Moed-A vs. Moed-B    -    a case study

What you may want to know about them

Here is an e-mail discussion, initiated by a student at the course Object-Oriented Programming (236703). The underlying subject is Is it reasonable to have a difference between the average grades of Moed-A and Moed-B ?
Since this discussion is relevant to any course (not necessarily academic course - as long as there is some normalization method for the grades), I have decided to put it on the Web, for the benefit of all of us: Students will have an answer before they ask, and lecturers will have less question to answer.
The only modifications I had made were line organization (mainly, in order to compensate over stupid mail-editors that break lines in an unreadable manner).
For obvious reasons, I have erased the student's identity.

• The student is asking a question:
``` Date: Wed, 29 Mar 2000 16:12:10 +0200 (IST)

Yechiel,Freddy, Hi,

While looking at the average grades of the two OOP exams Moed-A
and Moed-B, I noticed a huge gap between the two average grades:

Exam moed A 66.5
Exam moed B 55.7

A simple calculation leads to TEN points difference !
I don't have any idea why this is the gap and NO grade correction
was made, and why can't I get the same factored grade as
everybody else who was able to take the first exam ??

Why should students suffer from cold, unfair calculations?
Is it the overall average that leads to such a gap,
or is it a way to punish some of the students?

In my personal case, I simply couldn't take the first exam because
I was representing    XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX.
As a result, my grade is TEN POINTS lower then of those who took
the first exam.  As a student who worked hard to learn
and succeed on the exam, I feel deeply insulted by the results.
I think I recognize an unfairness when such is made.

==========================
(-) (a student)
```
• So I have answered:
``` Date: Wed, 29 Mar 2000 16:47:49 +0200 (IST)

What is the basis for your assumption that the population of students
attending Moed-A has the same properties (e.g, knowledge distribution)
as the population of students attending Moed-B ?

In fact, I see quite a few arguments to the contrary.

If I related to the rest, my response would be harsher.

Yechiel
```
• The student replies:
``` Date: Wed, 29 Mar 2000 17:09:59 +0200 (IST)

Yechiel,
The factor should be given according to the TEST DIFFICULTY
and not by the students distribution in the second exam.
If the tests where given in a FAIR (same) difficulty then
same factor should be given in both exams (or final grade),
but unfortunately this is not the case.

why there was a factor in the first exam final grades,
while in the second there wasn't ?
Is there a good reason for that?

There is NO need for harsh response,
we are dealing a principle matter ....
```
• If we were dealing about principles, I had to explain my view:
``` Date: Wed, 29 Mar 2000 20:53:54 +0200 (IST)

O.k., I did not want to respond about these matters,
but you're asking for it.

"suffer from cold, unfair calculations"
"a way to punish some of the students "

should be supported by something that is more
than your personal disappointment.
Without a support they are merely false accusations
and a personal attack on me, my character and my intentions.

For your other claim:
"my grade is TEN POINTS lower"

What do you know about the relative difficulty of the two exams ?
Among students who took both, one gained about 10pts
but another lost about 10pts.
What conclusion can you draw from this data ?

On Moed-A there were 74 students.
If there was a question that only 10 answered correctly
you may conclude that it was a difficult question
(but not necessarily unfair).
You may NOT conclude the difficulty of a question
if 13 students, out of 16, failed to answer it correctly
(because it may be the case that 10 of them just know very little
- see data below).

If ANYTHING can be said about Moed-B exam,
it is that it was VERY SUCCESSFUL in distinguishing
the level of knowledge of the students.
Here is the histogram, according to correct answers

There were total of 48 questions:

#correct-q	#stud
16		  1
18		  6
20		  1
21		  1
Note the GAP
27		  4
29		  1
32		  1
35		  1

It is evident that the average in Moed-B cannot be high,
and in fact - the average value is meaningless.

For the reference, in Moed-A there were 48 questions too,
and out of 74 students, 1 had correctly answered 40 of them
and 4 students had correctly answered 35 of them.
On the other end , there were 7 students who got 17-19 questions right.

you fall (whether intentionally or not) into the common, so convenient,
student fallacy, that a factor is per-course.
It is not.
The factor goal (at least in my opinion) is to try
to correct variations among exams (so it is per-exam).
Therefore, your request is erroneous BY-DEFINITION,
without even considering what sample we have,
and how representative it is (in our case - it was NOT representative).

You say:
"I think I recognize an unfairness when such is made"

Think again.  If there is unfairness here, it is YOURS,
drawing conclusions from vacuous assumptions.

You say:
"I feel deeply insulted by the results"

The only insult you should feel, in my view, is self inflicted,
by your irresponsible response.

To concluded, the only thing I am sorry about
is the time I had to spend - writing evident truth.
I have no intention to continue discussing this issue.

Unfortunately, this kind of discussion does not find its way to "Kuly",
so, once in a while I will have to repeat it again and again.

With all due respect
Yechiel
```

This page has been visited times since last reset.