Recent Researches in Educational Technologies

Learning Objects Types Dependability on Styles of Learning

BOYAN BONTCHEV, DESSISLAVA VASSILEVA
Department of Software Engineering,
Sofia University, Faculty of Mathematics and Informatics,
5, J. Baurchier blv., Sofia 1164
BULGARIA
bbontchev@fmi.uni-sofia.bg, ddessy@fmi.uni-sofia.bg

Abstract: - Theory of learning styles is steadily evolving in several decades and reveals importance of such
styles for construction and adoption of adaptive e-learning platforms. Courseware delivery adaptive to learning
styles needs precise definition of types of learning objects most appropriate for individual given learning
character with its specific learning style. The present article provides results of a case study of courseware
delivery of various types of learning objects such tasks, projects, essays, games and intermediate tests with
adaption based on learning styles. It presents evaluation of dependability of learning object types on the styles
of learning defined by Honey and Mumford and shows how practical results may differ from expected ones for
some of the cases.

Key-Words: - Adaptive, Learning Styles, Learning Objects, ADOPTA

1 Introduction [4] and the triangular conceptual model (TCM) [5].

The efforts of modern e-learning systems have been
mainly focused on increasing its efficiency both in
the context of delivery of learning materials and in
the context of developing the educational content.
These efforts have established the several of
contemporary trends in the creation and design of
learning materials, namely as they are following:
the content provided to the user to be
consistent with his/her profile (known as
learner model or user model);
educational content of a course to be created
in a way allowing its reuse in another course
and context, and import/export to another e-
learning system;
training materials to be organized so as to
enable import / export of the whole course
with the pedagogical strategy used in it.
Adaptive hypermedia systems (AHS) are most
commonly used to implement the above
requirements. According to Brusilovsky [1], AHS is
any system that uses hypertext and/or hypermedia,
which focuses on the user and implement its user
model to adapt various aspects of the system to
him/her. This type of systems is most wide
application in the e-learning area.
There are different reference models of AHS,
which implement the above conditions in different
level. Some of them are: Dexter Hypertext
Reference Model [2], AHAM Reference model [3],
Goldsmiths Adaptive Hypermedia Model (GAHM)
The last model, unlike the first three, supports
adaptation of educational content not only to goals,
preferences and knowledge of the learner but also to
different learning styles such as Honey and
Mumford, Gregoric-Mind, Dunn Dunn, etc. [6].
This is very important for modern e-learning,
because it has been established that using of
learning styles significantly increases the
effectiveness of teaching.
Other advantages of TCM are as follows:
availability of content metadata and
adaptation rules;
packaging of content to the standard IEEE
Learning Object Metadata (LOM) [7] and
OMV [8];
TCM consists of three sub-models the
learner model, the domain model and the
adaptation model, which are completely
independent of each other.
These additional features make the learning
process more efficient and provide opportunities for
more reliable interoperability, which entirely
satisfies modern requirements to create, design and
delivery of learning content.
This article will present results of a study
examining the preferences of students to the types of
educational content (narrative content, assignments,
essays, serious games, etc.), according to their
learning style. The study was conducted with the
platform ADOPTA (ADaptive technOlogy-enhanced

ISBN: 978-1-61804-021-3 227

 Recent Researches in Educational Technologies
Platform for eduTAinment), implemented in
accordance to the TCM. It involved 42 students of
the Faculty of Mathematics and Informatics at Sofia
University, Bulgaria.
The next section of the article makes a brief
introduction to the concept of creating educational
content in the form of learning objects and examine
different models of learning styles. The rest of paper
describes the experiment and analyzes the obtained
results.
2 Background
2.1 Learning Objects
Learning objects present a new paradigm for
creating teaching materials. In the old paradigm the
training is organized into lessons and courses that
implement predetermined objectives of the course or
lesson. In the new paradigm educational content is
divided into smaller independent units that can be
used both separately and combined (static or
dynamic) with other ones ([9], [10]).
Main properties that a learning object needs to
have are modularity, interoperability, opportunity
for reuse and accessibility. For the purpose each
learning object has to be described by metadata that
provide information about its context,
characteristics, intended and features. The use of
metadata improves and simplifies the extraction of
information. Moreover, they could support
interoperability, integration of a learning object and
its identification. The most popular and widely used
standards for metadata in the e-learning are two -
Dublin Core [11] and IEEE LOM [12].
Learning objects can contain text, video, audio,
graphics, and other learning objects. According to
the objective of the learning activity they could be
divided into the following types:
narrative content it represents real
learning content. The content author is
supposed to select the appropriate level
of learning objects complexity, which
helps the instructor to choose suitable
training materials.
test question it assesses learner
knowledge of a set (one or more) of
learning object (of type narrative
content) or a concept.
task contains text description of the
requirements for a project or a
description of the problem that the
learner has to solve.
ISBN: 978-1-61804-021-3
topic for essay specifies one or several
topics for essay, describes the
requirements for it, assigned evaluation
criteria and guidelines for work.
serious game - keeps students actively
engaged in the learning process and
assist them in developing and enhancing
their skills for critical thinking,
collaboration, decision making and
problem-solving.
2.2 Learning Styles
Learning style is a combination of intellectual,
emotional, psychological, physical and sociological
personal characteristics that allow students to
absorb, learn and gives meaning knowledge and
competences. Researchers ([13], [14]) have
established that familiarity of the learning style is
essential to improving learners performance ([15],
[16]).
In recent years there have been developed many
learning styles with several variations of them.
According some researchers ([17]) for identifying
different learning styles, there may be encountered
the following four basic types of approaches:
approaches dealing with personal cognitive
characteristics of the type field dependence /
independence [18];
approaches focused on concentration on
learning preferences [19];
approaches dealing with ways of learning
and combining elements of cognitive and
personal learning preferences - a
multidimensional model [20];
approaches dealing with ways of processing
information - based on the cyclical model of
Kolb [21] and the styles converger,
diverger, accommodator, assimilator and, as
well, on the Honey and Mumford [22]
model.
For the the present study, there are used
established and widespread pedagogical strategies
for adapative training, together the learning styles of
the Honey and Mumford [6]. The Honey and
Mumfords model is based on the theory of Kolb
[17] according to which the training has two
dimensions - perception (y-axis in fig. 2) and
processing of information (x axis in fig. 2). Each of
these dimensions is bipolar and, thus, four styles can
be represented by two-dimensional coordinate
system. According to Honey and Mumford the
preferred style is dominant but always adds to itself
elements of other.
228
 Recent Researches in Educational Technologies

The Honey and Mumford model includes the
following four predefined learning styles:
activist he/she is enthusiastic for new
ideas, experiments and seek challenges.
Activists prefer to be direct involvement in
carrying out a task rather than listening to
lectures and detailed descriptions;
reflector he/she does not proceed to
action, preferring to observe a situation
from different perspectives and to gather as
much information about it. Reflectors like to
analyze, to work with examples and detailed
plans to accomplish a task, project or
problem;
theorist he/she is the opposite of activist.
Theorists are relevant to the research,
formalization, concepts and logical theories.
Works well with symbols and abstract
concepts and fully oriented research;
pragmatist he/she is opposite the
reflector. Challenge for students with such
prevailing style is to apply theoretical ideas
into practice. Most important for them is to
acquire practical competences. Pragmatists
are attracted to work on real projects and to
deepen the study and analysis of abstract
ideas, concepts and theories.
matches converger and assimilator and benefits
from abstract conceptualization. On the other side,
the pragmatist corresponds to accommodator and
converger and adores active experimentation, while
the reflector matches diverger and assimilators and
relies on reflective observation.
2.3 E-Learning with Adaptation to Learning
Styles
Distribution of types of learning objects (LOs)
according their suitability for Honey and Mumford
learning style is presented in [24] and it is shown in
fig. 2. This distribution is a result of longstanding
practice of paper authors in teaching students.
According [24] learning objects of type game,
essay, project, problem-solving, comparative
analysis and observation task can also be used to
assess learners knowledge as well as classic tests.
Depending on the extent to which a LO is suitable
for a learning style, there may be selected a set of
candidates LO for evaluation. LOs for assessment
are presented in fig. 2 with elipses having black
labels. Mainly assessment approaches are as follows
[24]:

Activist

P R
r e
a f
g l
m
a c
t t
i o
s r
t
,???

Theorist

Fig.1: Relationship between learning styles

of Kolb and Honey & Mumford

Fig. 1 represents graphically the relation between

learning styles of Kolb and Honey and Mumford
found in [23]. The activist corresponds to Kolbs
styles of accommodator and diverger and feeds Fig. 2: Distribution of learning objects among
from concrete experience, while the theorist Honey and Mumford learning styles [24]

ISBN: 978-1-61804-021-3 229

 Recent Researches in Educational Technologies
self assessment LOs which are suitable for
this method are traditional tests and single
user games;
peer assessment LOs of type essay, task
observation, multi-user game and
comparative analysis tasks are the most
appropriate for this approach;
teacher assessment for this approach, the
most suitable types of LOs are comparative
analysis, task observation, essay, projects
and tests.
3 Courseware Delivery with
Adaptation to Learning Styles
In order to access an adaptive courseware delivery,
instructors have to develop an e-learning course
using authored contents and following pedagogical
strategies for adaptation to some features of the
learner model such as goals, emotions and
preferences, skills and performance shown by
assessment tests and, as well, learning styles [24].
This chapter of the paper presents a case study of
design and delivery of adaptive courseware to
bachelor students by using the ADOPTA platform
and, next, provides an analysis of obtained results.
3.1 Adaptive Courseware Design in
ADOPTA
The ADOPTA platform technology-enhanced
learning allows to learning style and assessed
knowledge of individual learners and comprises an
content authoring tool, an instructor tool and an
adaptation control engine [5]. Authors use the
authoring tool for creating learning objects (LOs)
organized within ontology and for annotating both
the LOs and ontology using respectively LOM and
OMV. Next, instructors create narrative storyboards
of adaptive courses with various paths within the
storyboard graph for different learning styles, by
specifying for what a style a given path is
appropriate. They allocate LOs of different types on
a path according their suitability for given learning
style as discussed above. Learning styles are
assessed in the very beginning of the course by
filling a questionnaire. As well, instructors dispose
LOs of different difficulty levels on pages of the
path, in order to allow the adaptation control engine
to select most appropriate LOs for learner
performance shown until the moment (adaptive
content selection). Each one of the paths finishes at
a control page where the engine generates an
assessment test using questions about LOs traversed
ISBN: 978-1-61804-021-3
by the individual learner. If the learner passes the
threshold defined for the control page, the engine
selects the most appropriate for him/her path to the
next control page and he/she starts traversing that
path (adaptive navigation). Otherwise, the learner
fails with the current assessment test and is returned
back to the previous control point.
For allowing adaptive courseware delivery,
instructors have not only to design the course
storyboard using appropriate types of LOs for
different learning styles but have also to calibrate
the delivery process by specifying weights for all
the paths for the learning styles (e.g., for the four
styles of Honey and Mumford), thresholds at control
pages, complexity levels of LOs for each assessment
note, annotations for links outgoing from the path,
etc. All these settings are used by the adaptation
control engine for management and monitoring the
adaptive learning process.
For the evaluation of the adaptive courseware
delivery of ADOPTA, there was used a bachelor
course at Sofia University about XML technologies
and languages. The instructional design of the
course has two main streams - one intended for
theorists, and another - for the opposite learning
style, i.e. activists [25]. These two main paths in
several sections are divided symmetrically in two
others sub-paths and next merged again, in order to
add LOs to them (respectively for activists and
theorists) intended for both pragmatists and
reflectors. Thus, there are provided sub-paths for
any one of the four combinations (quadrants) of
learning styles shown in fig. 2.
3.2 Practical Experiments and Results
The experimental field trial described in [25] aims at
evaluation of courseware delivery with learning
style adaptation. In practical experiments
participated 84 four-year students of the bachelor
program in Software engineering at Sofia
University, Bulgaria. The students were divided into
two groups with equivalent average student
performance shown by previous assessments. The
first group was taught by several modules of a
traditional, non-adaptive course in XML
technologies given by Moodle. On other hand, the
second one passed the same modules in adaptive
mode using the ADOPTA platform and having
appropriate instructional design for individual
learning styles, i.e. containing paths with LOs most
suitable for the four combinations of individual
learning styles shown in fig. 2.
For the adaptive course, there were provided two
assessment tests - one at intermediate control page
230
 Recent Researches in Educational Technologies

and another at the final one. Students taken the

adaptive version of the same course demonstrated
better performance - with 76,75% average result,
while the first group showed 67,14%. By a
questionnaire after the adaptive course there was
proven that students find adaptive learning much
more appealing and efficient. However, in order to
evaluate its efficiency, evaluation of appropriateness
of LOs types for different learning styles appears to
be crucially important.
First at all, such a research should find dominant
learning style in the context of the studied student
group, namely, bachelors in Software Engineering.
Fig. 3 represents average values of learning styles
Fig. 4: Preferable types of learning objects
for the 42 bachelors participated into adaptive
learning. As expected, such students are
Results shown in fig. 4 called into question
predominantly theorists and partially reflectors than
activists and pragmatists, due to their mathematical previously estimated level of suitability of LOs
background. types for different learning styles shown in fig. 2.
This imposed a more detailed analysis on studying
dependency of preferred LOs types on learning
styles. Fig. 5 reveals such dependency by showing
average learning styles of students appreciated
tasks, projects, essays, games and tests as valuable
types of learning objects. These results demonstrate
similar distribution of styles predominantly
theorist (about 80) and reflector (ca. 75) and less
activist and pragmatist (ca. 60), excepting for
essays. In fact, though only 6 students like essays as
LO type, all of them show very high theorist level
(for some it is the maximum value - 100) and rather
high reflector style, which does confirm our initial
Fig. 3: Learning style of an average bachelor student
in Software Engineering at Sofia University
assumptions as shown in fig. 2.

In conducted experiment, 38 of 42 students

passed successfully the adaptive course modules in
XML technology and have filled in a questionnaire.
They have been asked what main types of learning
objects do they prefer, where each student was
allowed to have multiple choice, i.e. to select
several LOs types. Next figure below shows
students prefer much more exercise tasks and
intermediate tests than writing essays and analyses,
though their theorist-reflector learning style (fig. 3).
As well, practical projects and educational games
are welcome for more that the half of the students,
which reveals their need of project- and game-based
learning.
Fig. 5: Dependency of preferred LOs types
on learning styles

ISBN: 978-1-61804-021-3 231

 Recent Researches in Educational Technologies
Fig. 6: Deviation of learning styles
for preferred LOs types from average values
The results presented in fig. 5 are similar as far
as they show average learning styles for students
selected given type of LOs, which cannot be far
from the average style presented in fig. 3. In order to
reveal the real differences in styles of students voted
for each one of the five object types, we have to
investigate the deviation of learning styles for
preferred LOs types from the average values. Fig. 6
presents extremely strong deviation from average
theorist and reflector style for students preferring
essays, which shows such type of LOs is strongly
desired by learners with such predominant style.
Theorists and reflectors like less projects and tasks
but not at all intermediate tests. On other hand,
activists and pragmatists like having tests but not
tasks. It is interesting to point out that only
reflective activists like playing educational games,
contrary to the assumption that games are suitable
for pragmatists.
25
20
15 Strong activists (>=80)
Strong reflectors (>=80)
10 Strong theorist (>=80)
Strong pragmatists (>=80)
5
0 different learning style. They are very important for
Tasks Projects Essays Games Tests
Fig. 7: Preference of LOs types by learners
with very predominant learning styles
Finally, it was interesting to find out what are the
preferences of learners with very prominent learning
styles, because strongly expressed learning style is
ISBN: 978-1-61804-021-3
important in finding clear trends. Here, under
strongly expressed learning style there is accepted a
style with value equal or greater than 80 according
the questionnaire of Honey and Mumford fiving 100
as a maximal value for a style. Fig. 7 shows the
preference of LOs types by learners with very
prominent learning styles. In fact, the numbers
resemble these shown in fig. 4 but with distribution
over styles.
Fig. 7 does not reveal the real dependency of
LOs types on predominant styles because the
vertical bars present only the number of learners
with given style. Fig. 8 gives the relative preference
of LOs types by learners with very prominent
learning styles. Results seem different because here
the bars show the percentage of learners with
prominent style voted for tasks, projects, essays,
games and tests. For example, 90% of strong
activists like executing tasks, while 90% of strong
pragmatists like solving tests. Games are preferred
mainly by strong activists and reflectors. Only few
of learners with prominent styles like essays but, if
any, they are strong reflectors and theorists.
100%
90%
80%
70%
60% Strong activists (>=80)
50% Strong reflectors (>=80)
40% Strong theorist (>=80)
30% Strong pragmatists (>=80)
20%
10%
0%
Tasks Projects Essays Games Tests
Fig.8: Relative preference of LOs types by learners
with very prominent learning styles
4 Discussion
Results presented in previous section reveal
important issues about preferences of learners with
adaptive courseware design as far as instructors
should know the level of suitability of various types
of learning objects for a given family of learning
styles.
Assumed distribution of LOs types on learning
style was proposed based on our teaching
experience. However, real preference of types of
LOs by learners is slightly different which imposes
232
 Recent Researches in Educational Technologies
some revisions over the initial distribution. Fig. 9
presents a revised view of the distribution taking
into account the obtained results. It shows games are
most suitable not for pragmatists but for reflectors,
while essays are preferred not by activists but by
theorists. As well, there are no available types of
LOs for assessment appropriate for pragmatists.
Fig. 9: Revised distribution of learning objects to
Honey&Mumford learning styles
5 Conclusion
The paper addressed practical evaluation of the role
of learning styles on adaptive courseware delivery
developed and executed in ADOPTA platform.
More specially, it presented results about
dependency of learning object types on learning
styles, which plays a crucial role while constructing
storyboards of adaptive e-learning courses. The
obtained results are based on questionnaire about
relevance of types of LOs filled by bachelor
students in Software Engineering after passing an
adaptive course developed and delivered by
ADOPTA. They reveal a strong dependency of LOs
types on styles. However, several remarks may be
concluded here:
First at all, the results are context-dependent
they depend on the average student
ISBN: 978-1-61804-021-3
profile. In cases where reflectors and
theorists are not predominant, results may
be rather different. As well, it has to pointed
out that learning style of an individual is not
fixed forever but upgrades with time.
Some of the obtained results differ from the
expected it remember us always to count
with real attitude of learners and their
feedback
Finally, it is under question whether
students have to get preferable types of
learning objects or, otherwise, types
imposed by instructors following given
pedagogical strategy. Probably, efficient
solution should take into account both the
pedagogical strategy and students
preferences.
Acknowledgments
This work is partially supported by both the SISTER
project funded by the European Commission in
FP7-SP4 Capacities via agreement no. 205030 and
the ADOPTA funded by the Bulgarian NSF under
agreement no. D002/155.
References:
[1] Brusilovsky, P., Methods and techniques of
adaptive hypermedia. In User Modeling and
User-Adapted Interaction 6(2-3), 1996, pp. 87-
129.
[2] Halasz, F., Schwartz, M., The Dexter hypertext
reference model. Communications of the ACM,
Vol. 37, Issue 2, ISSN:0001-0782, 1994, pp.
30-39.
[3] De Bra P., Houben G.-J., Wu H., AHAM: A
Dexter-based Reference Model for adaptive
Hypermedia. ACM Conf. on Hypertext and
Hypermedia, ISBN:1-58113-064-3, Darmstadt,
Germany, 1999, pp. 147-156.
[4] Ohene-Djan, J., Gorle, M., Bailey, C. P., Wills,
G. B., Davis, H. C., Understanding Adaptive
Hypermedia: An Architecture for
Personalisation and Adaptivity. Proc. of
Workshop on Adaptive Hypermedia AH2003,
August 26-30, 2003, Nottingham, UK.
[5] Bontchev B., Vassileva D., Software
architecture of a self adaptive hypermedia e-
learning system. Proc. of 3rd Int. Conf. on
Computer Science2006, Istanbul, Turkey, 12 -
15 October 2006, pp. 306-311.
[6] Karagiannidis, C., Sampson, D.,
Accommodating Learning Styles in Adaptation
233
 Recent Researches in Educational Technologies
Logics for Personalised Learning Systems,
Proc. of World Conf. on Educational
Multimedia, Hypermedia and
Telecommunication, 2002, pp. 1715-1726.
[7] IMS Global Learning Consortium, IMS Meta-
data Best Practice Guide for IEEE 1484.12.1-
2002 Standard for Learning Object Metadata
v1.3., 2004, available online at:
http://www.imsglobal.org/metadata/mdv1p3pd/
imsmd_bestv1p3pd.html
[8] Hartmann, J. et al, Ontology Metadata
Vocabulary and Applications. Proc. of Int.
Conf. on Ontologies, Databases and
Applications of Semantics, Workshop on Web
Semantics (SWWS), Springer, 2005, pp.906-
915.
[9] Collis, B., Strijker, A., New pedagogies and
reusable learning objects: Toward a new
economy in education. J. of Educational
Technology Systems, Vol. 30 (2), 2002, pp.
137-157.
[10] Friesen, N., Roberts, A., Fisher, S., CanCore:
Learning Object Metadata. Canadian Journal
of Learning and Technology, Vol. 28, no. 3,
2002, available online at:
http://www.cjlt.ca/index.php/cjlt/article/view/1
09/102.
[11] Dublin Core Metadata Initiative, Dublin Core
Metadata Initiative Overview, July 2009,
available online at:
http://dublincore.org/documents/dces/.
[12] IEEE LTSC, IEEE LTSC Working Group 12:
Learning Object Metadata, 2004, available
online at: http://ltsc.ieee.org/wg12/index.html
[13] Bostrom, R. P., Olfman, L., and Sein, M. K.,
The Importance of Learning Style in End-User
Training, MIS Quarterly, (14:1), March 1990,
pp. 101-119.
[14] Lindsay, E. K., An analysis of matches of
teaching styles learning styles and the uses of
educational technology, Doctoral dissertation,
North Carolina State University, Raleigh, 1999.
[15] Griggs, Sh., Learning Styles Counseling. ERIC
Digest, 1991.
[16] Triantafillou, E., Pomportsis, A., and
Demetriadis, S., The design and the formative
evaluation of an adaptive educational system
based on cognitive styles. Computers &
Education, Vol. 41, Issue 1, August 2003.
[17] Sadler-Smith, E., Learning Style: Frameworks
And Instruments. Educational Psychology,
Vol. 17, Issue 1, 1997.
[18] Riding, R. J. & Rayner, S. G., Cognitive styles
and learning strategies. London: David Fulton
Publishers, 1998.
ISBN: 978-1-61804-021-3
[19] Riechmann, S. W., & Grasha, A. F., A rational
approach to developing and assessing the
construct validity of a student learning styles
instrument. Journal of Psychology, 87, 1974,
pp.213-223.
[20] Murray-Harvey, R., Conceptual and
measurement properties of the productivity
environmental preference survey as a measure
of learning style. Educational and
Psychological Measurement, 54(4), 1994, pp.
1002-1012.
[21] Kolb D. Experiential learning: experience as
the source of learning and development.
Englewood Cliffs, New Jersey: Prentice Hall,
1984.
[22] Honey, P. and Mumford, A., The manual of
learning styles. Maidenhead: Peter Honey,
1992.
[23] Munoz-Seca, B., and Silva Santiago, C., Four
Dimensions to Induce Learning: The Challenge
Profile, IESE Working Paper No. D/520,
September 2003, available at SSRN:
http://ssrn.com/abstract=462282
[24] Bontchev, B., Vassileva, D., ADAPTIVE
ASSESSMENT BASED ON LEARNING
STYLES AND STUDENT KNOWLEDGE
LEVEL, Proc. of 3rd Int. Conf. on Computer
Supported Education (CSEDU2011), ISBN
978-989-8425-49-2, The Netherlands, May
2011.
[25] Vassileva, D., Evaluation of Learning Styles
Adaption in the ADOPTA E-learning Platform,
Proc. of CompSysTech2011, Vienna, Austria
(in print).
234