Utilizing Learning Styles for Interactive Tutorials

Utilizing Learning Styles for Interactive Tutorials

**************************

**************************

Abstract: Web-based learning environments, such as eLearning and distance learning systems, have become more common and popular in commercial and educational settings. They constitute a growing business today, and the concept of learning ‘anywhere, anytime’ has received much attention. These systems are still in their infancy, however, and are faced with various challenges to become fully effective tools. Their lack of individual preferred ways of instruction is a challenge we address in this paper.

We introduce a Web-based learning environment that teaches concepts from Artificial Intelligence to college students. This environment is intended to be used as a complementary tool for the standard lectures. By adapting the instruction and learning material to the individual student’s learning style, the tutorial gives the student a personal learning experience. The system was developed in collaboration with HCI students to combine the use of learning styles with the principles of user-centric design.

Keywords: learning styles, web-based learning system, personal instruction, HCI

1 Introduction

When the Internet boomed in the mid-nineties, Web-based instruction started to become high in demand from both corporations using it for employee training and educational institutions interested in meeting the students’ needs. This learning environment provides more flexibility with time, pace and place, and is often characterized as ‘anywhere, anytime.’

Other benefits of Web-based learning are support for classroom instruction and platform independence. An application installed and supported in one place can be used by learners equipped with any kind of computer connected to the Internet. This kind of learning environment is often referred to as eLearning, Web-based training, distance-learning, or online education system.

Developing successful Web-based learning environments, however, has shown to be a challenging and difficult endeavor that requires knowledge from multiple domains like pedagogy, psychology, knowledge engineering, software engineering, and Web technologies.

There are several Web-based learning systems available today that we will look at later, but research shows that there are many challenges to overcome before these systems are efficient enough. These systems usually assume the users to be a homogenous learning group and are therefore presenting interfaces, functionalities and the same course content uniformly to every user. They expect that all the different users learn the same way, and do not accommodate for the rich diversity of learning styles nor the user’s preferred

ways of learning. This can lead to dropping interests by the students, and failure in achieving the expected academic results.

2 Motivation

In this experiment we wanted to explore potential ways for improving such Web-based learning systems by utilizing learning styles, and develop a prototype, the Interactive Teaching Tutorial (ITT). We wanted to investigate the hypothesis that implementing a personalized learning environment would increase the student’s performance and motivation for learning.

The ITT was developed as a tool for the students enrolled in an Artificial Intelligence class at California Polytechnic State University, San Luis Obispo, as a complement to the standard classroom instruction. We wanted to discover how learning style theories could be utilized developing a Web-based learning system, and discover the tradeoffs of applying these theories.

3 Existing Systems

In the 1960’s the US military started developing training systems referred to as drill and practice and simulation activities, e.g. flight simulators. This form for training could be used instead of teacher-directed instruction, as a supplement to the traditional training, and as partial replacement of costly training in the real environment. In situations where testing of skill and knowledge is required, Computer Assisted Instruction (CAI) is used for management skills, industrial factory floor training, information technology products, health care, government services, and many other domains. CAI has been in use for several decades in various forms and applications.

Another area of development is Intelligent Tutoring Systems (ITS). ITS allow the emulation of a human tutor in the sense that an ITS can know what to teach (domain content), how to teach it (instructional strategies), and to acquire information about the student being taught (Capuano, 2000). These systems originated from the Artificial Intelligence (AI) movement of the late 1950's and early 1960's. It seemed reasonable to assume that, once we created machines that could think, they could perform any task we associate with human thought, such as instruction. This has proven to be harder than expected, and even today’s ongoing large ITS projects are not close to their forefathers’ high visions. They are especially struggling with high complexity and development costs.

Web-based learning (WBL) systems, however, are less ambitious and less complex than the Intelligent Tutoring Systems. WBL is one of the tools used by academic institutions and corporations (Berkeley, 2002). In traditional academic institutions, WBL systems are generally housed administratively in a "distance education" department alongside other at-distance delivery methods such as correspondence, satellite broadcast, two-way videoconferencing, videotape and CD-ROM/DVD delivery systems. All such systems seek to serve learners at some distance from their learning facilitator.

WebCT (WebCT, 2002) is a commercial course management system for higher education. According to (Diaz, 1999), WebCT is used at over 1700 institutions with University of Georgia as the largest user with 1,150 courses used by 32,117 students. WebCT provides utilities, components and tools to develop and maintain a virtual university. Some examples are: course builder, course appearance, student manager, file manager, course homepage, assignments, quiz and surveys.

Blackboard (Blackboard, 2002) began as collaboration among a team of students and faculty at Cornell University in 1996. Blackboard is a similar web-based product also used worldwide. Among 1600 institutions, Boston University is among Blackboard’s largest users with 896 courses used by 18,881 students.

LearningSpace (LeaLotus, 2002) is another commercial system developed by Lotus Development and currently marketed by IBM. This is a similar system to WebCT and Blackboard. However, this system has been built on top of the groupware product Lotus Notes and the Domino Web Server software, and takes advantage of the features that Lotus Notes offers for conferencing, e-mail and meeting-scheduling.

Figure 1: Course entrance for Lotus LearningSpace

(LeaLotus, 2002)

Taken to some extreme as so-called virtual universities, academic institutions have started offering more and more courses through their distance learning programs, or as conventional courses with substantial WBL capabilities.

4 Learning Styles

The idea that people learn differently is venerable and probably had its origin with the ancient Greeks (Wratcher, 1997). Educators have, for many years, noticed that some students prefer certain methods of learning to others. These dispositions, referred to as learning styles, form a student's unique learning preference and aid teachers in the planning of small-group and individualized instruction (Kemp, 1998).

According to (Felder, 2000), learning styles are “characteristic strengths and preferences in the ways we take in and process information”. There are a lot of discussions about what learning styles are and at which level of cognitive abstraction to describe these. Curry’s Onion Model (Curry, 1983), illustrates these different levels; see Figure 2 for an overview.

Figure 2: Curry’s Onion Model

The outer layer of Curry’s model examines instructional preference. Learning style models in this layer measure interactions with the environment. The main theory of instructional preference was proposed by Dunn & Dunn (Dunn, 1978). The middle layer of Curry’s model concerns an individual’s intellectual approach to assimilate information and encompasses many of the currently popular learning style theories. This layer is considered to be more stable than the outer layer because it does not directly interact with the environment. Kolb's experiential learning theory (Kolb, 1984) postulates the existence of four learning modes that combine to form two learning dimensions, concrete/abstract and active/reflective. The inner layer of Curry’s model examines cognitive personality style, addressing an individual’s approach to adapting and assimilating information. This layer is considered to be an underlying and relatively permanent personality dimension.

The Felder and Silverman Learning Style Model (Felder, 2000) overlaps the middle, information processing layer, and inner, cognitive personality layer. It classifies students along five spectra: sensing/intuitive, visual/verbal, inductive/ deductive, active/reflective, and sequential/global. Although students are classified on the five spectra, the assessment tool is a forty-four-question inventory developed by Felder and Silverman. It is available as a questionnaire that is submitted and automatically scored on the Web. The score is set of numbers that indicate preferences for each learning style.

Figure 3: Felder-Silverman Learning Style Spectra

(Felder, 2000)

Definitions / Dimensions / Definitions
Do it / Active / Reflective / Think about it
Learn Facts / Sensing / Intuitive / Learn Concepts
Require Pictures / Visual / Verbal / Require reading or lecture
Step by Step / Sequential / Global / Big Picture

Table 1: Preferences in the Felder-Silverman Model

(Felder, 2000)

Most of the research in learning styles theories in relation to educational technology are referred to as an aid in instructional improvement (Jarc, 1999, Ross, 1999, Howell, 2001). According to the literature, suitable educational software has the potential to benefit students with specific styles of learning. The reasons for this belief are the differences from learning with educational software to traditional classroom education. Unlike the classroom environment, learning with educational software is a self-paced type of learning, which enables the student to spend as much time as is needed to master the material. Educational software has the potential to provide an exploratory environment, in which students can learn through experimentation. Although scientific laboratory settings provide a similar environment, because of the lower cost of creating virtual environments, the range of possibilities can be much greater. Finally, with the development of the new multimedia technologies, educational software can become a highly visual learning environment, one that might also include pictures, animations, movies, and other visual modes of presenting information.

There has been some research on educational software as an instructional medium for meeting the different learning styles, but the experiences for improved learning have been mixed (Diaz, 1999) (Howell, 2001) (Grasha, 2000).

5 The Interactive Teaching Tutorial

The rationale for this experiment was to explore and find ways for improvements in educational software, especially teaching tutorials and distance learning frameworks. We therefore wanted to examine the potential of Learning Styles, and an Interactive Teaching Tutorial (ITT) was developed as a framework. Demonstrating learning styles in a distance-learning environment is a larger undertaking and went beyond the scope of this experiment. The results, however, are intended to be applicable to distance learning environments and might also give some ideas for next generation distance learning systems

Except for Intelligent Tutoring Systems (Burger, 2001), (Shang, 2001), there has been little research on systems that accommodate the needs and learning preferences of individuals separately. D. Howell (Howell, 2001) discusses that to increase the quality of the learning experience, the current “one-size fits all” systems need to be more personalized, and should try to accommodate the different learning styles of users.

Prior to designing a learning system that utilizes learning styles, however, a thorough domain analysis could be significant and save development time. Maybe the system would not benefit from learning styles, e.g. if the user group is homogeneous and the variation of learning styles is small. A system could then be developed concentrating on that particular user group without being concerned about a larger set of different types of learners.

The second objective of the ITT was to develop a system for the students in a computer science class at California Polytechnic State University, CSC 480 Artificial Intelligence.

The learning material contains concepts of artificial intelligence, in particular search algorithms. However, the framework of the ITT is intended to be flexible, so that it can be used for other majors and classes as well. The interactive part of the tutorial is also an interesting aspect, and is intended for teaching active learners that prefer “learning by doing” (which happens to be Cal Poly’s motto).

5.1 Feature Driven Development and User Centered Design

Developing the ITT fell into two major categories: establishing the feature set of the framework, and developing the learning material. The process for developing the framework is based on an agile development process, the Feature Driven Development (FDD) (Coad, 1999) for determining goals, analysis, design, and then implementing each feature in iterations. For developing the learning material, principles from user-centered design were used (Truchard, 1998). For both parts, Felder and Silverman’s Learning Style Model (Felder, 2000) was adopted.

An agile process like FDD is a team-oriented process that delivers working software early and welcomes changing requirements (Beck, 1999). Every iteration of the development was a new version of the ITT implemented with features that were discovered or stated. An iteration is a 1 to 2 week period of work done by the development team. A feature set is a group of related features ("evaluating a student" or "enrolling a new student"). Feature sets can further be grouped into major features ("student management"). This grouping of features aids with reporting on the progress of a project. Features are then assigned to iterations based on the user and development priorities and on the number of hours available. Features are similar to requirements, but focus also on how the software should perform, not only what the software should accomplish.

This FDD process is divided into five different steps by (Coad, 1999): 1.) Develop an overall model, 2.) Build a Feature list, 3.) Plan by Feature, 4.) Design by Feature, 5.) Build by Feature. The FDD followed the steps in the order requirements, system description, analysis, overall architecture and {design, implementation and testing} within each iteration.

5.2 Case Study – “ Artificial Intelligence Search Algorithms”

Since Artificial Intelligence is a large subject, the subtopic search algorithms was chosen for the case study. This included problem solving by search strategies and informed search algorithms (Russell, 95). Among the algorithms that were studied: Depth-First Search, Breadth-First Search, Uniform Cost Search, Best-First Search, Heuristic Functions, Memory Bounded Search, and Iterative Improvement Algorithms.

The class instructor maintains and updates the system. He provides learning material and quizzes, and also manages the quizzes. Since the instructor is normally very busy lecturing and grading, time and effort put into maintaining such a system, should be as effective and effortless as possible, or supported by administrative personnel.