This project combines both interactive computer-based instruction with data communications to dynamically create groups of distant education and on-campus students in a constructivist virtual classroom. Groups are formed to focus on design issues.
Students at Charles Sturt University, Australia, learn object-oriented design and programming using software environments which are stand-alone. A project is being undertaken to interlink the environments of students when needed. In each environment will be an "expert" monitor program that can determine if a group session is desirable for that student. If the student agrees, the "expert" monitor advises the "expert" monitors of other students so that a group session can be set up among students with similar problems. A human tutor from Charles Sturt University will also be involved. Commercial software will be used to exchange design and programming code so that each student sitting at their machine can not only see other students' code but also any contribution the tutor makes.
Smoothly combining individual learning with group environments focussed on specific design problems allow students time to both construct new design concepts from their own prior experience and time to construct new design concepts using social groups.
Traditionally system design and programming have been taught in different courses leading to a complete separation of the two activities in students minds. Object Oriented Technology overcomes that separation by iterating design and implementation through a number of prototype stages. Now the challenge for the teacher is to provide an environment where implementation is seen as a design explication by the student.
Constructivist educational learning theory indicates that people only radically reconstruct concepts if the concepts that they are using are unsuccessful. This reconstruction occurs for computing students only when the design implementation fails; the failure usually manifesting itself as an error - a bug. If we are not careful, students will see all their failures solely as programming failures and not as the design failures that many are. This misperception causes the student to concentrate their efforts on fixing the bug at the wrong level.
Analysis of the programming environment's log file show that attempting to correct bugs at the wrong level results in definite patterns in the log file. Students tend to persist in this behaviour when their understanding of design in programming is deficient or when the teaching support is inadequate in their environment.
In the former case, intervention by a teacher is particularly satisfying as often the continued failure to succeed prepares the student for a major change in design concepts. If the teacher does not intervene, at best a fudged solution is the result; at worst the frustration causes the student to drop out.
The latter case of poor teaching environment is particularly relevant to distance education students who have no easy access to lecturers or tutors. Even a part-time student can have difficulties if their role models at work are all structured programmers.
A project is underway to provide software monitoring of the log to allow intervention at the point where it is possible that a student is not design oriented. Small purpose-driven discussion groups of students with a tutor are then organised for students at approximately the same level. Because many of Charles Sturt University's students are distance education students, the groups are being organised via the Internet.
New concepts are structured from old concepts by specialisation, generalisation, association and abstraction (Ausubel, quoted by Novak (Novak 1978)). A reconstruction of concepts is only possible if
The software monitor evaluates the log while the student is implementing their program. When it appears that the student is not successfully progressing, the monitor negotiates with the student for a group session. If the student agrees, the monitor alerts other monitors which can evaluate if their students are "near enough" to the same situation to profitably use a group session. These monitors negotiate with their students for a group session. A group session will also include a human tutor from Charles Sturt University wherever possible.
The negotiation sessions between the students and the monitor programs fulfil the first two conditions by ensuring that the students are dissatisfied with their current progress and also believe that a group session will assist. In many cases the dissatisfaction may be sharpened when the monitor provides evidence of current poor progress, undesirable patterns of working, or past successes in group sessions.
The need for intelligibility of the group discussion is addressed by the requirement that discussions are based on students' work rather than on a lecturer/text-book sample. The comparison with other students' work, and perhaps with tutor-provided fragments, will enable the students to construct new ideas. Plausibility is provided by immediate testing against the compiler - a correctly executing program will suggest that the new ideas are better than the student's. Fruitfulness of the group session will be achieved because similar level students are dealing with similar problems in a session with a concrete goal.
The concreteness of the goal means that the software provided for communication can be more specific than Internet chat environments. The interchange of designs and reasonably large programs will need to be necessary to allow each student sitting at their machine can see both other students' code and any contribution the tutor makes. The interchange will be done using commercial software for distributed system development. The student interface will use modified virtual world software that networks learners and mentors who interact using designs and programs as props (Hughes & Moshell, 1995).
There has been considerable work on components of this project. Expert (rule-based) tutorial systems have been examined (e.g. Anderson, Boyle, Corbett and Lewis(1990)). The value of electronic group communication in a constructivist learning environment has been shown by such projects as ExploreNet Hughes and Moshell, 1995). We intend having a similar learning environment though with an interface suitable for mature students undertaking programming. There has also been a number of applications where remote students have interacted with teachers in real time.
However, the author is not aware of real-time constructivist classes where the members are dynamically selected on grounds of similarity in cognitive distance between currently held concepts and a particular target concept.
The School of Information Studies, Charles Sturt University, Australia has been teaching computing since 1979 for both on-campus and distance education students. We have been teaching Object Oriented Programming (Smalltalk) to undergraduate students since 1989. Indeed, all core subjects in the Bachelor of Information Technology (except database), now use the object-oriented paradigm. Currently approximately 260 students are taught Smalltalk in introductory programming subjects each year.
Since 1992, a research program has been examining ways to assist students to learn the paradigm of Object-Oriented Programming. A rule based tutor was found to involve more rules than could be handled on a personal computer, even if pruned to advise only on the most likely errors (Whitelaw 1993, Whitelaw 1994, Whitelaw & Weckert 1995). however, in a study at Carleton University, Ottawa, Canada in 1995, it was found that an evaluation of the log file could detect characteristics that were consistent with the student b being in difficulty. (For example, one indicator is excessive length of time spent on a particular section of the problem) . These difficulties could be resolved quicker by discussion with others rather than continuing an excessively long session working on a problem. Currently the monitor is being programmed and will be tested with on-campus students this year. We are seeking a grant to implement the whole system for the second half of 1997.
While this project will immediately support students in the second year subject of the Bachelor of Information Technology (Information Systems) it is also relevant to any subject where (i) a social constructivist situation will enhance learning, (ii) a significant amount of student activity is machine based (iii) measures can be defined that indicate a student is likely to be having difficulties, and (iv) students can be connected by electronic means. Such a learning environment will improve the design quality of assessment presentations. It will also reduce student frustration so will mean a lower drop out rate.
Creation of an effective learning situation that can be accessed at any time of the day has particular significance to distant education students who not only enrol because of geographical distance from an University but also because of access problems caused by shift work, disability, social setting, or financial limitations. Such a learning situation will reduce the need for other less effective tutorial media (phone, fax, email) and will reduce or remove the need for resource-expensive residential schools.
On-campus students will also benefit from a learning situation that can be accessed when they have a need rather than waiting for a session set by a timetable. Cross fertilisation between the two groups will benefit both and the need for tutorials on-campus will be reduced.
More generally we would see that the removal of geographical limitations on educationally effective teaching has implications for supra-national teaching.
Anderson, J.R., Boyle, C.F., Corbett, A.T., & Lewis M.W.(1990) Cognitive modelling and intelligent tutoring. In W.J. Clancey & E. Soloway (Eds.), Artificial intelligence and learning environments. Cambridge MA: MIT Press.
Gunstone, R.F., (1990) Children's science: A decade of developments in constructivist views of science teaching and learning. Australian Science Teachers Journal 36(4) 9-19.
Hughes,C.E., & Moshell, J.M., (1995) Shared virtual worlds for education: the ExploreNet experiment. to appear in ACM Multimedia Systems
Novak, J.D.(1978) An Alternative to Piagetian Psychology for Science and Mathematics Education Studies in Science Education 5 1-30
Posner, G.J., Strike, K.A., Hewson, P.W., & Gertzog, W.A., (1982) Accommodation of a scientific conception: Toward a theory of conceptual change. Science Education, 66(2), 211-227.
Whitelaw M W., (1993) Relating Programming Strategies Using Constructivism Proc 3rd Aust Conf on Reasoning 1993
Whitelaw M W., (1994) On the Cognitive Load of Learning Loop Construction Viewpoints 3
Whitelaw M W., & Weckert J., (1995) "The Humaness of Object-Oriented Programming". Proc First Conf Cognitive Tools
Michael Whitelaw, Msc Syd., is a lecturer in computing specialising in programming languages and artificial intelligence. He has been on the academic staff of Charles Sturt University since its inception. His current research interest is the changes in conceptual structure necessary for a structured programmer to develop into an object oriented programmer.