Full Paper

Lynette Schaverien and Mark Cosgrove

l.schaverien@uts.edu.au, mark.cosgrove@uts.edu.au

University of Technology, Sydney

The theme of this conference, What works and why? permits us to

• demonstrate two learning environments designed to confront well-known problems in education,
• describe these products and reverse engineer them back through the design stages,
• present some evidence of successful use, and
• discuss these learning systems from the point of view of the underlying educational theory on which their design is based.

We locate our stories at a cross-road where the educational practices of one era are intersecting with those being formulated for another. As Tiffin and Rajasingham (1995) point out (Table 1),

It is our view that present technologies can deal effectively with the provision of education anytime and anywhere. Here, we aim to provoke discussion of information era teaching methods and learners' roles by suggesting that the two learning environments we will demonstrate, underpinned as they are by a particular learning theory, might well constitute prototypical examples of what Tiffin and Rajasingham (1995) envisage.

We have developed a theory of learning (Schaverien and Cosgrove, 1997) which posits that the heuristic, generate-test-regenerate (Minsky, 1985; Plotkin, 1994), is natural, and, as such, available to all learners since it is built into the brains and genes of all living things. This theory grew out of two major empirical studies, reports of which have now been published in international journals (Cosgrove, 1995; Schaverien and Cosgrove, 1997). The findings of these two projects suggest

1. that students' learning of seminal science ideas is well-explained as a process of generating and testing (hence selecting) of ideas on their value, that is by means of a Darwinian generate-test-regenerate heuristic, and
2. that teachers' learning to teach (in ways that support such natural learning) can be understood in terms of that selectionist heuristic as well.

However, these reports demonstrate that although the radical re-learning required in both cases is possible, it occurs at a price: those interventions by a teacher-cum-mentor which brought this re-learning about were intensive, making them expensive and difficult to replicate. Many teachers lacked insight into them and lecturers considered them undignified in tertiary lectures.

In demonstration 1 the learning environment is a Generative Virtual Classroom. This World-Wide-Web-based learning environment permits students to observe and discuss authentic examples of children learning. Here, students can deepen their insights, over time, into teaching events by:

1. viewing digitised video excerpts of children learning, with their transcripts,
2. recording, in a database, features of children's learning, including the key ideas children generate, how they test those ideas, the progression of those ideas during the chosen video excerpt, characteristics of children's conjectures and other insights they have into those learning events (such features being critical to an appreciation of science learning as knowledge generation and refinement),
3. searching the expanding database archive for their own views and those of other students,
4. listening to pre-recorded audio commentaries which cast each learning event as a significant example of knowledge generation,
5. conversing with each other and with their tertiary teacher (by means of a supporting threaded email discussion group) with respect to the learning they observe, and
6. reflecting upon their own learning and teaching as a conceptualising process, that is, as knowledge generation.

The project has links with three established World Wide Web based learning environments:

1. The Virtual Hospital at the University of Iowa Medical School.
2. The Australia Street Archive, in which users write to the archive, adding to its richness and versatility.
3. Learning Constellations which employs video-ethnographic methods, emanating from the Media Laboratory at MIT to capture authentic science learning and teaching in local classrooms

As well, the project has a well-recognised research base. That research suggested that in order to adopt innovative science teaching approaches, teachers needed to be helped to relinquish their view of learning as occurring only by being instructed. Only then could they perceive learning as generative (after Wittrock, 1974; Minsky, 1985; Edelman, 1992 and Plotkin, 1994), as occurring when learners generated ideas and tested them on their value. The support of a mentor appeared to be critical in assisting teachers to see more fruitful ways of teaching. Furthermore, that mentor support, with its sustained and frequent conversations between teachers and the mentor, appeared to contribute, by its very nature, to teachers' developing understanding of generative styles of teaching. So, this project preserves those aspects critical to successful mentoring, designing them into an electronically mediated learning environment. The GVC encourages students to attend to the development of children's ideas through electronically mediated conversations.

In demonstration 2 the learning environment is a free-standing (= anywhere, anytime) laboratory in which individual students can identify and examine their ideas about a conservation principle. This environment, incorporating screen-based images (including animation) and simple laboratory equipment, assists students, first, to diagnose the status of their own knowledge and to compare that with scientists', and secondly, to generate and test a new theory from which accurate predictions of circuit phenomena can be made.

Evidence of the success of these two learning environments is accumulating.

The effectiveness of the first of these learning environments is being investigated by case study. At the time of writing, the first phase of development of the Generative Virtual Classroom has been completed. It exists as a fragile, working prototype. As part of this first phase of development, we have conducted two case studies so far.

In the second of those case-studies two students worked with the Generative Virtual Classroom, unassisted, in sessions ranging from one to three hours in length, once or twice a week, over a period of three weeks, spending approximately six hours altogether. We did not observe these sessions. At the end of the sessions, we audio-taped a conversation with one of the students (Susan). On analysis, this conversation reveals evidence of learning reminiscent of teachers in the mentor-supported research projects on which the design of the Generative Virtual Classroom was based. For example, in that previous research, one teacher (Jan) came to realise that her daughter Sarah's actions in connecting up a circuit, far from being "primitive" as she had earlier described them, were actually tests of Sarah's particular theory. Similarly, Susan not only began to deepen her criteria for appreciating the worth of what children did in the Generative Virtual Classroom, but she was aware that she was doing so.

Jan's and Susan's appreciation that these children were actually generating and testing their ideas, in circumstances where these teachers had earlier failed to recognise that, is a critical indication that they are beginning to develop a generative view of learning. That Susan can come to this same significant realisation as Jan, but in a computer-mediated learning environment, is a promising indication that the design of this learning environment has been successful in preserving those features of the mentor-supported learning environment that are critical for developing a generative view of learning.

The principal users are graduate students in Science and Technology Education subjects in the Graduate Diploma of Education. The laboratory was presented in a flexible learning module.

Volunteers in a study of learning through interactive multimedia. N = 20, Names = 01, 02, 03 .... 20. Mean age = 26 years, mode 22 ± 1 years, with 10 females. The survey enquired into their physics education; first, participants were asked to indicate the most advanced level at which they had studied physics (varying from year 10 to year 15). Next, the participants were asked if they enjoyed learning Physics. Of those (10) who did, eight were male. The security of their knowledge about electricity was established in a question, 'Do you have a deep understanding of the topic, 'Electricity'?

Next, students completed six questions about simple circuit phenomena, such as drawing a circuit for a torch, placing lights in series and parallel, commenting on relative brightness levels of lights and expiry rates for cells in different circuits. Typical responses for one question is listed in Table 2.

Such a proliferation of responses is an effective probe of students' confused knowledge states and some would say, a profound worry on two scores; first, that even after extensive teaching these students cannot use the scientific way of thinking about even very simple phenomena, and second, that these people were expected to teach secondary school students about this phenomenon.

Once this survey had been analysed the results were returned to the students in the form of a report which presented

1. a summary of research literature on the issue of deep versus shallow learning,
2. the results for all members of the class, reported anonymously, and

suggestions for further work, including individual diagnosis using the information in Table 3 with which students diagnosed their own learning state and designed the treatment needed.

With this information students were asked to comment upon the results of the first survey and then to devise a plan to achieve deeper understanding. An extract of one student's comments in presented in Table 4.

Some weeks later, the students were presented with a second survey with questions similar to those in the first survey, but with the further requirement that reasons or explanations be provided for each answer. The full data for the second survey is summarised in Table 5. Here, the first listed y indicates a correct answer, and a second listed y indicates a correct reason. X indicates wrong answer or wrong reason.

Compare the outcomes of survey 2 question 3 with the information in Table 2. In the second survey 18 students wrote the correct answer and 13 provided a correct explanation of the answer.

Research in multimedia and learning has been criticised by Clark and Craig (1992) on the basis that, "aside from obvious research design problems, the main obstacle to multi-media and learning studies is that they are conducted without any benefit of why one would expect differences in the first place". By contrast, the design of these two learning environments explicitly incorporates findings of previous research, leading us to expect that they will be effective in promoting student learning. Features designed into these learning systems which may be significant to the re-learning that both groups of students demonstrated include

• time
• multiple passes
• interaction
• interactivity with other students and others' views
• conceptual clarity
• individual responsibility
• self-diagnosis
• idiosyncratic pathways
• asynchronicity.

(c) Lynette Schaverien and Mark Cosgrove

The author(s) assign to ASCILITE and educational and non-profit institutions a non-exclusive licence to use this document for personal use and in courses of instruction provided that the article is used in full and this copyright statement is reproduced. The author(s) also grant a non-exclusive licence to ASCILITE to publish this document in full on the World Wide Web and on CD-ROM and in printed form with the ASCILITE 97 conference papers, and for the documents to be published on mirrors on the World Wide Web. Any other usage is prohibited without the express permission of the authors.