ar-
gumentative script components are able to support argumentative knowledge
construction in both the formal argumentation process dimension during dis-
course and individual knowledge acquisition. Scripting the construction of
arguments may support learners in elaborating the learning material. By con-
structing formally complete arguments with claim, datum, warrant, and
qualifier, learners need to self-explain the learning material, which may fa-
cilitate the acquisition of knowledge (Baker, 2003). Learners supported with
the argumentative script component may have elaborated the learning mate-
rial better than learners without the script. Learners did not always use the
appropriate concepts to solve the task, however, and may thus have elabo-
12. Scripting argumentative knowledge construction 207
rated prior knowledge using misconceptions rather than the knowledge that
was to be learned. Argumentative script components in this way may func-
tion as a thinking tool to amplify elaboration, but fail to prompt learners to
use the relevant knowledge concepts that are to be learned.
The results further indicate that social script components may facilitate
social modes of co-construction, epistemic activities, and the individual ac-
quisition of knowledge. Collaborative learners without support from a social
script component often build a minimal consensus in order to hastily com-
plete collaborative tasks or do not collaborate on the learning task at all. In
contrast, social script components support learners in inquiring about the
contributions of the learning partners more critically and thereby help them
acquire more knowledge individually than learners without additional sup-
port in the dimension of social modes of co-construction (see King, 1999;
Palincsar & Herrenkohl, 1999). This critical approach to the contributions of
the learning partners has also appeared to facilitate the elaboration of the
learning material. One explanation could be that socially scripted learners
engaged in more transactive discussions and appeared to benefit to a greater
extent from the contributions of their learning partners (Teasley, 1997). An-
other explanation is that learners with the social script component elaborated
the learning material to a greater extent, because they anticipated critique
from their learning partners. This explanation is in line with studies that in-
dicate that only the expectation of externalizing knowledge facilitates learn-
ing (Renkl, 1997).
In summary, computer-supported script components can be designed to
facilitate specific process dimensions of argumentative knowledge construc-
tion. Script components that “micromanage” discourse on an epistemic di-
mension may cause learners to focus on solving the task at hand without
elaborating the learning material. In order to foster the elaboration of the
learning material and individual knowledge acquisition, script components
may need to target not only the epistemic activities, but also focus on social
modes of co-construction in argumentative discourse (see Herrenkohl &
Guerra, 1998; Palincsar & Herrenkohl, 1999). Conversely, script compo-
nents that are aimed at formal aspects of argument construction without ad-
ditionally fostering epistemic activities or social modes of co-construction
may not be able to help learners achieve better results than without support
from a script. Content-independent argumentative script components may
aid elaboration, but hold the danger that learners may not be able to select
the appropriate concepts that are supposed to be elaborated. The social script
component of this study, in contrast, managed to not only facilitate transac-
tive discourse, but also supported the epistemic activities of learners. This
indicates that transactivity can be essential to argumentative knowledge con-
208 Chapter 12
struction and can be facilitated beyond the levels that collaborative learners
would spontaneously achieve (Teasley, 1997).
7. FUTURE RESEARCH
CSCL environments offer a suitable context for scripting the interaction
of learners. Clearly, there is further need to examine beneficial interactions
of script components for CSCL, for example, investigating epistemic script
components which do not micromanage interaction of learners in combina-
tion with social and argumentative script components (see Ertl, Kopp, &
Mandl, this volume). Individual cognitive processes and their relationship to
various process dimensions and outcomes of argumentative knowledge con-
struction may explain to a greater extent how learners benefit from argu-
mentative knowledge construction scenarios. Text-based CSCL may provide
a unique opportunity for investigating the cognitive processes of learners.
While engaging in written communication, learners may simultaneously
provide information about their cognitive activities through think-aloud-
techniques. We also need to better understand how scripts interact with
learners’ prior knowledge and skills, which may be represented as internal
scripts in contrast to external, computer-supported scripts (Carmien, Kollar,
G. Fischer, & F. Fischer, this volume). Therefore, an important question for
future research of CSCL environments is how scripts can be designed not to
substitute, but to facilitate discourse and cognitive activities related to indi-
vidual acquisition of knowledge. In these terms, we need to further investi-
gate the interaction of different script components that may be adapted to the
already existing internal scripts. Design environments need to be developed
in order to improve the impact of computer-supported script research in
practice at schools and universities. These environments should facilitate the
adaptive combination of script components with different representations
that can be used relatively independent of the computer support available in
The studies have been funded by the Deutsche Forschungsgemeinschaft
(DFG) and the Kaleidoscope Network of Excellence within a European re-
search team entitled “Computer-Supported Scripting of Interaction in Col-
laborative Learning Environments”.
12. Scripting argumentative knowledge construction 209
Andriessen, J. E. B., Baker, M., & Suthers, D. (Eds.). (2003). Arguing to learn. Confronting
cognitions in computer-supported collaborative learning environments
. Dordrecht: Klu-
Baker, M. (2003). Computer-mediated argumentative interactions for the co-elaboration of
scientific notions. In J. Andriessen, M. Baker, & D. Suthers (Eds.),
Arguing to learn: con-
fronting cognitions in computer-supported collaborative learning environments
pp. 1-25). Dordrecht: Kluwer.
Baker, M., & Lund, K. (1997). Promoting reflective interactions in a CSCL environment.
Journal of Computer Assisted Learning, 13, 175-193.
Brown, A. L., & Palincsar, A. S. (1989). Guided, cooperative learning and individual knowl-
edge construction. In L. B. Resnick (Ed.),
Knowing, learning, and instruction. Essays in
the honour of Robert Glaser
(pp. 393-451). Hillsdale, NJ: Lawrence Erlbaum Associates.
Chi, M. T. H., Bassok, M., Lewis, M. W., Reimann, P., & Glaser, R. (1989). Self-explana-
tions: How students study and use examples in learning to solve problems.
Cohen, E. G. (1994). Restructuring the classroom: Conditions for productive small groups.
Review of Educational Research, 64, 1-35.
Dillenbourg, P. (2002). Over-scripting CSCL: The risks of blending collaborative learning
with instructional design. In P. A. Kirschner (Ed.),
Three worlds of CSCL: Can we support
(pp. 61-91). Heerlen: Open Universiteit Nederland.
Dufresne, R. J., Gerace, W. J., Thibodeau Hardiman, P., & Mestre, J. P. (1992). Constraining
novices to perform expertlike problem analyses: Effects on schema acquisition.
nal of the Learning Sciences, 2
Fischer, F., Bruhn, J., Gräsel, C., & Mandl, H. (2002). Fostering collaborative knowledge
construction with visualization tools.
Learning and Instruction, 12, 213-232.
Herrenkohl, L. R., & Guerra, M. R. (1998). Participant structures, scientific discourse, and
student engagement in fourth grade.
Cognition and Instruction, 16, 433-475.
Jeong, H., & Chi, M. T. H. (1999).
Constructing shared knowledge during collaboration and
Paper presented at the AERA Annual Meeting, Montreal, Canada.
King, A. (1999). Discourse patterns for mediating peer learning. In A. M. O’Donnell & A.
Cognitive perspectives on peer learning (pp. 87-115). Mahwah, NJ: Law-
rence Erlbaum Associates
Kollar, I., Fischer, F., & Slotta, J. D. (2005). Internal and external collaboration scripts in
web-based science learning at schools. In T. Koschman, D. Suthers, & T.W. Chan (Eds.),
Computer Supported Collaborative Learning 2005: The Next 10 Years! (pp. 331-340).
Mahwah, NJ: Lawrence Erlbaum Associates.
Kuhn, D., Shaw, V., & Felton, M. (1997). Effects of dyadic interaction on argumentative
Cognition and Instruction, 15(3), 287-315.
Larson, C. O., Dansereau, D. F., O’Donnell, A. M., Hytecker, V. I., Lambiotte, J. G., &
Rocklin, T. R. (1985). Effects of metacognitive and elaborative activity on cooperative
learning and transfer.
Contemporary Educational Psychology, 10, 342-348.
Leitão, S. (2000). The potential of argument in knowledge building.
Human Development, 43,
Mäkitalo, K., Weinberger, A., Häkkinen, P., Järvelä, S., & Fischer, F. (2005). Epistemic
cooperation scripts in online learning environments: Fostering learning by reducing
uncertainty in discourse?
Computers in Human Behavior, 21(4), 603-622.
Mandl, H., Gruber, H., & Renkl, A. (1996). Communities of practice toward expertise: Social
foundation of university instruction. In P. B. Baltes & U. Staudinger (Eds.),
210 Chapter 12
minds. Life-span perspectives on the social foundation of cognition (pp. 394-411). Cam-
bridge, MA: Cambridge University Press.
Marttunen, M., & Laurinen, L. (2001). Learning of argumentation skills in networked and
Instructional Science, 29, 127-153.
Means, M. L., & Voss, J. F. (1996). Who reasons well? Two studies of informal reasoning
among children of different grade, ability and knowledge levels.
Cognition and Instruc-
O’Donnell, A. N., & Dansereau, D. F. (1992). Scripted cooperation in student dyads: A
method for analyzing and enhancing academic learning and performance. In R. Hertz-
Lazarowitz & N. Miller (Eds.),
Interactions in cooperative groups. The theoretical anat-
omy of group learning
(pp. 120-141). Cambridge, MA: Cambridge University Press.
Palincsar, A. S., & Herrenkohl, L. R. (1999). Designing collaborative contexts: Lessons from
three research programs. In A. M. O’Donnell & A. King (Eds.),
Cognitive perspectives on
(pp. 151-177). Mahwah, NJ: Lawrence Erlbaum Associates.
Pea, R. D. (1994). Seeing what we build together: Distributed multimedia learning environ-
ments for transformative communications. Special Issue: Computer support for collabora-
Journal of the Learning Sciences, 3(3), 285-299.
Reiser, B. J. (2002). Why scaffolding should sometimes make tasks more difficult for learn-
ers. In G. Stahl (Ed.),
Proceedings of the Computer Support for Collaborative Learning:
Foundations for a CSCL Community - CSCL 2002
(pp. 255-264). Boulder, CO: Lawrence
Renkl, A. (1997).
Lernen durch Lehren - Zentrale Wirkmechanismen beim kooperativen
Lernen [Learning through teaching - central mechanisms in cooperative learning]
baden: Deutscher Universitäts-Verlag.
Scardamalia, M., & Bereiter, C. (1996). Computer support for knowledge-building communi-
ties. In T. Koschmann (Ed.),
CSCL: Theory and practice of an emerging paradigm (pp.
249-268). Mahwah, NJ: Lawrence Erlbaum Associates.
Teasley, S. (1997). Talking about reasoning: How important is the peer in peer collaboration?
In L. B. Resnick, R. Säljö, C. Pontecorvo, & B. Burge (Eds.),
Discourse, tools and reason-
ing: Essays on situated cognition
(pp. 361-384). Berlin: Springer.
Toulmin, S. (1958).
The uses of argument. Cambridge, MA: Cambridge University Press.
Voss, J. F., & van Dyke, J. A. (2001). Argumentation in psychology: Background comments.
Discourse Processes, 32(2&3), 89-111.
Weinberger, A. (2003).
Scripts for computer-supported collaborative learning. Effects of
social and epistemic collaboration scripts on collaborative knowledge construction.
toral thesis. Ludwig-Maximilian University, Munich. Available at: http://edoc.ub.uni-
Weinberger, A., Ertl, B., Fischer, F., & Mandl, H. (2005). Epistemic and social scripts in
computer-supported collaborative learning.
Instructional Science, 33(1), 1-30.
Weinberger, A., & Fischer, F. (2006). A framework to analyze argumentative knowledge
construction in computer-supported collaborative learning.
Computers & Education, 46,
Weinberger, A., Reiserer, M., Ertl, B., Fischer, F., & Mandl, H. (2005). Facilitating collabo-
rative knowledge construction in computer-mediated learning with cooperation scripts. In
R. Bromme, F. Hesse, & H. Spada (Eds.),
Barriers and biases in computer-mediated
knowledge communication - and how they may be overcome
(pp. 15-37). Boston: Kluwer.
Weinberger, A., Stegmann, K., & Fischer, F. (2005). Computer-supported collaborative
learning in higher education: Scripts for argumentative knowledge construction in distrib-
uted groups. In T. Koschmann, D. Suthers, & T. W. Chan (Eds.),
12. Scripting argumentative knowledge construction 211
Collaborative Learning 2005: The Next 10 Years! (pp. 717-726). Mahwah, NJ: Lawrence
Weiner, B. (1985). An attributional theory of achievement motivation and emotion.
logical Review, 92