Documentos de Académico
Documentos de Profesional
Documentos de Cultura
Abstract User models are essential to e-learning systems, giving students learning continuity, tutors evidence of
students progress, and both a way to personalize students learning materials to their abilities and preferences.
Personalizing information has long been the motivation behind developing e-learning systems. Adaptive educational
systems attempt to maintain a learning style profile for each student and use this profile to adapt the presentation and
navigation of instructional content to each student. This kind of system adapts the learning process on the basis of
the students learning preferences, knowledge, and availability. One such Web-based tool is built at the Business
School of Professional Studies in Blace (the system of intelligent evaluation using tests), which infers student
knowledge using adaptive testing. The knowledge will not be evaluated according to fixed standards, but it will
depend on individual characteristics of each student, therefore the learning style which is the most suitable. This
system is part of the integrated assessment system inside of integrated curriculum.
Keywords Student profile, adaptive system, adaptive testing, learning style, integrated assessment.
1. Introduction
Todays dominant e-learning systems are LMS (Learning Management Systems) systems
like Blackboard [1], Moodle [2]. They are integrated systems which offer support for a large
range of activities in the e-learning process. Teachers can use Learning Content Management
Systems (LCMS) for creating courses and tests and LMS for communication with students,
monitoring and assessment of their knowledge. Students can learn and communicate with each
other through LMS. However, the problem lies in the fact that LMS usually does not offer
services for personalization, which means that students have access to the same set of
educational tools and resources, and personalization aspects like the differences in knowledge,
interests, motivation and goals are ignored [3].
According to ALFANET [4], adaptability in the context of e-learning represents the
creation of students experiences under different circumstances (personal characteristics,
pedagogical knowledge, his/her interaction, the result of the previous learning process) in a
certain period with a tendency of increasing the pre-defined criteria of success (efficiency of
learning: result, time, the price, satisfaction of the user, etc.). Since the system is adapted to the
person, i.e. the user, this type of adapting is called personalization.
There are three factors which must be taken in consideration when we speak about the
adaptability of e-learning systems:
Student (whose characteristics are: degree of knowledge, technical education, educational
goals, interests, motivation, learning style, personal characteristics, general knowledge,
etc.)
Hardware and software platform (PC/laptop/PDA mobile phone, the size of the screen,
available input devices, connection speed, the performance of the processor, memory
size, operating system, Web browser, etc.),
Environment (physical environment in which we observe the interaction-lights, noise,
geographical location and other external factors).
1
Testing is directly related to education and training as a way to measure the performance
levels of students. Different methods of assessment have been used in different contexts, but the
most common tools are oral test and the paper-and-pencil test. Given that the computer has been
an educational tool in the last few decades, there are a number of benefits in using computers for
assessing performance [12]: large numbers can be marked quickly and accurately, students
response can be monitored, assessment can be offered in an open-access environment,
assessments can be stored and reused, immediate feedback can be given, assessment items can be
randomly selected to provide a different paper to each student.
Software tools and web-based sources are frequently used to support the learning process,
so it seems reasonable to use similar computer-based technologies in the assessment process.
The basic motive of our research is to build a flexible system which will depend on the
performances of students (the differences in levels of previous knowledge, competence, learning
styles, communication abilities, cognitive style, etc.).
In this paper, the starting point is the fact that the traditional model of knowledge
assessment does not have enough influence on the motivation for success. Namely, the teacher
does not content himself with ranking students according to their knowledge but he wants to
show with the grade how they are different from each other. The suggested model of knowledge
assessment should lie on the success as the final experience, because teachers should establish
the requirements which only determine positive goals.
This model will be based on knowledge evaluation, where knowledge will not be evaluated
according to fixed standards, but it will depend on individual characteristics of each student,
therefore the learning style which is the most suitable.
The suggested model will be presented in the continuation.
2. Literature review
User models are essential to e-learning systems, giving students learning continuity,
tutors evidence of students progress, and both a way to personalize students learning materials
to their abilities and preferences. Personalizing information has long been the motivation behind
developing e-learning systems.
Few attempts have been made to model user cognitive and affective attributes in order to
achieve systems adaptively according to the needs of individual user. And while researchers
agree on the importance of adaptation towards user cognitive and affective characteristics, there
is little agreement on which features can and should be used and how to use them [13].
Guidelines and examples on content adaptation and presentation depending on various
learning style in combination with instructional design theories are presented in [14]. Lessons are
designed based on combinations of educational material modules, supporting several levels of
adaptation towards individual learning style. Paper [15], gives guidelines for preparing learning
materials according to different learners characteristics, based on pedagogical strategy and
motivation factor with a strong psychological background, applying categories of Kolbs learning
styles.
The most generic ITSs (Intelligent tutoring systems) architectures suggest building a good
student model that reflects systems beliefs about learners mastery level in certain concepts.
Moreover, such architecture is based on different domains, pedagogical strategies and enables
systems to perform individualized tutoring for learners [16].
3
With the Internets evolution, researchers have attempted to deploy ITSs on the Web. These
Web-based systems retain most generic ITS architecture features, such as AH (Adaptive
Hypermedia), which generates content with different levels of detail according to users
knowledge. This is known as adaptive presentation. AH system can offer adaptive navigation by
giving users directional assistance in selecting the most relevant link. Such adaptive methods
main purpose is to support students in hyperspace orientation and navigation [16]. Two wellknown examples are: ISIS-Tutor [17] and Hypadapter [18].
In addition, Web-based adaptive and intelligent educational systems (AIESs) have begun
adopting and benefiting from AH technologies. Examples include Web-administered multiplechoice tutors [19]. The three most popular techniques used in Web-based AIES are direct
guidance, adaptive link annotation, and adaptive link hiding. Representative examples reflecting
these include ELM Adaptive Remote Tutor (ELM-ART), Adaptive Statistics Tutor (AST), and
InterBook.
SIETTE [20] is an example of a Web-based adaptive testing system. The only kind of
learning material it possesses is questions. The system generates an adaptive sequence of
questions to assess student's knowledge. SIETTE is not complete AIES but it has to be used as
component in distributed Web-based AIES.
2.1.
The terms learning styles and cognitive styles have been often used interchangeably in
literature. Jonassen and Grabowski [21] distinguish between learning and cognitive styles by
explaining that learning style instruments are typically self-report instruments, whereas cognitive
style instruments require the learner to do a task which is then measured to some trait or
preference. However, there is a difference between their uses. Cognitive style deals with the
form of cognitive activity (i.e., thinking, perceiving, remembering), not it's content. Learning
style, on the other hand, is seen as a broader construct, which includes cognitive along with
affective and psychological styles.
Learning style is a distinctive and habitual manner of acquiring knowledge, skills or
attitudes through study or experience while learning preference is favouring of one particular
mode of teaching over another [22].
There are a lot of types of learning styles. Some of them are:
Kolbs Learning Styles - The Kolb Learning Style Inventory (LSI) is a commercially
available questionnaire (www.learningfromexperience.com) with twelve items where
respondents rank-order four sentence endings that correspond to the four separate learning styles:
diverger, converger, assimilator, and accomodator. These four styles of learning are assessed by
two dimensions (abstract/concrete and active/reflective), which describe four abilities required to
be an effective learner. Scores on the Kolbs Learning Style Inventory identify the learners
preferred style of receiving and organizing [23].
Gregorc Learning/Teaching Style Model based in phenomenological research as well as
Kolbs experiential learning cycle, that defines learning style as distinctive and observable
behaviours that provide clues about the mediation abilities of individuals and how their minds
relate to the world and, therefore, how they learn [24]. The Gregorc Style Delineator (GSD) is
commercially available (www.gregorc.com) and asks the respondent to rank order ten sets of
four words that correspond to the four poles of the two mind qualities.
4
Dunn and Dunn Learning Styles - there are 5 main categories (environmental, emotional,
sociological, physiological, and psychological) and 21 elements in considering a learning style
when using their model [25]. iWeaver [26] uses Dunn & Dunn model - the application of a
Bayesian network is planned to predict and recommend media representations to the learner.
The Honey and Mumford Approach - A more promising alternative than the Kolb LSI may
be a measure developed by Honey and Mumford [27], named the Learning Styles Questionnaire
(LSQ). The Kolb model is the theoretical background to Honey and Mumfords LSQ, which has
four styles: theorist, activist, reflector and pragmatist. INSPIRE [28] uses Honey and Mumford
model.
According to the teaching model of Felder-Silverman [29], learners can be divided into
four categories according to their characteristics: active learners (they like team work) and
reflective learners (students who think and like individual work); sensing learners (facts, details
and practical work is important to them) and intuitive learners (they like abstract materials,
innovations, revealing connections and possibilities); visual learners (they remember what they
see: pictures, diagrams etc.) and verbal (they use spoken or written words), sequential learners
(they solve the problem by using small steps) and global learners (they learn in large leaps,
details are unclear to them but they can create connections between objects).
Tangow [30] uses sensing-intuitive dimension of Felder-Silverman model. Hong and
Kinshuk [31], develop a mechanism to model students learning styles and present the matching
content to individual student, based on Felder-Silverman Learning Style Theory. CIMEL-ITS
[32] is a novel pedagogical framework that facilitates integrating the Felder-Silverman Learning
Style Theory in an intelligent tutoring system (ITS).
VARK Model: The acronym VARK stands for Visual (V), Aural (A), Read/Write (R), and
Kinesthetic (K). Fleming [33] defines learning style as an individuals characteristics and
preferred ways of gathering, organizing, and thinking about information. VARK is in the
category of instructional preference because it deals with perceptual modes. It is focused on the
different ways that we take in and give out information. The only perceptual modes, or senses,
it does not address are taste and smell. The VARK Inventory provides metrics in each of the four
perceptual modes, with individuals having preferences for anywhere from one to all four.
Cognitive styles refer to the preferred way an individual processes information. Unlike
individual differences in abilities (e.g., Gardner, Guilford, and Sternberg) which describe peak
performance, styles describe a person's typical mode of thinking, remembering or problem
solving. The research literature classifies cognitive styles in two major subgroups [21]:
Information Gathering and Information Organizing. Within the subgrouping of Information
Gathering, there are visual/haptic, visualizer/verbalizer, and levelling/sharpening styles [34].
Within the subgroup of Information organizing, there are serialist/holist and analytical/relational
styles.
Barber and Milone [35] developed a simpler instrument to measure cognitive styles (VAK
Instrument), which simply classifies learners as being visual, auditory and kinesthetic or a
combination of these. Cognitive styles are recognized as describing learner traits and are mostly
related to their preferred way to process information. Persons with a visual preference tend to
show a greater ability to analyze and integrate visual information. An auditory learner would
prefer to process information in the form of verbs, either written or spoken [34]. Kinesthetic
learners prefer to process information through tactile means such as interactive media.
Cognitive characteristics of students in fact affect the learning process. However, a very
small number of adaptive systems are created to take into consideration these cognitive
5
characteristics because it is difficult to gather this type of data and then design the adaptive rules
according to them.
Developing e-learning systems that adapt to student learning style is not a trivial task.
Design and development challenges include selecting the appropriate learning style model and
instrument, creating course content consistent with the various learning styles, and determining
the level and degree of adaptation of domain content. The generic pedagogical framework would
allow easy integration of learning styles in an e-learning system.
Hawk and Shah [36] performed comparative analyze of five learning style instruments (the
Kolb Learning Style Indicator, the Gregorc Style Delineator, the FelderSilverman Index of
Learning Styles, the VARK Questionnaire, and the Dunn and Dunn Productivity Environmental
Preference Survey). But only two Web-based instruments that the students could do on their own
time and report the results to the instructor would be the Felder Silverman and the VARK, with
only the VARK having a moderate support for validity and reliability. An advantage for using the
Felder Silverman would be it is an instrument designed for engineering students.
In the end there are inevitable questions:
Can separate dimensions be considered independent and is it possible to reduce the
mutual overlaps onto the smaller number of more basic styles?
Are tests and questionnaires used for the estimation of different students styles valid and
reliable?
The task of the practical teachers is, on the basis of the presented ideas, to ensure and
create as diverse as possible forms of lectures, as well as to monitor carefully which of them suit
the concrete conditions best.
Generating Student Profile begins with students filling out their basic data and tests. The
data are:
Fill out their personal data, data about the school and the major they are attending
(registration),
Fill out VARK questionnaire (available on address www.vark-learn.com) offers sixteen
statements in order to determine their learning style. On the basis of these data the system
recommends the student teaching materials which would suit best his learning style.
VARK questionnaire (version 7.0) is used as it is free and available. The total of all four
scores ranges from 13 to 48, with individuals having a preference for one, two, three, or all four
of the learning channels. Students and faculty can self-administer, self-score, and self-interpret
the VARK Inventory.
The basic course is distributed online in different forms: as text (doc and pdf format) and as
a multi-media course (presentation with audio and video file and simulations).
Instructors modified their teaching styles to accommodate the learning styles of all the
students in their classes. When a lecture plan is created, it is desirable that as many activities as
possible are utilized which reflect different students learning styles. This means that beside the
theory (text and pictures) it is necessary to include the application of the theory (multimedia
simulation). Web laboratory [40] completely accomplishes those requirements.
WnetSim [41] presents an environment which allows students to visualize and simulate a process
in computer network with any topology. The purpose of the simulator is to help students
comprehend basic principles of acting of one TCP/IP computers network as a part of the
educational system of computer network.
Web-based educational computer system simulator SIMAS [42] provides for entering and
compiling an assembly language program, as well as for loading and executing the resulting
machine code. The simulator enables user to visualize the whole execution process of an
instruction and to compare side-by-side assembly and machine language instructions.
Table below summarizes a number of learning activities to support each learning style
[33].
Visual
Aural
Read/Write
Kinesthetic
Diagrams
Graphs
Debates, Arguments
Discussions
Books, Texts
Handouts
Real-Life Examples
Examples
Colors
Charts
Written Texts
Different
Fonts
Spatial
Arrangement
Designs
Conversations
Audio Tapes
Video+Audio
Seminars
Reading
Written Feedback
Note Taking
Essays
Guest Lecturers
Demonstrations
Physical Activity
Constructing
Music
Multiple Choice
Role Play
Drama
Bibliographies
Working Models
in the blank and drag and drop. Each of the options dynamically generates the form for filling in
the answer. Figure 3 illustrates the use of multiple choices.
As soon as questions and correct/incorrect answers are defined and tests cases are supplied,
it is possible to perform the assessment in the delivery stage.
The delivery stage includes a question presentation, an interface for the student to answer
the question and the retrieval of the answer for evaluation. This stage depends on the technology
used for the learning system. For the delivery stage of an assessment in ADES-TS system, there
are: windows and web-based learning environment. An assessment starts as soon as a student has
logged on to the system and has entered a user name and password. The module for the
presentation of an assessment question displays different question types differently but preserves
the unique assessment page presentational characteristics. An assessment can be automatically
evaluated at any moment. Student can check his/her answer at any time (it depends on teacher
decision). After the check, the student cant change his answer and can submit the assessment for
automatic evaluation.
At the assessment stage, the system should evaluate answers as correct/incorrect, deliver
feedback to student, grade the question. System should provide a summary score as well as a
suggestion for learning those modules for which the answers were incorrect and record the
students' performance. When an assessment is submitted, students cannot access it anymore.
After the automatic assessment evaluation, students are provided with summary assessment
results containing all the assessment questions together with the number of points acquired
question by question and for the whole assessment.
Personalized Testing System Model is suggested in paper [44]. The model suggests four
types of identification: preliminary (determines learning/cognitive styles of students), self-testing
(self-assessment of knowledge conducted online, important interactive element for students),
progressive (assessment of knowledge in different phases of the learning process) and final
testing. This system is part of the integrated assessment system inside of integrated curriculum
(Fig.3).
Each assessment module has its own database. Those databases store two different groups
of data. In the first group are data used to create questions questions, answers, parameters,
ranges, audio/video materials, etc. The second group are data generated for specific testing
instance: student's details, answers, problem's parameters, etc. Despite this, communication
between system and module during the assessment is limited to assessment request (system gives
10
student id to logon to a specialized module) and a test result response (quantitative score,
percents). Detailed testing instance results can be requested from specialized module after the
end of testing.
Each of the mentioned types of testing is conducted in school (windows application) except
the online testing. Short quizzes in the form of multiple choice questions are made available
online for students who want to self-test their knowledge or learning. Students, therefore, receive
immediate feedback on whether their answers are correct, and what the correct response should
be for each question.
To create an assessment, the following parameters have to be determined: accessibility students use the test to help them study (online test), grading - knowledge assessment, timing
(limit for answering) - the criteria of stopping the test, assessment scale, question selection
criterion - for dynamic selecting of questions from database (fixed, random, or adaptive) and
test-finalization. In the progressive and the final test the number of easy and more difficult
questions is varied (adaptive testing).
and subjectively by asking participants about their perceived enjoyment, progress, and
motivation. It was expected that giving learners a choice of media experiences would have a
positive effect.
The distribution of the four perceptual styles visual (V), auditory (A), read/write (R) and
tactile-kinesthetic (K), is displayed in Fig 6. The light bars represent percentages for all 63
students, whereas the dark bars represent percentages for the group of 25 students used for the
statistical analysis. It was interesting to note that 12 of all 63 (19%) participants expressed
equivalent scores for multiple (two or three) most-preferred styles compared to 6 of 25 (24%) of
the analyzed students.
The R style was rare: only 2% of all students expressed R style and none of the analyzed
students. Similarly, only 6% of all students expressed an A style, compared to 8% of the analyzed
students. The second most common perceptual style was V with 40% of all students and 36% of
the analyzed students. Learners were most commonly assessed with K as their most-preferred
perceptual style: 52% of all students, compared to 54% of the analyzed students. In summary,
94% of the students expressed either K or V as their most-preferred style.
Examining the influence of learning styles on the learning process itself, i. e. the adoption
of new knowledge, we reached the conclusion that the testing process itself would crucially
influence the learning style preferred by the student.
Students, who during the learning process take notes and draw pictures or diagrams in
order to remember the information more easily, use the visual learning style. They prefer to learn
alone. They prefer to get questions formulated through picture illustrations, diagrams and the
like.
The auditory learning style is dominant for those who easily learn when information is
presented through a text (written or spoken). They learn in the easiest way by listening to
lectures, discussions, and exchange of ideas. They find the textual questions most suitable.
Learning in a group or pair is a good way to learn for these students. That is why they cooperate
perfectly during the group work (e.g. group projects).
Tactile/kinesthetic students learn the best through movement, game, acting and action,
actively researching the world around them. They prefer practical work.
The following table shows the students style and recommended representation of tests.
12
According to this, students were divided into three groups (V-visual students, A-auditory,
K-practical students).
The testing showed that the system described in this paper proved to be useful to all the
three groups of students. We describe our research in [41].
13
References
[1] Blackboard - www.blackboard.com
[2] Moodle http://moodle.org
[3] P.Brusilovsky, Knowledgetree: A Distributed Architecture for Adaptive E-Learning,
Proceedings Www 2004, New York, USA, 2004.
[4] Popescu, E.; Trigano, P.; Badica, C, Evaluation of a Learning Management System for
Adaptivity Purposes, Computing in the Global Information Technology, ICCGI 2007.
[5] E. A. Edmonds, "Adaptive Man-Computer Interfaces", Computing skills in the user interface,
Coobs M.J &Alty J.L Eds., Computer and People series, New York, Academic Press, 1981.
[6] Jia-Jiunn Lo, Ya-Chen Chan, Relationships between User Cognitive Styles and Browsing
Behaviors of an
Online Learning Web Site, International Conference on Cyberworlds 2008.
[7] B. Dara-Abrams, Applying Multi-Intelligent Adaptive Hypermedia to Online Learning, Ph.D.
Dissertation, Union Institute & University, 2002.
[8] S. Benadi, Structuration Des Donnes Et Des Services Pour Le Tlenseignement, Phd.
Thesis, Insa Lyon, 2004.
[9] Felder, R. M. Matters of Styles. ASEE Prism, 6(4), 18-23, 1996.
[10] Rasmussen, K. L. Hypermedia and learning styles: Can performance be influenced?
Journal of Multimedia and Hypermedia, 7(4), 1998.
[11] Felder, Richard, "Reaching the Second Tier: Learning and Teaching Styles in College
Science Education." J. College Science Teaching, 23(5), 286-290. 1993.
[12] Harvey, J. & Mogey, N. (1999). Pragmatic issues when integrating technology into the
assessment of students. In S. Brown, P. Race & J. Bull (Eds), Computer-assisted assessment in
higher education (pp. 720). London: Kogan-Page.
[13] Brusilovsky P., Peylo C., Adaptive Hypermedia User Modeling and User-Adapted
Interaction 11, Kluwer Academic Publishers, 2001, pp. 87-110.
[14] G. Magoulas, K. Papanikolaou , M. Grigoriadou, Adaptive web-based learning:
accommodating individual differences through systems adaptation, British Journal of
Educational
Technology.Vol
34(4)
2003,
[Online]
Available:
http://www.dcs.bbk.ac.uk/~gmagoulas/bjet.pdf
[15] Milosevic D., Brkovic M., Bjekic D., Designing Lesson Content in Adaptive Learning
Environments, iJET International Journal of Emerging Technologies in Learning, 2006.
Available at: www.i-jet.org
[16] Qing Li, Rynson W.H. Lau, Elvis W.C. Leung, Frederick Li, and Victor Lee, Benjamin W.
Wah, Helen Ashman, Emerging Internet Technologies for E-Learning, Published by the IEEE
Computer Society, 2009.
[17] cs.joensuu.fi/~mtuki/www_clce.270296/Brusilov.html
[18] www.springerlink.com/content/jkh0k872vu318135
14
16