The Bell Curve

Question

Some of the points you raised actually point to the administrative hurdles which are experienced by teachers. One difficulty is the bell curve. That is okay with essay type questions. You can give marks of 80% or 60% depending on what you expect. But if you have a large number of problem-solving questions, there is no way of predicting what the distribution of marks will be. What you need is a kind of politically flexible relation between marks and grades. In the absence of that, you will find the curve shifted a little to the left or right and then you have to do all kinds of tricks to adjust the marks to reflect the bell curve. Actually, I do not understand why there should be a fixed marks-grade relationship. That can be left to the teacher.

Answer

The problem lies with the fact that this is not Cambridge University, U.C. Berkeley or Stanford, where every Asingle teacher can be safely empowered and the student abilities are also more homogeneous. Here, past experience has shown that the moment we relax, there are misuses and problems. Let me give you some indirect evidence to show that people need time to adjust to the flexibility of empowerment. When I was a head of department, I removed the need to record xeroxing. One month later, xeroxing cost went up by 300%. Some staff were actually bringing their children’s music books to copy. The problem is that the moment you give that kind of freedom some staff will suffer. While staff who are fair will spend 50% more of their time grading, others will spend more time on consulting. I think our spectrum of quality of teachers is not as uniformed. But I am just painting you the problem. It can be done, but with proper management. The first thing is to train the heads of department on how to deal with difficult staff. All these are systemic problems that we have to address.

Rigid Exam Rules

Question

In many departments, there are very rigid rules about the number of questions and the type of questions and so on. For example, there is a rule that you have to have five questions and only one of them can be obligatory. I do not see the rationale for such rules particularly when each department is different.

Answer

I think we have to provide training for heads of department to be able to accept that kind of situation. You see, Aif the head of department just follows the convention, he is there just to safe-guard the conventional approach. If we are going to train creative people, some of these will become creative lecturers and some will become creative heads. A creative head will realise that there is a group of very good teachers who should be given more freedom; the weaker teachers should be guided by the system. So we should introduce a flexible system where if a teacher is voted best teacher, or there are a few top researchers, you would want to empower them with flexibility. Then these few teachers will be the ones who will spearhead innovation.

They could try out their own types of questions and format. After a year or two, they can then share their experience with others and there will be team-learning. The better, more creative staff who are top researchers and able to integrate teaching with research should be encouraged to go ahead. However, anyone who wants to deviate should be prepared to take a risk. If he fails badly, he should apologise and tell the head of department his new thinking and plans to adjust. Staff should be allowed the chance to fail without penalty so long as they are responsible. Otherwise, who would dare to stick their neck out? Again, the head has to be trained on this.

Change

Question

Change here has been very slow. Even if heads were willing to go through training, it would take a great effort to change. And also, when you say we should stand up and say things, change takes place only when implemented from the top. A few years ago, I ended up having a fight with an external examiner and the head of department when I tried to make changes to the exam. It is not easy to change because the people are still there—the same people who have been there a long time.

Answer

I am sending a paper on Learning Organisations to the deans, asking them to share it with their vice-deans and Aheads and expecting the more dynamic deans and heads to share that paper with the staff. In that paper, I raised the subject of upward appraisal which I started in the EE Department in 1989. I allowed staff to carry out confidential evaluation of their division heads. All these would come to me, and I would then provide feedback to the division heads on what their colleagues thought about them. I went a step further to allow all the forty-five staff to evaluate me as a head of department. It was confidential, and the assessment went to the dean. The dean would summarise the rating and discuss it with me. This is upward appraisal. Till today, not all engineering departments have adopted it. So, it is not easy to change. Without training, incentives and clear guidance, the heads cannot do it. But we want to share with the heads and the deans that those who are daring and ready can go ahead.

Types of Assessment

Question

I was looking at the types of assessment in different universities. There is one assessment system used by the University of Queensland. They call it criterion-referenced assessment system as opposed to what we are using—the norm-referenced system. In the norm-referenced system, we measure students’ achievement with respect to their peers’. As a result, if we have very good students with two A’s and two B’s going through our system, we will eventually drop some of these first class into second class and some of the third class to pass degree. But in the criterion-based system, we first set our criteria and conditions for grades A, B, C, and students will not have to pitch against themselves but will try to measure up to our yardstick. Also, in this criterion-based system, we have to think very deeply about assessment because we have to do a lot of planning before we assess.

But of course, this criterion-based system has drawbacks as well because, it is more difficult to sort out first class, second class and so on. So a lot of thought is needed if we adopt this system for NUS. Secondly, in this system, the setting of conditions for the different grades can be abused. Finally, adopting this system will be difficult for our staff at present because they are so used to the norm-referenced system. But I think this system addresses a lot of our concerns.

Answer

Agreed. But this is a major issue from a management perspective. It means the head of department would Aknow where he wants to go. A very strong department would set a target of three or four years, allow half the staff to try it and then they will become role models for the rest. So in five years, the department as a whole would reach the goal. A weak department would allow five or ten percent of its best staff to try it out and then evolve slowly. It may take ten or more years. But it will not be acceptable if nothing much is done after ten years. Again, it boils down to the training of heads and deans.

Norm-Reference vs Criterion-Reference

Question

The issue about norm-referencing and criterion-referencing is more urgent because the crush is already being felt. For example, in CA’s we set the learning goals and we set the criteria—what would get an A, what would get a B and then according to the criteria, students work very hard. Many of them actually achieve an A. As a result, most of the students get A’s and B’s. But when it comes to the final exam, we are told to satisfy the bell curve. In that sense, those students deserving A have to be marked down and I find that rather difficult to take. The bell curve may be true in statistics, but it can be done as a description after the event. Yet we say that there should always be a bell curve, therefore we should fit whatever population we have got into that bell curve. In this case, the already deserving A and B students are not properly assessed.

We have to think what sort of information we get from this norm-referenced system. In norm-referencing, we can only tell that out of a batch of students, who are the top thirty, who are the second group from the top. But it does not tell us what the students can do. In criterion-referencing, we can tell what a student has achieved; we know exactly what he can do, or what he cannot do. I think top international companies would be more interested in what the student can do rather than what percentage of NUS he has achieved. Criterion-referencing is the trend in the world today; sticking to the bell curve in norm-referencing seems backward.

Answer

Let me tell you that there are some structural constraints which I find very hard to remove. I am starting to Araise this question: In the knowledge economy and the age of life-long learning in the 21st century, the concept of a first class honours may not hold water anymore. So you may find that one day when the British abandons the honours system, NUS, because of tradition, might hang on. The issue about the old titles “lecturer” and “senior lecturer” is the same. The world had shifted to the U.S. style and we were still hanging on to the old system. Our very senior person was only called “Associate Professor” and I was queried by my Japanese counterpart for sending such a ‘junior’ person. It took us a few years of discussion before we switched to the U.S. titles recently. So these are the things we are willing to change but I am afraid that there is still hesitation regarding the honours system. If you really think about it, the Grade Point Average (GPA) scoring is good enough.

Many people who rise up to become bosses or leaders were second upper or second lower students. Our first class honours may not be the boss or leader. So all these are only academic achievements. I would suggest a halfway solution in the first instance. That is, instead of a bell curve, we will have a modified bell. That means because our students in some departments are better, relative to other universities, our middle section should be larger. Then the first class honours should be those who are very outstanding. So now, what is outstanding? Scoring an A is not automatically outstanding; to have produced something publishable or something unique would be. For example, a student who takes the initiative to tell me what he is interested in, and that he has formulated a problem and wants to solve it is outstanding. Actually, the Undergraduate Research Opportunity Programme (UROP) gives you a lot of opportunity to single out the outstanding students who will deserve first class honours. I am actually challenging the heads and deans to see how many on their list of first class honours have had the UROP experience.