Self-directed Learning and Tutorial Assistance in a ... - CiteSeerX

Self-directed Learning and Tutorial Assistance in a Remote Laboratory Andreas Böhne1, Nils Faltin and Bernardo Wagner Learning Lab Lower Saxony

Key words: collaborative learning, remote laboratory, self-directed learning, tutorial assistance. Abstract  A shortcoming of most remote laboratories is that they do not support self-directed team learning and that they do not provide the help of a remote tutor. We describe the technical and pedagogical concept of a distributed internet assisted laboratory experiment (I-Lab). The experiment will be used in the laboratory for process control for students of electrical engineering in the main section of their study. Students will develop an embedded controller program which serves the deflection unit of a laser system and generates a picture. Participants act in distributed virtual teams and communicate with each other and a tutor by synchronous tools like videoconference and chat. They observe the lab devices by a user-controlled web cam. Student motivation, learning process and learning outcomes will be evaluated comparing a selfdirected and a teacher-directed learning setting.

1 Introduction Laboratories are important elements in technical education. They allow applying and testing theoretical knowledge in practical learning situations. Students working in a local lab can directly see, hear, touch and smell the laboratory devices. This full impression of being in the lab cannot be provided by remote labs where students interact with remote devices through a computer user interface. That is why remote labs are called by Aktan the “Second Best to Being There” [1]. So students should have the opportunity to work with local labs to get a direct “hands on” experience. But there are good reasons to provide students with remote labs too. Remote labs can be shared by many institutions and students worldwide. Given a constant amount of money, students can have the opportunity of access to many more remote labs than local labs. As an example, in the German LearnNet network each of the seven universities provides a remote lab to all members [8]. Distance education and learning-on-the-job students benefit even more from remote labs, as the cost and time needed for traveling to a local lab would often prevent them from using such labs. Working with remote experiments will develop engineering skills like remote maintenance of devices. When students work in distributed teams they also learn how to act in remote

1

Learning Lab Lower Saxony (L3S), Expo Plaza 1, 30539 Hannover, Germany {boehne, faltin, wagner}@learninglab.de

Interactive Computer Aided Learning Conference 1

September 25-27, 2002, Villach, Austria

collaboration. Some experiments in areas like nano physics [7] can not be accessed by human hands and eyes anyway, so students do not loose anything by working remotely. Instead of building a remote lab one can in many cases develop a software based simulation of the experiment. While the cost of developing a veridical simulation can be high, once developed it can be copied and distributed at low cost, so cost does not limit the number of simulation users. While a remote lab can be shared by many students working at different times, only one student at a time can use a lab device. To provide access for several parallel users to one experiment type, one has to build and maintain several instances of the experiment devices, which may result in a substantial cost per student. But remote labs do have some advantages over simulations which stem from the point that remote labs include reality into the experiment: (1) A goal of engineering education is to prepare students to cope with problems of real devices and systems. While simulations tend to be too simplistic, remote experiments naturally demonstrate problems of measurement tolerance, device failure and difficulties to return to a start state. (2) Remote labs allow the verification of scientific theories. In the Scientific Method, theory has to be evaluated by testing hypothesis in real experiments. (3) As an evaluation of a remote lab showed, students like to perceive and influence reality [13]. Remote laboratories are currently being developed and used at many places around the world. A recent survey of remote labs on the internet identified 19 projects [12]. Our experience with remote labs is that you often do not get the feeling of being in the lab or of working with real devices. It is hard to tell if the experiment is performed in reality or just faked by prerecorded pictures and videos or generated by a simulation. One would wish to get instant feedback and to see and hear the remote lab location through live video and audio. To get a feeling of working with a real experiment it is helpful not only to see controls and displays but also the complete experiment desk and the surrounding lab with people sitting there or walking by. Another limitation of many current remote labs is the lack of support for distributed group work and tutor help. Even when working independently to perform experiments or program devices, students typically need help at some time by fellow students or a tutor. So there should be some form of synchronous communication available and the user interface at the different locations should be synchronized to reflect the current lab state.

2 The I-Labs Project and its Remote Experiments Our project “Internet assisted Laboratories (I-Labs)” is a cooperation of the Stanford Center for Innovations in Learning (http://scil.stanford.edu), California, and the Learning Lab Lower Saxony (http://www.learninglab.de), Germany, within the Wallenberg Global Learning Network. In Stanford optical experiments for physics education and in Hanover mechatronic devices for engineering education are developed. By working with a remote lab students shall learn to program, maintain and supervise remote devices. The I-Labs project will develop didactic concepts for online laboratory usage and reusable software and hardware components. Together with newly gained design knowledge it will make it easier to develop new online labs. In one of our experiments, students program an industrial controller (PLC) for a process engineering plant. In this process, a fluid can be moved between several tanks and mixed and heated in a reactor. The experiment has been evaluated with university students of electrical Interactive Computer Aided Learning Conference 2

September 25-27, 2002, Villach, Austria

engineering [13]. It has subsequently been adapted for use in technical schools. We will measure students’ knowledge gain, problem solving strategies and motivation comparing a local and a remote learning setting. In the local setting students and the experiment are in one room. In the remote setting, students and the experiment are in separate locations [14]. We are currently developing an experiment for picture generation by laser deflection. Students will write embedded Java programs to generate pictures. The new remote lab reuses several components of the previous experiment for handling fluids. Both labs are used in a problem based learning setting. We decided to develop the picture generation experiment because of the following criteria: • The learning subject fits into the course on industrial process control for engineering education. • It is maintenance free, so it can be provided for long term use. • It gives users a visible result and provides high feedback, which we expect to lead to high student motivation. • It can be displayed and run in a public place.

Figure 1: Laser experiment components The experiment system consists of a green 3mW laser beam which is deflected by two mirrors of a galvanometer scanner in X- and Y-direction (fig. 1 and 2). The scanner can move the beam to 30,000 positions per second. The deflection is controlled by two analogue inputs to the scanner. They are driven by DA-converters controlled by an I/O-Card in an embedded system. A web server is running on the embedded system. It allows users to control the I/O signals through a web browser. Real time Java programs running on the embedded system generate pictures with the laser system. Programs can quickly turn the laser beam on and off and position the mirrors of the scanner. The embedded system uses a real time earliest deadline first scheduler. When a new task is installed, CPU time and deadline have to be provided. Jobs without real time requirements can be performed with standard Java threads. Interactive Computer Aided Learning Conference 3

September 25-27, 2002, Villach, Austria

Figure 2: Block diagram laser experiment Programs for the embedded system are developed in a web based programming environment (figure 3, lower part). The programming environment allows editing of Java program files in the editor window. Files can be loaded and saved on the lab server with the buttons and input fields on the left. The lower left of the screen is for compiling and running Java programs. Java files are compiled on the lab server (button “compile”) and compiler messages are displayed in a separate window. To start a Java method, the user asks for a list of Java methods in the file (“inspect”), chooses a method (“select”), enters actual parameters in a separate window and calls the method (“start”). If the program works correctly, the laser system will generate a picture on the canvas.

Interactive Computer Aided Learning Conference 4

September 25-27, 2002, Villach, Austria

Camera control functions (pan, tilt, zoom)

Camera selection

Video picture Telepresence Java program editor Files and folders

Methods and parameters

Programming environment

Compiler messages

Figure 3: Web based lab environment - telepresence and programming

The upper part of the window in figure 3 allows students remote access to the lab and is thus a substitute for being present in the lab (“telepresence”). They can select among four cameras to see a live video of the experiment desk, the canvas, the status lights of the devices or the moving galvanometer scanner. The canvas camera allows users to pan, tilt and zoom in on details. Users with access right can switch on and off the embedded system. In future, a microphone will transmit the sound of the operating scanner via an audio server. In a first camera setup the canvas camera always showed the entire canvas. However, fine details of the generated picture were not discernible, so we had to replace it with a user-controllable camera. Allowing users to pan, tilt and zoom also adds to the joy of use and to the feeling of being in the remote lab. Figure 4 and 5 shows an example of zooming in on a detail of a generated picture. The right picture was taken slightly after the left picture. In both video pictures the laser beam drew the lines twice during the snapshot. As the beam positioning is not exactly the same some lines are visible twice. The web based lab environment also contains a scheduling component (not shown in figure 3). It allows administrators and users to reserve time for using the lab. The program development environment and the scheduling component are implemented with Java Server Pages (JSP) and a MySQL database.

Interactive Computer Aided Learning Conference 5

September 25-27, 2002, Villach, Austria

Figure 4: Video picture of whole canvas

Figure 5: Zooming in on a detail. Taken slightly later.

3 Educational Concept and Research Questions 3.1 Problem-based Learning A major goal in engineering education is that students acquire problem solving and creativity strategies so that they become able to construct technical systems. Such strategies can be learned by working on small problems and construction assignments in a problem-based learning environment. Reinmann-Rothmeier and Mandl [11] have developed a concept of problem-based learning that combines principles from situated learning theories with instructional support. They follow a moderate constructivist view of learning where learning is seen as an active, self-directed, constructive, situated and social process: • Students have to become active and to develop motivation and interest in the subject. • Students are responsible for planning and controlling their learning process, even when part of the planning and control is provided by a teacher or learning environment. • Students construct new knowledge by integrating information and experiences in their existing knowledge structure. • Learning always takes place in a specific situation. It gives a background for interpreting experiences and will enable and restrict the experiences. • Most learning situations include social interaction with a teacher or tutor and other students. Reinmann-Rothmeier and Mandl propose the following guidelines for problem based learning environments which they derived from the learning theories of anchored instruction, cognitive flexibility and cognitive apprenticeship. 1)

Situated and authentic learning. Letting students solve realistic problems will enable them to apply their knowledge in practical situations. Authentic problems will motivate students.

Interactive Computer Aided Learning Conference 6

September 25-27, 2002, Villach, Austria

2)

Learning from multiple perspectives and contexts. Solving several related problems which all require the same basic skill will make it more likely that the skill can be applied in a different practical situation.

3)

Learning in a social context. Learning and problem solving should be done in student groups. If students have experts as models they can acquire skills, knowledge and orientation.

4)

Learning with instructional support. Complex problem solving tasks without instructional support will often demand too much from the students and will lead to ineffective learning. Many students need instructions and help for solving problems. The learning environment shall provide knowledge for solving the problem.

Several pedagogical psychology experiments confirm these guidelines. For example one experiment tested the influence of multiple perspectives and instructional support on learning results for students learning bookkeeping [5]. Four groups of students received different pedagogical treatments. Students learned under single perspective by reading one example per topic or multiple perspectives by reading several examples per topic. Another difference was the instructional support for learning problem solving strategies. Some groups were instructed to describe the problem solving strategy they just had seen in the solution example, whereas other groups did not receive any such guidance. Students learning under multiple perspectives and receiving no support where overstrained with the task and had the worst learning results. Best learning results were obtained by the group learning under multiple perspectives and receiving instructional support The educational concept for our new laser experiment lab is problem-based learning. The context is situated and authentic because programming an embedded system is a relevant task in engineering practice. We will provide students with multiple perspectives and contexts by letting them work on different levels of our laser experiment. They will analyze the mechanical dynamics of laser deflection, program real time control, analyze the limits of webcam video and audio transmission, program part of the vector graphics library and program static pictures and dynamic animations. Learning takes place in a social setting because students will work in groups of two or three with the experiment. Students receive instructional support through work tasks with appropriate difficulty, background information hypertext and through a human tutor who assists them in the problem solving and programming task where necessary. The constructivist view of learning demands that students become responsible for planning and controlling their learning process and thus become self-directed learners. In the following sections we describe our educational concept of self-directed learning and how it relates to tutorial assistance and collaborative learning. 3.2 Self-directed Learning Self-directed learning seeks to put the learner as much as possible in control of the learning process. It is the learner rather than an institution, a teacher or a learning program who chooses the educational objectives, emphases and learning strategies. Learners set their own goals, analyze a given problem, observe the learning progress and assess the learning results. However, it is not practical to demand completely autonomous learning. For example even when a web based course gives the learner a maximum of autonomy the contents and structures are typically

Interactive Computer Aided Learning Conference 7

September 25-27, 2002, Villach, Austria

fixed and thus limit self-directed learning. Self-directed learning seeks to maximize the learners’ freedom of choice in the following dimensions [9]: • learning goals (relating to the subject or the aspiration level) • media and resources • learning place • learning time, speed and duration • social setting (e.g.: group learning or learning alone) • learning strategy (e.g. depth and sequence of processing subject elements). Our remote labs are web based. Concerning the dimensions above, web based learning has a high potential to support self-directed learning in the following aspects [4][3]: • Learning is independent of time and place. • The hypermedia structure supports the exploration of the learning content. • Learners need feedback to sustain motivation. When no teacher is available, interactive elements can provide some feedback. • The learning environment can adapt the learning materials to the needs of the learner. Learners in continuing education have a special need for flexible, self-directed learning. They need to choose the learning content according to their actual needs. It is not practical to demand that they work through a large predefined curriculum. As they have a fixed work schedule they need to choose the learning time and place. Web based learning will save travel time and cost. Self-directed learning poses high demands on the personality of the learners because of their responsibility for the learning process. That is why the demands of the learning environment should be in balance with the competences of the learners for self-directed learning: “Consequently, self-direction in learning is a term recognizing both external factors that facilitate a learner taking primary responsibility, and internal factors that predispose an adult accepting responsibility for learning-related thoughts and actions. At the same time there is a strong connection between self-directed learning and learner self-direction. Both internal and external aspects of self-direction can be viewed on a continuum and optimal learning conditions exist when a learner's level of self-direction is balanced with the extent to which self-directed learning opportunities are possible.” [6]. For self-directed learning, a learner should have the following competences [9][10]: • self-observation, self-reflection, self-judgment • development of own objectives • reactivation of initial knowledge • autonomous organization of learning (e.g.: time management) • autonomous development of learning motivation and concentration Interactive Computer Aided Learning Conference 8

September 25-27, 2002, Villach, Austria

• development of learning and problem solving strategies • knowing when to seek the assistance of other learners or the tutor Often students are only used to learn in a teacher-directed setting, so that a long term goal should be to develop these competences for self-directed learning. Many students abort their learning when they get stuck in a web based learning environment and can not get help. To allow typical learners to perform well in a self-directed setting, they need support by a tutor. A tutor can help with organizing the learning process, solve technical problems and answer questions regarding the content. We think that a human tutor is needed as it is not possible to prepare every thinkable learning task and support information in advance. The general question in designing environments for self-directed learning is which degree of selfdirected learning is suitable for which goals and which students under which conditions. The evaluation of our remote laboratory described below will contribute to answering this question by comparing the effectiveness of two learning settings with a varying degree of self-directedness. 3.3 Tutorial Assistance In our educational concept a tutor assists the learners in their self-directed learning. The tutor resides at a remote location and communicates through synchronous media. He can support the learners in several ways: • He can give information and assistance in questions concerning the content or referring to technical problems. • He can stimulate the meta-cognition of the learners. • He can advise on the choice of the learning goals and acting goals. • He can give feedback to motivate the students. • He can organize the learning process and set time limits. Another task of a tutor in a remote lab can be the design, realization and administration of the learning environment. Moreover he can form groups, moderate group discussions and coordinate the interaction within the group. In our laser experiment the students will act in distributed teams and communicate with each other and a tutor by synchronous tools like videoconference, chat and application sharing. In our evaluation we will compare two settings with different tutor roles: • In the self-directed setting the tutor acts as an e-coach or e-moderator. He supports the students in learning to learn, moderates the group discussions and answers the emerging questions of the students. • In the teacher-directed setting the tutor acts as an e-instructor: He explains the tasks, guides the group in the learning process along the structure of the subject rather than the actual task (i.e. subject-oriented) and answers the questions in a subject-oriented way. We also want to investigate which technical tools for communication and collaboration are suitable for a remote lab with a remote tutor. Interactive Computer Aided Learning Conference 9

September 25-27, 2002, Villach, Austria

3.4 Collaborative Learning In our remote labs, groups of students work with an experiment. We decided to implement a collaborative approach because of the following reasons: • The social relatedness promotes the motivation of the students. • Students can help each other. • Students acquire social competence in team work. When students are located at different places they also acquire skills in remote collaboration. • The consulting effort per student decreases. Given a constant amount of tutor time, the possible number of participants, who can use the experiment, rises. Working in a group limits the freedom and self-directedness of a single student. The student must coordinate the preparation and implementation of the lab with the other students in his group. But the character of self-directed learning is not lost in a collaborative setting. Self-directed learning should not be confused with absolute autonomy. In our evaluation we will compare a setting where students are co-located with a setting where they are distributed over several locations. This will give us evidence whether student collaboration works sufficiently well in a distributed setting.

Figure 6: Variables of the research question

4 Evaluation Approach We will perform a comparative evaluation study. We are following the evaluation methodology of Borz/ Döring and distinguish independent, dependent, mediator and control variables [2]. Our main research question is: How does the distribution form and the degree of self control considering initial knowledge influence problem solving strategy, motivation, task success, consulting effort and communication behaviour? To this end, we will vary the independent variables distribution form and degree of self-directed learning (see figure 6).

Interactive Computer Aided Learning Conference 10

September 25-27, 2002, Villach, Austria

We will distinguish two different distribution forms: In the first setting all students are in one location, the experiment and the tutor are at remote locations. In the second distribution form each learner, tutor and the experiment are at separate locations (see figure 7).

Figure 7: 2x2 design of independent variables Regarding the degree of self-directed learning we will implement two different educational settings (see table I): 1)

Self-directed. High degree of self-direction of the students in the dimensions learning goals, tasks and tutor role.

2)

Teacher-directed. Low degree of self-direction in these dimensions. TABLE I: Degrees of self-directed learning Self-directed setting

Teacher-directed setting

Learning goals After some predetermined tasks, and tasks the group develops the learning goals and tasks. The tutor approves goals and tasks. Role of the tutor

e-coach, emoderator

Predetermined tasks with increasing difficulty.

e-instructor

An important point in self-directed web based learning environments is the students’ ability to freely choose learning time and place. Unfortunately it is very difficult to perform a comparative study with such a setting. Several variables like work environment, work time and available Interactive Computer Aided Learning Conference 11

September 25-27, 2002, Villach, Austria

support materials would be very different from student group to student group and hard to measure. The variability of these variables would endanger the reliability of our study. We plan to perform a qualitative study in the future, where students are free to choose learning time and place. We have defined the dependent, control and mediator variables in the following way: When students are distributed they have to communicate over synchronous communication media like video, audio, whiteboard, chat and application sharing. When students are co-located they can communicate directly. The communication opportunities are the media, which are available for the students for communication among the students and for communication with the tutor. The media used for the interaction student-experiment (e.g.: video, audio, programming environment) are the interaction opportunities. With communication behavior we describe the media, which are really used by the students during the experiment. The content of the communication will be logged and summarized qualitatively. The consulting effort is the time effort of the tutor for the support of the students. We will differentiate between the support for technical questions, questions concerning the content, learning process and moderation of the collaborative work. The task success is evaluated by the tutor. He assesses the results (e.g.: laser picture, Java program) of the group by comparing the results with other groups and with the pre-determined goals. The motivation is distinguished in types of extrinsic and intrinsic motivation. Problem solving strategies are mental represented schemes to find operations for solving concrete problems. The initial knowledge is measured with a questionnaire before the experiment, because we think, that it is an important factor in a self-directed learning setting. For example, students, who only have little knowledge of Java programming will have problems with selfdirected learning and need special support of the tutor. Students with experience in programming will be able to work mostly independently. We will use qualitative and quantitative methods for the acquisition of the necessary data. To measure the dependent and control variables we will design questionnaires and pre tests. We will observe the learners’ activities during the experiment by logging all communication and interaction. In the winter term we will test our laser experiment and the assignments with a small number of students. The main focus will be the usability of the lab environment and the difficulty level of the assignments. The results will help us to optimize the remote lab and the assignments. In the summer term 2003, students of electrical engineering will take part in the laboratory for industrial control. There we will perform the evaluation described above. In case only few students take part in the laboratory, we will only vary the degree of self-directed learning with co-located students (settings D1H and D1L in figure 4).

5 Conclusion In the I-Labs project, we develop an educational concept for collaborative, self-directed learning with tutorial assistance in online laboratories. Also, reusable software and hardware components for remote laboratories will be delivered. We are currently working on a experiment for picture generation by laser deflection. Up to date only a small number of remote labs have been built in research projects. The future will show which organizational model will succeed in providing a large pool of remote labs for Interactive Computer Aided Learning Conference 12

September 25-27, 2002, Villach, Austria

continuing use. Remote labs can be shared by a network of educational institutions where each institution provides labs for the participating partners. Another model for the provision of remote labs is that commercial providers maintain the labs and lease lab time to educational institutions. In any case remote labs will be used by institutions with different curricula and student background. This calls for a flexible educational setting like the one described above based on self-directed learning and tutorial assistance.

6 Acknowledgements We like to thank Christoph Richter from the Learning Lab Lower Saxony evaluation team for his support in designing the pedagogical evaluation. Nils Hagge and Jörg Tuttas from the University of Hanover supported us by developing the programming environment and basic embedded system software. Patrick Hohmann from the Learning Lab wrote test programs for picture generation with the laser system.

7 References [1] Aktan, B., Bohus, C.A., Crowl, L.A., Shor, M.H., „Distance Learning Applied to Control Engineering Laboratories”, IEEE Transactions on Education, Vol. 39, No 3., August 1996. [2] Bortz, J., Döring, N., “Forschungsmethoden und Evaluation für Human und Sozialwissenschaftler”, Springer, 2002 [3] Euler, D., “From connectivity to community – Elektronische Medien als Katalysator einer Kultur des selbstorganisierten Lernens im Team”, bwp@ Berufs- und Wirtschaftspädagogik - online, No 2 , June 2002. http://www.bwpat.de [4] Geyken, A., Mandl, H., Reiter, W., “Selbstgesteuertes Lernen mit Tele-Tutoring“, In: R. Schwarzer (ed.), MultiMedia und T eleLearning. Lernern im Cyberspace, Frankfurt, 1998 [5] Gruber, H., Mandl, H., Renkl, A.: „Was lernen wir in Schule und Hochschule: Träges Wissen?“ pp. 139-157 in: Mandl, H. and Gerstenmaier (Ed..): „Die Kluft zwischen Wissen und Handeln“. 2000. Hogrefe. [6] Hiemstra, R., “Self-Directed Learning”, In: Husen, T., Postlethwaite, T.N. (Eds.), The International Encyclopedia of Education, Second Edition, Oxford, Pergamon Press, 1994. [7] Kapitan, C., “Electrical engineer imagines future where laboratories fit on a dime”, Stanford Report, November 28, 2001 http://www.stanford.edu/dept/news/report/news/november28/hesselink-1128.html [8] LearnNet project web page. http://learnet.esr.ruhr-uni-bochum.de/ [9] Niegemann, H. M., “Computergestützte Instruktion in Schule, Aus- und Weiterbildung”, Peter Lang, 1995. [10] Rautenstrauch, C., “Tele-Tutoren: Qualifizierungsmerkmale einer neu entstehenden Profession”. Bertelsmann, 2001. [11] Reinmann-Rothmeier,G., Mandl, H.: „Unterrichten und Lernumgebungen gestalten“. Chapter 13 in: Krapp, Weidemann (ed.): „Pädagogische Psychologie“. 2001. Beltz. [12] Teichmann, T.: “Übersicht über Fernlabore”. 2002. http://www.learninglab.de/i_labs/docs/Uebers_Fernlabore/uebersicht_fernlabore.html [13] Tuttas, J., Wagner, B., “Distributed Online Laboratories”; International Conference on Engineering Education (ICEE01), Oslo, Norway, August 6-10, 2001. [14] Tuttas, J.,Wagner, B., “The Relevance of Haptic Experience in Remote Experiments”, World Conference on Educational Multimedia, Hypermedia & Telecomunications (ED Media 2002), Denver, USA, June 24-29, 2002

Interactive Computer Aided Learning Conference 13

September 25-27, 2002, Villach, Austria