Regardless
of whether we have marks or grades, students still have to sit for
the exams, The exam papers will still be corrected and marks given to
them. Are we not just artificially converting marks into grades,
e.g. above 90 is an A, below 40 is an F? Are
we not being unfair to a student who earned more marks but there is no
indication of it in grades?
Let
us start by acknowledging that grading scheme is not infallible.
However,
even a small subtle change in the objectives or philosophy can have a
dramatic impact overall. It changes the way we look at problems and
the type of solutions we seek to overcome these problems.
Grading,
relative grading to be precise, is philosophically different from
marks. It assumes that each group of students is similar and the
relative performance of the group will be statistically similar. As a
Professor of mine said that about 15% of you will get an A – even
if the class is full of Newtons. He hastened to add, “What is the
probability that all 25 of you are Isaac Newtons?”. The
distribution of grades is not rigid and precise. Rather it is a
guideline. In certain circumstances, an A grade may be given to 20%
students instead of 15%. But giving an “A” to one third of the
class may be regarded as unreasonable.
The
assignment of grades to marks is not very difficult and we can follow
processes which are transparent and self-correcting. Before getting
into the details of a grading scheme, it is worth demonstrating that
this approach minimises some of the serious problems which we have
with our existing system. We will illustrate the issues with specific
examples drawn from my experiences in industry and academics. We will
also switch roles as an examiner or as a member of a recruitment
panel, hopefully, without causing too much confusion.
Orals
and Practicals :
Whenever
I have been a part of the oral exams, we have found it very easy to
rank students into categories. We found it easy to assign grades from
A+ to the best group, A the next group and so on. University wants
marks not grades; so, these groupings needed to be converted into
marks. The conversions have invariably been very generous because
there is nothing concrete on which to base our judgement and
generosity is the path of least resistance. In fact in one instance,
the conversion spread recommended was 95% for A+ and 80% for C!
Exactly
the same problem is faced when dealing with project work. While
recruiting, we came across a student with first class in the final
year, who, on close inspection was barely passing in theory. The
reason was that his project marks were 95%! We decided after that to
eliminate the project marks before calling a student to appear for
our test. Very good students suffer. As one student mentioned that in
retrospect, it was silly of them to put so much effort into their
project because all the effort was not worth much because everyone in
their class got very-very good marks in their projects.
Exams:
We
often hear of a hard question paper or that questions were out of the
syllabus. Do we ever hear that the question paper was unfair because
it was too easy? We have seen results where nearly everyone got over
60% with half of them getting distinction. Did the students suddenly
show brilliance? There is nothing wrong with such question papers,
except that they distort results. Without having a yardstick with
which to compare e.g. the distribution of marks in a given year at a
given university, judging candidates' performance based on marks
becomes very difficult. 'Grade-flation' can be a serious issue with
grading systems though the guidelines about the distribution of
grades is used to monitor the teachers to ensure that high grades are
not devalued by the 'generosity' of teachers. In 1960's, some of the
US university administrations faced this dilemma of high grades
because 'all our students are very good'. The reaction from the
industry was swift and approximately that “since you can't tell us
who are your best students, we can't either. So, we won't recruit
any”. Guidelines were automatically followed – no enforcement was
needed.
Part II
Examinations
are a very serious business. Almost every year, many universities
must be facing problems with some question paper, which contained
erroneous questions, questions that were out of syllabus or the
paper was too long. Soultions applied are arbitrary and if a
newspaper report was correct, bizzare. For a B,Ed program, the Maths
exam paper was annulled and the students were given marks in Maths
based on the average marks in the other subjects, none related to
Maths! What shocked me was that even the newspaper reporter felt it
was fair and students should not have to go through the agony of
appearing in an exam again. I do not recall why the paper was
cancelled. Grading will not ensure that a paper is not cancelled. If
a paper has leaked, there is little else that can be done.
Grading
allows many problems with examinations to be resolved swiftly and
internally without public glare or ire. This allows the university to
take tough decisions on the rare occasions when there are serious
problems with the exam system. I am convinced that the students would
be fully supportive of the university in those instances beacuse the
perception will not be that “University is always blundering”.
Consider
the scenario of a tough exam paper. With grading, a hard paper will
not cause any problem because we can give an “A” to a person
scoring only 50 marks should the need arise. If a question is out of
the syllabus, that can be ignored and the paper corrected and grades
given on the balance paper. There is rarely any need for the issue to
get into public debate or the university being faced with pressure
about the problem with the paper. “Although that question was worth
only 6 marks, we wasted so much time that we could not finish the
paper and we must be given 15 marks.” Negotiations take place,
bargaining goes on and a compromise value is agreed upon. The problem
is solved to the satisfaction of no one and everyone is aware that
the same scenario will be faced the coming semester with some subject
or the other.
Net
effect:
A
friend of the family was waiting for a perfect wife. Never found
her. In life we make choices based upon the options available. We
will look at the candidates' performance and the performance of the
University/College where studied. As employers, we based on
our experience and the performance of our employees from an
institution, develop a mental model of the quality of an institution
and its students. It becomes imperative for the students studying at
an institution that employers trust their institution.
Based
on the candidates' results and our knowledge of the institutions, we
rank the candidates in different categories. During my entire
industry experience, a few percent marks difference was never
considered seriously. We did, however, look at consistency.
Suppose
a candidate has very high marks in one subject. What signal do we
get? We will assume that the candidate knows this subject well. We
will start our interview questions with the strength of the
candidate. Now, suppose that the question paper that year had been
very easy and the candidate is not able to answer our questions
effectively. Human nature is such that we talk more about the
negative experiences than the positive ones. The interviewer is
likely to spread the word. “Last week, I interviewed a candidate
from Goa University. He had 75% marks in object design but could not
even explain ...” The credibility of the University will suffer.
Had it been grading system, the candidate may have got a “B-”
inspite of scoring 75% marks.
Employers are searching
for answers to the following two questions:
How well did a
student learn a subject?
How does his
performance compare with that of his peers?
Both the questions can be
answered with greater consistency using grading. Grading is by no
means perfect – just better.