Quiz and question banks

Here are my own thoughts. Sorry if anyone else has posted similar ones without me referring to them. You can also disagree with me if you like.

In the Undergraduate studies regulations of my school, one can read:

The evaluation of the student in a course is the responsibility of the teacher.

Well, it's good all the same that they mention it.

The evaluation is the assessment of the level of learning achieved by the student in relation to the objectives of the courses.

It is obvious that the grades serve this purpose. Assess students, give them a grade, and rank them from weakest to strongest. Sure, but there are other purposes of evaluations.

I would say there are three types of quizzes. The first aims to help students become familiar with and learn a subject, the second aims to help them master it and the third aims to encourage them to strengthen their understanding of the subject and to let them check for themselves that they have understood it correctly. This third type is often described as a "summative" assessment.

But suppose an important summative exam is canceled at the very last minute, say just as the students are taking their places in the exam room. The students will complain and some will even say that they studied "for nothing". Obviously, the assessment of the level of learning achieved by the students is not made. However, while this may sound awkward, much of the goal of the exam has been achieved. Indeed, the students prepared for the exam and thus strengthened their understanding of the subject. This is the most important. Mind you, I'm not suggesting that canceling exams at the last minute would be a good thing.

In my view, school is not a place where a competition is held to find a winner, which would be a degenerate view of it. School is a place where students learn and that should be its main goal.

One can also read in the regulations:

Not all students should receive the same grade.

The average for a class of the results attributed should not be too far from the probable or normal average for such a course.

These two points are interesting. 1) The grades must be reasonably well distributed and 2) their average must be as expected. The regulations further says that if this is not the case, the teacher must revise the grades at the request of the Dean of Studies, but in reality most probably of the Head of the Department. Note that nothing is said about the skewness of the distribution.

Cont'd in my next post.

A nice, slightly negatively skewed curve may be an ideal, but several factors make it not easy to achieve.

Let's take a given course as an example. Factors that influence the grades of the class include among others:

• The level of the students, which can vary from one class to another. It may happen that some groups are measurably stronger or weaker than others, but in general the composition of the student population is fairly constant session after session.
• The difficulty of the course. First-year courses in a program are arguably more challenging for students than final-year courses. Indeed, many things are new in the first year, and the rate of failure or dropping out is surely higher than in the last year, hence the averages lower.
• How the teacher assesses and grades students. This is probably by far the most important factor influencing the grades for the class. Some teachers are harsh, others are lenient in awarding grades. So much so that our School provides teachers with an in-house program, prepared by our statistics teachers, to work on final grades. Professors can do whatever they want with this tool, for example automatically adjust the average, its standard deviation or its kurtosis (leptokurtic or platykurtic) and, manually, its skewness. But the marks are always submitted to the Head of the department for approval .

As a teacher, you are free, to a certain extent, to assign grades as you see fit. One thing is certain, there will be a certain distribution of grades. Is there one that is natural? Maybe yes under some conditions, but the conditions are so many and varied that the answer is maybe that there are several natural curves or maybe none at all.

As far as I am concerned, I give marks whose average and distribution correspond to what my School's administration expects. The curves also have a negative skew which comes from the way I assess students in my course.

By the way, I worked in industry for several years before becoming a teacher. For various reasons, I had to change jobs often and therefore had to go through many job interviews. The only company where I was asked for my academic transcript was one where, as I quickly realized, I felt lucky not to get the job.

Some students don't give a hoot about the grades they get, as long as they pass the course. In a way, I envy these students. Maybe they will have a better life.