ICT: Puzzled peers are teaching tools, too

Vicarious learning, in which people learn from watching others learn, helps students improve their clinical reasoning, writes Richard Cox

Vicarious learning is the idea that people can learn via exposure to the learning experiences of others. An everyday example is the radio programme Gardeners' Question Time , in which a novice gardener engages in dialogue with an expert, but many overhearing gardeners benefit from their interaction.

The history of vicarious-learning research dates from the 1960s and 1970s, when Albert Bandura studied the conditions under which children copied aggressive behaviours modelled by adults or children. His work was driven by concerns about the effects of violence on television. One of Bandura's many findings was that children are motivated to imitate aggressive behaviour if the model is seen to be rewarded.

Now a research project that is part of the Teaching and Learning Research Programme is looking at students' vicarious learning in more complex and arguably positive contexts, making use of information technology to do so.

The project - which involves academics from Sussex, Sheffield, Newcastle and Edinburgh universities - is studying the vicarious learning of complex cognitive skills. The group has developed vicarious learning resources to support health science students in the challenging task of learning to reason clinically.

In the first phase of the project, the group studied speech and language therapy students at Sheffield and Newcastle as they diagnosed virtual patients who were using an established e-learning resource, the Patient Assessment Training System, or Patsy ( www.patsy.ac.uk ). This system is used in about 30 university departments in the UK and Ireland. It contains more than 60 virtual cases in four discipline areas: developmental reading disorders, neuropsychology, neurology and medical rehabilitation, and speech and language pathologies. Learners can use Patsy to "administer"

psychometric and cognitive assessment tests to the virtual patients. The system is typically used in blended learning contexts in which students attend clinical placements and tutors integrate online Patsy sessions with lecture content.

The students were given Patsy virtual patient cases that they had not seen before, under exam conditions. In the research, Patsy served as an interactive e-learning resource and as a research instrument. Students'

interactions with the virtual patients were analysed, and a wide range of clinical reasoning difficulties was observed.

The research identified two types of diagnostic reasoning impasse. One was a failure of general reasoning, such as not using efficient general strategies to hone in on a diagnosis. The second type of problem was subject specific - for example, when a student was not familiar with common clinical tests or had too little knowledge of speech and language disorders.

In the second phase of the project, vicarious-learning resources were developed to target each source of difficulty. Each resource consists of a short video clip of two students (or a student and a tutor) discussing a clinical-reasoning topic. The participants engage in "task-directed dialogue", a type of language game commonly used in teaching English as a second language.

In a typical "task-directed dialogue" activity, participants might be instructed to discuss a scenario in which a client seems to cope well in conversation with the therapist despite having very obvious expressive difficulties. What would be the first area of communication that you would look at and why? About 190 clips have been produced, and these are stored in a database extension to the Patsy system together with rule-based software for detecting when learners are having clinical-reasoning difficulties.

The choice of students rather than experts as models was deliberate. The traditional model of learning is that of direct instruction of students by tutors. But students also learn well from listening to and observing fellow students who are a little further down the learning track than they are.

Students often identify with their peers more than they do with their teachers. The student who is being observed often raises learning issues that are directly relevant to the observing student - ones that the student doing the observing can often identify closely with. Every teacher knows that if one student raises his hand to ask a question, it is probably one that several other shyer, quieter students might also have asked.

Another rationale for vicarious learning arises from indications that, in e-learning contexts, students often have less opportunity to observe and "overhear" the dialogue of others than in traditional teaching and learning. The case for educational dialogue forming a crucial part of the learning cycle in higher education has been convincingly made by Diana Laurillard in her book Rethinking University Teaching . The project group is also exploring vicarious learning's potential for facilitating professional enculturation and students' ability to learn to "talk the talk" of the profession.

In another line of research, the group is studying whether exposure to the clinical conversations of expert clinicians with students helps vicariously learning students to master their profession's "register" and its professional forms of discourse.

The project has already had an effect on clinical learning. The most recent version of Patsy detects when a student is having difficulties and can offer an appropriate selection of video clips to vicariously learn from.

Richard Cox is reader in computer science and artificial intelligence, Sussex University, and leader of the Vicarious Learning and Teaching of Clinical Reasoning Skills project.

Details: www.tlrp.org/proj/phase111/cox.htm

A screenshot of Patsy's vicarious learning interface is at www.vicarious.ac.uk/vl-patsy.jpg