The introduction of a final year undergraduate option module in the new area of object oriented methods provided an opportunity to adopt a student centred approach. The module lasts two semesters, is timetabled for two hours 'contact time' per week and is available to a group of associated degree courses across two faculties. The module provides around 14% of the final degree marks for the students concerned. The module assumes that the students have no prior knowledge of the specific subject area, although all have a foundation in 'traditional' computing methods.

In introducing this new subject onto the curriculum it was decided that student-centred learning would facilitate matching the syllabus to appropriate learning techniques and to our objective of producing an industrially relevant education for the students. Some of these techniques were novel and thus need to be evaluated in terms of:

• their educational value as effective learning techniques
• their value as reliable indicators of student ability
• their acceptability to the students
• their industrial relevance.

The main data to enable evaluation was from student performance in assessments and three detailed evaluation feedback questionnaires. In addition, other questionnaires were distributed (see Hand & Chandler, 1992). Ad hoc comments from students and other members of staff were also taken into account. The ongoing evaluation of the module's initial run in 1992/3 has been used to modify and improve the module for the second year.

The module entitled Object Oriented Methods is structured to allow a 'student-centred' approach, so that the students have opportunities to guide their own progress in object oriented methodologies. Therefore (within limits) they can choose how proficient they want to become in programming, how deep they wished to take the design exercises and which object oriented areas they wish to research. Additionally, lecturer-moderated peer assessment plays a major role within the module. There were large case-studies for design and programming and substantial research areas which formed the central elements of the module - with the presence of other parts (such as programming exercises) primarily to provide underpinning. As part of the module's 'real life' emphasis the students work extensively in groups. In view of this, the importance of "traditional" examinations has been minimised.

The assessment philosophy underlying the module design was that grading should primarily act both as an incentive for the students to do the work and as a way of enabling them to monitor their progress. It was only secondarily considered to be a means of providing the University with a means of classifying the student's final degree. The process of assessment has formed an integral part of the students' learning experience. Generally most of the assessment was coursework which were designed to test different skills inherent in object oriented methods and to enable students to practise their transferable skills.

On one version of the module (for the Business School) a short answer test was included. This was open book and purely the practical application of theory. On the second version of the module (for the Science Faculty) the external examiner required a 2 hour examination to be included.

Evaluations occurred throughout the module based on a pre-course questionnaire, a mid-course questionnaire and an end-course questionnaire. The number of respondents was consistently high (usually in excess of 75%). There are also indications of the students' understanding in the coursework and the test/examination answers. These results are presented elsewhere (Chandler & Hand, 1993; Hand & Chandler, 1994)

The pre-course questionnaire indicated that the module was being taken by a mixed group of students from differing backgrounds who indicated a variety of positive and negative reasons for doing the module. The students had a range of experience in object oriented methods with 56% claiming to know nothing about object oriented methods. 44% thought that student-centred nature of module was O.K. (despite having little knowledge of the methods that were to be used) and 34% thought it was good news. 62% thought that the range of assessment methods was good news.

Object Oriented Theory

The module's limited time and its concentration on the practical usage of object oriented techniques inevitably meant that the theoretical basis of the object oriented paradigm was under-emphasised, with only three hours specifically allocated to theoretical underpinning at the start of the module.

The students received a basic introduction to the major concepts. Most students felt that they had at least an adequate or better introduction to the principal object oriented concepts. This lack of theoretical basis is probably the major criticism that could be directed at the module from an academic - although not a vocational -perspective.

A major element of the module was concerned with new developments within object oriented methodologies. The students organised themselves into groups and were provided with open learning study packs consisting of recent references on a particular topic. Using these, together with their own research, they made a verbal presentation to the rest of the class and produced a two page summary document. The verbal presentations were marked by a combination of peer assessment and lecturer assessment. The students were provided with guidance stressing the importance of content (compared with presentation skills) on how to assess the verbal presentations. The students in each group were given an opportunity to (anonymously) redistribute the marks allocated to the group according to the work put in by each member. The written summaries were assessed solely by the lecturers.

The lecturers running the module felt that the students performed well in both the verbal presentations and in the written summaries. The students themselves also felt that they had gained from the experience of preparing the material. As well as the students being assessed on the quality of their own presentations (verbal and written), all students were also assessed on this material as part of the short answer test questions (to ensure their support for the presentations and that they read the summaries). The students were generally unhappy about the main source of some material being that produced by other students; nevertheless, the students performed well in the short answer tests and the type of question was popular.

Object Oriented Design

The object oriented design part of the module was centred around a single class-wide case study, with simple case studies used for illustration. The students worked in self-selected 'design groups' consisting of four or five students. In addition, the lecturers instigated a 'cross-over' group system, such that each member of a design group was placed in a 'discussion group' consisting of representatives from four or five different design groups. The way the discussion groups were formed ensured that each student continually viewed the development of at least three other groups' designs and had indirect feedback from all the other designs being produced in parallel.

Each hour of contact time commenced with discussion group meetings to review their progress on the case study during the previous week and to exchange ideas. This was followed by a short formal introduction to the next stage of the methodology and finally an opportunity for the students to start work in their design groups on this next stage of the design.

The design was assessed by peer review (from the discussion groups) with moderation by the lecturers. Students were provided with guidelines to assist them in their assessment technique. The students were required to rank the other designs in their discussion groups. The (anonymously-produced) rankings were collated and formed the basis of the final marks for each design. It was felt that peer assessment of this type was an essential element to ensure that students actively contributed to the discussion groups throughout the designs' development.

The assessment guidelines for the students asked them to judge their colleagues designs in terms of breadth, length, depth and presentation. In order to evaluate and to calibrate this assessment method, the lecturers also marked the designs. The student rankings of the designs was close to that of the lecturers: there was good agreement on the best few and worst few designs. Therefore, the final decision was that the lecturers felt confident in converting the students ranks into percentage marks. Over 50% of the students did not like the peer assessment and/or considered it unfair, despite it resulting in similar marks to the lecturers.

The design was covered by a large continuous case study approach using design groups and discussion groups. Most students felt that the case study was too large and/or too difficult. The module lecturers were in agreement on this issue. Most of the problems experienced on the module as a whole derived from the excessive size of the design case study. This impacted on the other modules the students were doing and thus caused concern to other members of staff.

However, the approach was generally well received; 74% felt that they had been introduced to object oriented design adequately or better and 72% liked undertaking the module work as a group exercise, with 49% thought that the discussion groups were a good idea. The method seemed to inculcate a high degree of understanding amongst the student group. The technique of discussion groups appeared to the lecturers to be successful, but it seemed clear that more guidance was necessary to make the group discussions more effective. Additionally the use of a cumulative case study allowed the students to spread their coursework throughout the term, get regular feedback about their progress and experiment with and change their designs in the light of their own and others' experiences.

Object Oriented Programming

To introduce object oriented programming (using Smalltalk/VDOS) the students were encouraged to work through the Smalltalk on-line/manual tutorials, a set of internally-produced 'Smalltalk Quick Reference Sheets' and a Logbook containing graded exercises. The logbook approach to assessment and learning is being used increasingly within our Department in a variety of ways.

Twelve essential competencies were identified for Smalltalk programming. Each competence was then broken into the skills necessary to achieve 'Novice', 'Competent' and 'Expert' levels. Exercises were designed to cover the whole range of skills and a table was drawn up showing which exercises needed to be successfully completed in order to gain recognition of a particular competence level. The students then self-assessed their skills with their fellow students confirming their ability at each level. The lecturers provided moderation, checking all the students' claimed achievements by means of a small selection of test cases. The logbook approach was popular (only 10% disliked it) and they liked the peer assessment element of the logbooks (only 13% felt it was unfair).

Advanced Smalltalk programming was taught and assessed by means of a programming case study, which was supported by the lecturers in machine-based 'trouble-shooting' tutorials. The students organised themselves into small groups to carry out a significant programming assignment using Smalltalk/VDOS. The original intention was to continue with the design case study since seeing the full development life cycle through was considered to be important. However, the design exercise proved to be too large in the time available and so a smaller programming case study was substituted. There was no peer assessment element for this coursework. Lecturer assessment was based on the working programs and took into account both the coding and the interface. Students reported that the case study provided a good but difficult learning experience. However, by the end of the module, most students felt confident in applying Smalltalk to a new application.

Overall feedback from first year

The overall feelings about the teaching methods indicated that 81% thought that the non-traditional teaching methods used on the module were O.K. or better. 36% thought it was better or much better than other modules. However, 57% thought the module was much more work than other options. On being asked how highly they would recommend the module, 26% said highly or very highly if it remained as it was but this increased to 60% if it had small changes (such as a reduced design case study). Thus, apart from peer assessment which was disliked, the students generally indicated a high level of satisfaction and achievement on the module as a whole. The general knowledge of object oriented techniques indicated by the test results was good and reflected the large amount of work that most students put into this module.

The general conclusion to be drawn from the first year of experience of implementing this module is that student-centred learning techniques are an ideal vehicle for teaching object oriented methods, provided appropriate care is taken over their implementation and the staff are responsive to the needs of the students. The use of new teaching and assessment techniques in the object oriented methods modules had proven controversial with students and other members of staff in the University. Nevertheless the lecturers involved feel that the module as a whole has been very successful (as indicated by several measures, including the high proportion of students who now recommend the module), and that the module would continue into the second year, with relatively minor amendments.

The sections below describe the major modifications made to the course for its second year of running, together with some initial responses. The background of the students was similar to the previous year.

Object Oriented Theory

In the light of the previous year's experiences major modifications were made to the introductory theory lectures. These were increased from three to seven hours, including a practical session in which the students were asked to provide examples of object oriented concepts in a given situation. The lecturers felt this greater emphasis on theory gave more substance to the academic side of the subject.

The second element of theory - recent developments in object oriented technology - was adapted to allow the students to present the research material in the form of posters or debates instead of in lecture-style presentations. This was very successful and the quality of many of the posters very high. The quality of the written summaries also seem to have improved this year, perhaps because the proscriptions imposed in creating posters allowed the students to understand the principles of creating effective summaries.

Object Oriented Design

The size of the design exercise was decreased substantially. In addition, the lecture component was reduced to provide more time for the discussion groups to view the designs. The lecturers also played a bigger role in guiding discussion, but this was still limited due to the large group size. Finally the assessment guidelines added a fifth 'dimension' - that of design quality (in terms of whether good software engineering practice has been adopted). The peer assessment was less successful for this group, although there was an underlying agreement between student ranks and staff ranking (especially on the best and worst). The final decision was to combine the student and staff marks in a 50:50 split. There is no doubt that the case study itself was more popular than the previous year: 78% liked the exercise being in groups and 72% thought it being cumulative was good. However, they still generally considered it too big and too difficult. They still disliked peer assessment (57%) and felt they did not gain from it (79%). The quality of the designs was better than the previous year; the top three designs certainly being implementable. This is no doubt because the case study was considerably more manageable enabling the students to concentrate on the quality of the design.

Object Oriented Programming

The principal changes on the introductory programming were enforced by the change to Smalltalk/VWindows. The lecturers themselves did not have opportunity to try it out prior to the course, but were led to believe that it would not affect the content of the exercises they used in the Logbooks. This did not prove to be correct. For this reason several examples needed to be changed during the course. However, overall the Logbooks seemed successful and gave a good spread of marks. The students were rather less supportive than the previous year with 40% disliking them. However, they claimed to find Smalltalk easy. They did not find the exercises too difficult (despite the problems mentioned above) and generally liked them. They liked setting their level of attainment (88%) and thought the method of assessment was fair.

For the advanced programming the students were set a coursework based on the object oriented design case study they had previously undertaken. The students were not constrained to use their own design, with the result that a number of students with weak designs were able to utilise designs produced by other groups.

The lecturers teaching the course feel that they have learnt a great deal from teaching this course in a more student-centred way. They have become more aware of their own pedagogical methods and views, especially in considering the role of assessment in teaching. They have - almost inadvertently - put assessment in the role of a fulcrum to 'force' the students to set their own agenda and to become active learners.

In addition, the course has emphasised transferable skills a great deal. As well as knowledge about object oriented methods, the students have been tested on a range of other abilities including technical presentation skills, critiquing and appraisal skills, ability to synthesise new ideas, research skills, summarising skills, group working skills and so on. These skills are ones that potential employers highly value and thus provide increased relevance to the module.

There are clear cultural variations between students from the different departments which affected their attitude towards the module and to their colleagues. Given the trend towards 'mixed background' teaching it is interesting that we have detected significant differences between students in 'similar' subjects such as Statistics & Computing compared with Computer Science. This suggests that classes mixing more diverse elements may have considerable problems. Careful consideration of future courses is needed to ensure that additional time and resources are available to overcome these 'cultural' differences.

Thanks go to the EHE who have supported this work, to Dr Angela Brew for her encouragement throughout and to Kate Cook and Larry Dunn for their invaluable aid in analysing the data so effectively and especially to all the students who have participated in this course.