Content uploaded by Dietrich Albert
Author content
All content in this area was uploaded by Dietrich Albert
Content may be subject to copyright.
Int. J. Knowledge and Learning, Vol. 4, Nos. 2/3, 2008 18
9
Copyright © 2008 Inderscience Enterprises Ltd.
Activity- and taxonomy-based knowledge
representation framework
Birgit Marte and Christina M. Steiner*
Cognitive Science Section
Department of Psychology
University of Graz
Universitaetsplatz 2, 8010 Graz, Austria
Fax: +43 316 380–9806
E-mail: birgit.marte@uni-graz.at
E-mail: chr.steiner@uni-graz.at
*Corresponding author
Juergen Heller
Psychological Institute
University of Tuebingen
Friedrichstraße 21, 72072 Tuebingen, Germany
Fax: +49 7071 29–3363
E-mail: juergen.heller@uni-tuebingen.de
Dietrich Albert
Cognitive Science Section
Department of Psychology
University of Graz
Universitaetsplatz 2, 8010 Graz, Austria
Fax: +43 316 380–9806
E-mail: dietrich.albert@uni-graz.at
Abstract: Elaborations of Competence-based Knowledge Space Theory
(CbKST) incorporate skills that refer to the conceptual information of the
domain as well as to the activities learners are expected to perform in this
context. Thus, they are suggested as a formal knowledge representation
framework that is able to take into account current activity-oriented
pedagogical trends in designing effective Units of Learning (UoL). The broad
array of required behaviour to be achieved by learners demands a search for
instruments like taxonomies that allow for conceptualising activities, and
consequently, skills and learning objectives. It is shown that the availability of
such a taxonomy-based framework may be utilised in order to enhance the
access and interface functionalities of learning systems. In particular, the
selection of proper learning units and the delivery of effective feedback
mechanisms on the teaching and learning progress are facilitated.
Keywords: knowledge representation; knowledge space theory; skills and
competences; learning objectives; learning activities; taxonomies.
190 B. Marte, C.M. Steiner, J. Heller and D. Albert
Reference to this paper should be made as follows: Marte, B., Steiner, C.M.,
Heller, J. and Albert, D. (2008) ‘Activity- and taxonomy-based knowledge
representation framework’, Int. J. Knowledge and Learning, Vol. 4, Nos. 2/3,
pp.189–202.
Biographical notes: Birgit Marte has a Diploma in Psychology from the
University of Graz (Austria). From April 2004 to April 2008 she was working
at the Cognitive Science Section (http://css.uni-graz.at) of the University
of Graz on several European Commission-funded R&D projects that focused
on e-learning. Her research addresses the representation and assessment
of knowledge and competences. Her recent research interests include the
evaluation of the effectiveness of e-learning, especially the evaluation of
game-based learning based on sound psychological methods and techniques.
Christina M. Steiner completed her Diploma (MS) in Psychology at the
University of Graz. She is currently a Researcher at the Cognitive Science
Section of the University of Graz. Her work within several European R&D
projects on e-learning focuses on the representation, modelling and assessment
of knowledge and competence. In particular, she is doing research on the use of
concept maps as a means to build prerequisite structures among learning
objects and competences, and on the validation of concept maps and their
application as a learning and teaching strategy.
Juergen Heller received his Diploma in Psychology, his PhD (Major:
Psychology, Minor: Mathematics) and his postdoctoral degree (Habilitation in
Psychology) from the University of Regensburg (Germany). Since 2006 he
is holding a Full Professorship in methods of psychological research at
the University of Tuebingen (Germany). His research interests are mainly
centred around developing psychological theories, psychological modelling
and measurement in various areas of cognitive psychology and perception.
His recent work addresses probabilistic knowledge structures, individual
semantic structures, lightness and brightness perception, as well as binocular
space perception.
Dietrich Albert has a Diploma in Psychology from the University of Goettingen
(Germany), and received his PhD and his Habilitation in Psychology at the
University of Marburg/Lahn (Germany). Since 1993 he has been a Professor of
Psychology at the University of Graz (Austria) and Head of the Cognitive
Science Section at the Department of Psychology. His current research focus is
on knowledge and competence structures, especially competence-based
theories of knowledge spaces, their applications and empirical research. He and
his team are currently involved in several European e-learning R&D projects
(e.g., iClass, ELEKTRA).
1 Introduction
The learning effectiveness of a virtual learning environment depends to a large extent on
how well the knowledge represented in the system is able to match the knowledge of a
domain and of individual learners. At the first European Learning Grid Infrastructure
(EleGI) conference in March 2005, we introduced the basic ideas of Competence-based
Knowledge Space Theory (CbKST) (Heller et al., 2005), which directly refers to the
Activity- and taxonomy-based knowledge representation framework 191
underlying skills required for solving problems or that are taught by learning objects.
Moreover, their application in the context of distributed resources and Virtual Scientific
Experiments (VSE) was discussed.
The present paper outlines how skills in terms of CbKST can be elaborated to adopt a
more activity-oriented and learner-centred perspective which is currently predominant in
the educational field. To this end, domain and learner knowledge are represented via
skills that are characterised by both conceptual information and information related to the
learning activities and objectives. Finally, the implications of using taxonomies that may
underlie such a skill representation for enhancing the access and interface functionalities
of learning systems (e.g., within ELeGI) are presented.
2 Learning objectives
In the design of Units of Learning (UoL, e.g., a lesson, a course), learning objectives play
an essential role. Since recent developments in the educational field focus more on the
teaching and learning processes than on the pure learning content, they are usually stated
at a more fine-grained level than just at the conceptual dimension, as commonly practised
in existing learning environments. Thus, learning objectives have to precisely specify the
skills and competences that need to be acquired by the learner.
2.1 On the role of activity-oriented learning objectives
Instead of the traditional approach of directing instruction to the transmission of
knowledge and defining objectives in terms of content to be learnt, learner-centred
instruction acknowledges what the learner does.
Hence, learning objectives are formulated to express the intended learning outcome
(e.g., skills and competences) and what the learners will be able to do as a result of the
instruction (Anderson et al., 2001). Teaching consists in providing appropriate strategies
that enable learners to achieve the respective knowledge and skills. The activity-oriented
approach is also considered in the development of current standards, such as the IMS
Learning Design (for details refer for example, to IMS Learning Design Best Practice and
Implementation Guide, 2003).
In literature, much has been written about the nature of learning objectives. For
example, Tyler (1949) suggested phrasing learning objectives in terms of the behaviour
to be developed and the content in which this behaviour is to be operated. Another
influential approach relies on three components: behaviour, standard and conditions
(Mager, 1962; 1984). The behaviour describes what the learner is expected to be able to
do and is expressed by specific verbs such as ‘calculate’ and ‘differentiate’. Via the
standard (also criterion), the acceptable level of performance is determined, for example,
by a certain proportion of correct answers. The conditions indicate the constraints (e.g.,
available resources) under which the learner will perform in the learning situation.
According to Krathwohl and Payne (1971), the specificity of learning objectives can
vary from rather global to more specific objectives. They differentiate between global,
educational and instructional objectives. Global objectives (as in curricula, for example)
are identified with rather broadly stated learning outcomes, such as ‘The student shall
develop the fundamental skills of reading and writing.’ More specific curricula can be
defined at the level of educational objectives, which can be used by the teacher to plan
192 B. Marte, C.M. Steiner, J. Heller and D. Albert
their classroom activities for a longer period, for example, ‘can name and recognise the
letters of the alphabet’. In order to plan daily activities, still more specific instructional
objectives are needed that focus on narrow teaching and learning in specific domains
(e.g., ‘ability to discriminate between pairs of similar and easily confused letters, e.g.,
P from R’). These are in line with Tyler’s (1949) specification. Such specific
instructional objectives can be identified with the skills and competences the students are
going to achieve during a UoL. They can even be characterised by single learning objects
in a virtual learning environment.
Including activity-related information into the specification of learning objectives
emphasises the fact that content alone is only part of what is to be learnt. Many possible
activities may be related to each content item. For example, a student may be requested to
explain a concept or to apply it in a specific context. Thus, it is important that learning
objectives refer to both – information on the content and required activities.
Note that learning objectives define what the learner is going to achieve, not how the
learner learns and accomplishes them. The way to achieve these objectives (i.e., learning
activities) is mainly determined by the respective pedagogical approach the teacher relies
on. In order to differentiate better between learning activities (means) and learning
objectives (targeted end states), Anderson et al. (2001) suggest indicating learning
objectives by the phrase ‘to be able to’ or ‘learn to’.
To conclude, learning objectives are crucial for both the design process of effective
teaching, on the one hand, and the assessment of the learning outcome (i.e., skills and
competences), on the other hand. The more specific these objectives are, the easier one
can assess their achievement. The alignment between clearly stated learning objectives
and learning activities rooted in the respective pedagogical background facilitates
learning as well as its assessment, and hence the identification of the skills and
competences gained by the learners.
2.2 Taxonomies for learning objectives
The use of learning objectives for designing and assessing learning is complicated by the
availability of a broad range of possible activities that may be part of an objective. Thus,
there is a need for a framework according to which learning objectives can be organised
in order to increase their manageability. An instrument like a taxonomy can be used for
planning instruction, learning and assessment.
A number of approaches have been devised for the classification of learning
objectives and activities. The most popular taxonomy for the educational practice was
developed by Bloom (1956), whose influential framework was later updated by Anderson
et al. (2001). Bloom and his colleagues classified intended behaviours related to mental
acts or thinking that occurred as a result of educational experiences. The purpose of this
taxonomy was to enhance the exchange and communication of ideas and material by
using a common language for educational objectives. The taxonomy should also serve
as a tool for determining the congruence between objectives, learning activities
and assessments related to a lesson, course, etc., as well as for revealing the array
of possible instructional options. The taxonomy comprises six categories: knowledge,
comprehension, application, analysis, synthesis and evaluation. These categories
characterise different levels of cognitive processing and are assumed to form a
Activity- and taxonomy-based knowledge representation framework 19
3
cumulative hierarchy. Bloom’s taxonomy has been used worldwide for designing
educational instruction and has influenced many other researchers (e.g., Ausubel and
Robinson, 1969; Gagné, 1985; Marzano, 2001; Anderson et al., 2001).
The revision of Bloom’s taxonomy by Anderson et al. (2001) retains the six cognitive
process categories. However, while the original framework uses nouns (e.g., analysis), in
the revised taxonomy verbs (e.g., analyse) label the different categories, reflecting the
prevalent activity-centred approach in teaching and learning. Moreover, Bloom’s
category knowledge was renamed to remember, comprehension to understand, and
synthesis to create. Additionally, the order of the last two categories was reversed (see
Table 1). Anderson et al. (2001) also introduced a second dimension, which lies along a
continuum from concrete to abstract knowledge. It consists of four categories: factual,
conceptual, procedural and metacognitive knowledge. The revised and activity-oriented
version of the taxonomy remains hierarchical in overall complexity, too. It is also more
applicable for the planning of educational instruction and assessment because the two
dimensions, knowledge and cognitive process, displayed in a table form a useful
representation of any UoL (Krathwohl, 2002).
Table 1 The six categories of cognitive processes
Process categories Description and examples
Remember Retrieve relevant knowledge from long-term memory, e.g., reorganise dates
of important events in US history
Understand Construct meaning from instructional messages, including oral, written,
graphical communication, e.g., classify mental disorders
Apply Carry out or use a procedure in a given situation, e.g., divide one whole
number by another whole number
Analyse Break material down into constituent parts and determine how parts relate to
one another and to an overall structure, e.g., differentiate between
(ir)relevant numbers in a word problem
Evaluate Make judgements based on criteria and standards, e.g., judge which of two
methods is the best way to solve a given problem
Create Put elements together to form a coherent or functional whole; reorganise
elements into a new pattern or structure, e.g., plan a research paper on a
given historical topic
Source: Adapted from Anderson et al. (2001)
A more recent approach for classifying learning outcomes is the Structure of the
Observed Learning Outcome (SOLO) taxonomy developed by Biggs and Collis (1982;
Biggs, 1999). This framework is primarily an assessment tool looking at the structure of
the observed learning outcome. Its purpose is to provide a systematic way of describing
how a learner’s performance grows in complexity when mastering a range of tasks. It can
be used to define objectives that describe performance goals or targets, as well as to
evaluate the level of learning outcomes.
Still another approach for classifying learning activities was devised by Vermunt and
Verloop (1999), who built their framework around cognitive, affective and metacognitive
(regulative) dimensions. Since the categories of the cognitive process dimensions they
distinguish are neither exhaustive nor mutually exclusive according to the authors, it can
be seen as a framework rather than a taxonomy.
194 B. Marte, C.M. Steiner, J. Heller and D. Albert
Since there are various frameworks and taxonomies available that focus on different
aspects of learning and teaching, one has to define the criteria according to which the
most appropriate approach can be selected, given particular interests and purposes.
A comprehensive review and evaluation of existing frameworks for teaching, learning
and thinking skills is provided in a report by Mosely et al. (2004). A hierarchical
categorisation such as Bloom’s revised taxonomy (Anderson et al., 2001) seems to be the
most appropriate for the purpose presented in this paper.
3 Knowledge space theory and competence-based extensions
CbKST is a knowledge representation model that is able to incorporate the
activity-oriented understanding of teaching and learning. After outlining the basic notions
of Knowledge Space Theory and its competence-based extensions, recent activity – and
taxonomy-based considerations are presented.
3.1 Basic notions of knowledge space theory
Knowledge Space Theory (Doignon and Falmagne, 1985; 1999) provides a set-theoretic
framework for representing the knowledge of a learner in a certain domain, which is
characterised by a set of problems (subsequently denoted by Q). The knowledge state of
an individual is identified with the subset of problems this person is capable of solving.
Due to mutual dependencies between the problems, which are captured by a so-called
prerequisite relation, not all potential knowledge states will actually be observed.
Any prerequisite relation can be illustrated by the so-called Hasse diagram, where the
relation is depicted by ascending sequences of line segments. According to the diagram
shown in Figure 1(a), for example, problems b and d are in a prerequisite relation, which
means that problem b is a prerequisite to the solution of problem d. The collection of
possible knowledge states of a given domain Q is called a knowledge structure, whenever
it contains the empty set Ø and the whole set Q. The knowledge structure induced by the
prerequisite relation depicted in Figure 1(a) given by:
K = {Ø, {a}, {b}, {a, b}, {b, d}, {a, b, c}, {a, b, d}, {a, b, c, d}, Q}.
The resulting order on this collection of knowledge states is based on set-inclusion and is
shown in Figure 1(b). A knowledge structure offers a range of possible learning paths
from the naive knowledge state to the expert knowledge state. Besides offering
personalised learning paths dependent on the knowledge state of an individual, a
knowledge structure is at the core of an efficient adaptive procedure for knowledge
assessment. It allows for uniquely determining the knowledge state by presenting the
learner with only a subset of the problems.
Activity- and taxonomy-based knowledge representation framework 19
5
Figure 1 Example of a prerequisite relation on a set of problems Q illustrated as a Hasse
diagram (a) and corresponding knowledge structure (b) with the dashed arrows
representing a possible learning path
3.2 Competence-based knowledge space theory
The original Knowledge Space Theory exclusively focuses on the observable solution
behaviour of learners and does not refer to skills that are required for solving the
problems or that are taught by learning objects. Since these issues are of special interest
for practical application in educational settings, CbKST explicitly refers to learning
objects as well as skills and competencies. The following considerations are based on
work by Falmagne et al. (1990), Doignon (1994), Düntsch and Gediga (1995), Korossy
(1997; 1999), Albert and Held (1994; 1999), Hockemeyer (2003) and Hockemeyer et al.
(2003). The basic assumption is the existence of a set of skills that are relevant for
solving the problems of a particular domain, and that are taught by the learning objects of
the respective domain.
In CbKST, skills are assigned to both the problems and learning objects of a
knowledge domain. Note that skills are meant to provide a fine-grained, low-level
description of students’ capabilities. Generally, these assignments represent the
assignment of (semantic) metadata to the problems and learning objects. The relation
between assessment problems and skills is realised by two mappings. The mapping s
(skill function) associates with each problem a collection of subsets of skills. Each of
these subsets (i.e., each competence) consists of those skills that are sufficient for solving
the problem. Assigning more than one competence to a problem takes care of the fact that
there may be more than one way to solve it. The mapping p (problem function) associates
{b}
{a}
{a,b}{b,d}
{a,b,d}
{a,b,c}
{a,b,c,d}
{a,b,c,d,e}
Ø
a
c
e
d
b
(a) (b)
196 B. Marte, C.M. Steiner, J. Heller and D. Albert
with each subset of skills the set of problems that can be solved with it. It defines a
knowledge structure K because the associated subsets actually are nothing else but the
possible knowledge states (for an example, see Heller et al., 2005).
The association of skills with the problems of a domain allows uncovering a learner’s
skills in the frame of an efficient assessment. The collection of skills a person has
available is then called the competence state of this individual. It is not directly
observable but can be inferred on the basis of the knowledge state. Once the knowledge
and competence state of an individual is identified, low-level learning objectives may
indicate the targeted skills to be learned next. To bridge the gap between the actual
knowledge and competence state and the targeted learning objectives, in CbKST skills
are also associated with the learning objects of a domain. This relationship is mediated by
two mappings. The mapping r associates with each learning object a subset of skills
(required skills), which characterise the prerequisites for dealing with it, or understanding
it. The mapping t associates with each learning object a subset of skills (taught skills),
which refer to the content actually taught by the learning objects. Given the competence
state of a learner, personalised learning paths can be built that teach the skills this learner
is ready to learn next.
A further extension is to assume dependencies between the skills (e.g., Korossy,
1999), inducing a competence structure on the set of skills. A competence structure
further restricts the number of possible knowledge states that can occur and may be
explicitly established by identifying relationships between skills, for example, by
querying experts, or indirectly via the assignments of skills to the problems of a domain
as described above. Obviously, pedagogical aspects, e.g., curriculum frameworks,
educational standards and learning objectives, also have to be taken into account when
building a competence structure.
3.3 Activity-based skill characterisation
As mentioned above, fine-grained learning objectives can be identified with the skills and
competences learners are expected to achieve as a result of a UoL. Since learning
objectives refer to both the information on the content and the required activities, the
representation of skills has to reflect this characterisation, too. Thus, we suggest
characterising a skill as a pair consisting of a concept and an activity (e.g., apply
Pythagorean Theorem). Both entities hold structural information, which has to be
combined in order to derive a structure of the set of skills (Heller et al., 2006).
The concepts for the skill definition may be derived from a domain ontology.
Common tools for representing the ontological information of a domain are concept
maps, which depict structural relations between the basic concepts. For example, a
concept map may illustrate that the concept (c2) Pythagorean Theorem is a prerequisite to
the concept (c1) Altitude Theorem. This induces an order on the set of concepts, which
can be graphically represented as in Figure 2(a).
The relations between activities associated with the respective skills may be based on
a proper taxonomy, according to which they can also be organised. The revised taxonomy
by Anderson et al. (2001) seems to be appropriate for at least two reasons. First, the
cognitive process dimension focuses directly on learning and required cognitive efforts.
Second, the hierarchical structure of the framework provides the desired information on
the relation between given activities. For example, the prerequisite relation between an
activity (a2) state (a certain theorem) and (a1) apply (a certain theorem) may be derived
Activity- and taxonomy-based knowledge representation framework 19
7
by contrasting these verbs with the categories of the taxonomy. In doing this, it may be
revealed that state can be associated with the category remember and apply with the
category apply. Since, according to the structure of the taxonomy, remember is a
prerequisite to apply, this information can be adopted for the relation between state (a2)
and apply (a1), as can be seen from the graphical representation shown in Figure 2(b).
Figure 2 Concept structure (a) and structure on activities (b)
Obviously, the principle for classifying activities (verbs) into categories or levels of the
cognitive process dimension is not a straightforward procedure. It requires the definition
of clear features that characterise each category, possibly enabling an automatic
assignment of activities to appropriate categories. Further research is required on such
issues as how to determine such features.
Another crucial question is how the structures on the concepts and activities can be
merged in order to derive the competence structure. We suggest resolving this issue by
using the component-attribute approach (Albert and Held, 1994; 1999). Within this
approach, components are understood as dimensions, and attributes are the different
values the dimensions can take on. For the given context, the set C of concepts and the
set A of activities can be seen as the components, and the attributes can be identified with
the different elements (e.g., c1, c2 in C and a1, a2 in A). For each component a relation is
defined according to which the attributes are ordered. The structure of the set of skills is
then built by forming the direct product of the components, inducing a prerequisite
relation on the Cartesian Product C x A by component-wise ordering. The product of
the two graphs shown in Figure 2 is the structure illustrated in Figure 3. From this it
can be read off, for example, that the skill (c2a2) is a prerequisite to skills (c2a1), (c1a1)
and (c1a2).
c1
c2
c4
c3 a1
a2
(a) (b)
198 B. Marte, C.M. Steiner, J. Heller and D. Albert
Figure 3 Example of the prerequisite relation induced by the structures shown in Figure 2
If learning objectives are formulated at the level of skills, and skills are associated with
the problems and learning objects of a knowledge domain, it can easily be assessed
whether the respective learning objectives have been achieved by learners. This may
provide the basis for identifying the gaps that are to be filled, by devising a personalised
learning path. Moreover, the association of skills to a sound taxonomy may also provide a
method for describing skills in an aggregated form. In the next section the practical
implications for using taxonomies and a skill characterisation, as outlined above, within a
virtual learning environment will be outlined.
4 Application of an activity- and taxonomy-based skill representation
The relation of skills to conceptual information and required activities, as well as to
the concrete learning objects and assessment problems that make up a learning
system, can be utilised in order to enhance the access and interface functionalities of
e-learning applications. In fact, the skills’ association to taxonomies of cognitive
processes facilitates the selection of a proper UoL and the delivery of effective
feedback mechanisms.
4.1 Access modalities: building units of learning
In virtual environments, the options for selecting a UoL are commonly limited to the
conceptual level. CbKST can provide additional options for defining a learning unit that
are driven by the above-introduced skill assignments.
First of all, the learner or teacher may choose a certain skill to be taught. In this case
there are two options for building a UoL. First, based on the selected skill, the set of
learning objects that actually teach that skill can be identified. Then it has to be checked
(c2a2)
(c2a1)
(c1a2)
(c1a1)
(c3a1)
(c3a2)
(c4a1)
(c4a2)
Activity- and taxonomy-based knowledge representation framework 19
9
whether the learner has already acquired the skills needed for understanding these
learning objects. If this is not the case, the learning objects that teach these required skills
have to be selected. This procedure recursively builds up a structure of learning objects,
and thus a collection of possible learning paths. This procedure continues until the
required skills in at least one of the resulting paths match the learner’s competence. The
second way of producing a UoL is to identify the respective concepts associated with
the chosen skill. Which option actually is selected depends, for example, on didactical
considerations (e.g., the type of knowledge that is to be learnt).
Due to the broad range of possible skills and learning objectives, respectively, their
identification for building a UoL may be facilitated by the availability of a taxonomy as
described in Section 2.2. Based on such a framework, teachers or even learners may
simply choose the level of the skills and objectives with respect to specific concepts
instead of defining particular skills. The system then has to search for available skills and
associated learning objects or concepts to build up a UoL.
4.2 Interface modalities: reporting mechanism
A taxonomy underlying the skills and learning objectives may be useful for teachers
and learners in getting an overview of either the spectrum of their teaching and/or
their learning progress. These considerations are mostly in line with the purpose for
which Bloom (1956), and later Anderson et al. (2001), introduced the taxonomy of
educational objectives.
A so-called reporting tool may inform the teacher about the range of skills and
objectives with respect to certain concepts that they already covered in the prepared UoL.
It may quote the levels of skills the teacher has never considered as well as dominant skill
levels. Such a report may foster a teacher’s metacognition about his or her teaching
strategies and may inspire the teacher to build an effective UoL covering a broad range of
possible learning objectives and skills. Via a taxonomy, it would also be possible to
notify the teacher in an aggregated way about the skills related to the particular concepts
the learners have achieved so far. This may also influence the selection of skill levels for
designing subsequent instructional units.
Via an underlying taxonomy, the learner could also be notified about the amount
of achieved skills and competences in an efficient way, promoting the learner’s
metacognitive reflection on his or her learning progress. In case that the learner uses
the learning system to edit new learning units, he or she may also be notified about the
range of disregarded and preferred skill levels, in order to consider this information for
the next time.
For both teachers and learners, a concise way to deliver and represent the feedback
may be to visualise the report information on achieved and/or covered skill levels with
respect to certain concepts. Two kinds of information are of interest. Since the total sum
of skills contained in each category will likely vary, first, it is relevant to know the
amount of skills contained in each category. Second, for the teacher or learner the number
of actually achieved and/or covered skills per category or level is of significance.
200 B. Marte, C.M. Steiner, J. Heller and D. Albert
Figure 4 Example of a possible visualisation (pie chart) for reporting covered or achieved skills
per taxonomy level for a particular UoL. The whole UoL comprises 28 skills. The
fractions next to each slice indicate how many of the available skills per category have
already been achieved and/or covered
For illustrating this information at a glance, histograms or pie charts seem to be
favourable. In the diagram shown in Figure 4, each slice corresponds to one category of
the taxonomy (i.e., skill level, remember, understand, apply, etc.). The fraction per slice
shows how many of the available skills per category have already been covered by a
teacher or achieved by a learner. For example, according to Figure 4, a learner has
achieved six out of a total of eight skills that are associated with the first level (i.e.,
remember) of the taxonomy and three out of the five skills that are contained in the
category apply. The information on the skills can be provided either for a particular
concept, closely interrelated concepts, or even a whole UoL. In each case the aim is
to increase the number of covered and/or achieved skills for each skill level of
the taxonomy.
5 Conclusion and implications
The paper shows that learner and content knowledge modelling based on CbKST is able
to take into account current pedagogical trends in designing effective UoL, by
incorporating skills that refer to the conceptual information of the domain as well as to
the activities learners are expected to perform in this context. Via the component-attribute
approach, it is possible to merge these two kinds of structural information into a unified
representation of the skill structure. This type of skill characterisation is further in line
with the definition of fine-grained learning objectives for building narrow instructional
units. The broad array of possible verbs that express the required behaviour to be
achieved by learners requires a search for mechanisms that allow for conceptualising
activities and consequently, skills and learning objectives. The revision of Bloom’s
remember
understand
apply
analyse
evaluate
create
6/8
0/2
1/4
1/2
3/5
5
/
7
Activity- and taxonomy-based knowledge representation framework 201
taxonomy by Anderson et al. (2001) is suggested as a tool for organising and structuring
learning objectives and skills. It was shown that the availability of such a framework
may facilitate building a UoL (see Section 4.1), as well as the reporting on the
teaching and learning progress (see Section 4.2) with respect to covered and achieved
skills, respectively.
However, there is a need for further research. Principles have to be elicited according
to which activities can be associated with one of the six levels of cognitive processing.
The aim is to find features for each category in order to be able to differentiate between
them. Based on such characteristics, it would also be possible to automatically assign
activities to certain levels of the taxonomy. Obviously, it is not reasonable to assume that
there will be always clear and unambiguous assignments. There may be some overlaps
between the categories, which means that a single activity may possibly be associated
with more than exactly one level. In this case, mechanisms like fuzzy assignments
according to predefined criteria may be useful. Another option would be to ask the
teacher for clarification about the category the respective activity should be assigned to.
Another open research question is whether, at all times, all levels of the taxonomy have
to be included for defining learning objectives and skills. The coverage of the different
levels of cognitive processing may, for example, depend on what is to be achieved for
different groups of learners in a class.
Acknowledgements
Part of the work presented in this paper is supported by the European Commission
(EC) under the Information Society Technologies (IST) programme of the 6th FP for
RTD-project ELeGI (contract no. 002205). The ideas discussed here evolved from
research done in the scope of the EC-funded project iClass (contract no. 507922) and
have been elaborated to show potential applications within ELeGI. The authors are solely
responsible for the content of this paper. It does not represent the opinion of the EC, and
the EC is not responsible for any use that might be made of data appearing therein.
References
Albert, D. and Held, T. (1994) ‘Establishing knowledge spaces by systematical problem
construction’, in D. Albert (Ed.) Knowledge Structures, New York: Springer Verlag,
pp.78–112.
Albert, D. and Held, T. (1999) ‘Component based knowledge spaces in problem solving and
inductive reasoning’, in D. Albert and J. Lukas (Eds.) Knowledge Spaces: Theories, Empirical
Research, and Applications, Mahwah, NJ: Lawrence Erlbaum Associates, pp.15–40.
Anderson, L.W., Krathwohl, D.R., Airasian, P.W., Cruikshank, K.A., Mayer, R.E., Pintrich, P.R.,
et al. (Eds.) (2001) A Taxonomy for Learning, Teaching, and Assessing: A Revision of
Bloom’s Taxonomy of Educational Objectives, New York: Longman.
Ausubel, D.P. and Robinson, F.G. (1969) School Learning: An Introduction to Educational
Psychology, New York: Holt, Rinehart and Winston.
Biggs, J. (1999) Teaching for Quality Learning at University, Buckingham: SRHE and Open
University Press.
Biggs, J.B. and Collis, K.F. (1982) Evaluating the Quality of Learning – The SOLO Taxonomy,
1st ed., New York: Academic Press.
202 B. Marte, C.M. Steiner, J. Heller and D. Albert
Bloom, B.S. (Ed.) (1956) Taxonomy of Educational Objectives: The Classification of Educational
Goals: Handbook I, Cognitive Domain, New York/Toronto: Longmans, Green.
Doignon, J. (1994) ‘Knowledge spaces and skill assignments’, in G.H. Fischer and D. Laming
(Eds.) Contributions to Mathematical Psychology, Psychometrics and Methodology,
New York: Springer-Verlag, pp.111–121.
Doignon, J-P. and Falmagne, J-C. (1985) ‘Spaces for the assessment of knowledge’, International
Journal of Man-Machine Studies, Vol. 23, pp.175–196.
Doignon, J-P. and Falmagne, J-C. (1999) Knowledge Spaces, Berlin: Springer.
Düntsch, I. and Gediga, G. (1995) ‘Skills and knowledge structures’, British Journal of
Mathematical and Statistical Psychology, Vol. 48, pp.9–27.
Falmagne, J-C., Koppen, M., Villano, M., Doignon, J-P. and Johannesen, L. (1990) ‘Introduction
to knowledge spaces: how to build, test and search them’, Psychological Review, Vol. 97,
pp.201–224.
Gagné, R.M. (1985) The Conditions of Learning, 4th ed., New York: Holt, Rinehart and Winston.
Heller, J., Mayer, B., Hockemeyer, C. and Albert, D. (2005) ‘Competence-based knowledge
structures for personalised learning: distributed resources and virtual experiments’, in
G. Albano, P. Ritrovato and S. Salerno (Eds.) Proceedings of the 1st ELeGI Conference, Vico
Equense, March, pp.14–16.
Heller, J., Steiner, C., Hockemeyer, C. and Albert, D. (2006) ‘Competence-based knowledge
structures for personalised learning’, International Journal on E-learning, Vol. 5, pp.75–88.
Hockemeyer, C. (2003) ‘Structures for teaching and testing competencies’, Talk at the 33rd
European Mathematical Psychology Group (EMPG) Meeting, International University
Madrid, Spain, 26–30 August.
Hockemeyer, C., Conlan, O., Wade, V. and Albert, D. (2003) ‘Applying competence prerequisite
structures for eLearning and skill management’, Universal Journal of Computer Science,
Vol. 9, pp.1428–1436.
IMS learning design best practice and implementation guide (2003) Version 1.0 final
specification, http://www.imsglobal.org/learningdesign/ldv1p0/imsld_bestv1p0.html (retrieved
6 November 2007).
Korossy, K. (1997) ‘Extending the theory of knowledge spaces: a competence-performance
approach’, Zeitschrift für Psychologie, Vol. 205, pp.53–82.
Korossy, K. (1999) ‘Modelling knowledge as competence and performance’, in D. Albert and
J. Lukas (Eds.) Knowledge Spaces: Theories, Empirical Research, and Applications, Mahwah,
NJ: Lawrence Erlbaum Associates, pp.103–132.
Krathwohl, D.R. (2002) ‘A revision of Bloom’s taxonomy. An overview’, Theory Into Practice,
Vol. 41, pp.212–218.
Krathwohl, D.R. and Payne, D.A. (1971) ‘Defining and assessing educational objectives’, in
R.L. Thorndike (Ed.) Educational Measurement, Washington, DC: American Council on
Education, pp.17–45.
Mager, R.F. (1962) Preparing Instructional Objectives, Palo Alto, CA: Fearon Publishers.
Mager, R.F. (1984) Preparing Instructional Objectives, 2nd ed., Belmont, CA: David S. Lake.
Marzano, R.J. (2001) Designing a New Taxonomy of Educational Objective, Thousand Oaks, CA:
Corwin Press.
Mosely, D., Baumfield, V., Higgins, S., Lin, M., Miller, J., Newton, D., et al. (2004) ‘Thinking
skill frameworks for post-16 learners: an evaluation’, A Research Report for the Learning and
Skills Research Centre, Trowbridge: Cromwell Press, http://www.lsneducation.org.uk/pubs/
pages/031541.aspx (retrieved 5 November 2007).
Tyler, R.W. (1949) Basic Principles of Curriculum and Instruction, Chicago: University of
Chicago Press.
Vermunt, J.D. and Verloop, N. (1999) ‘Congruence and friction between learning and teaching’,
Learning and Instruction, Vol. 9, pp.257–280.
