welsch@houxu.UUCP (Larry Welsch) (01/01/84)
To that eater of first lines.
This is in response to Robert S. Giansiracusa's article "Grades -- do they
help or hurt??"
General Philosophical Comments on Grades
In a very real sense what students pay for are grades that represent a
degree of attained knowledge and ability. To not give grades or to give
everyone automatic grades is a cop out ie. the grades are determined
before the class begins. To give grades based on a student's relative
performance to other students is a cop out, ie. to grade on the curve.
To not provide the student an opportunity to display his/hers (I wish we
had a good pronoun! I hate this his/hers */?$%) knowledge and ability is a
cop out, ie. to grade based on 1 dimension.
Each course is unique and before starting to prepare a course, there are
two questions that teachers should answer.
* What are the students expected to learn?
* How are the students going to be graded?
The two questions go hand in hand. If something is important enough to
learn then it should be graded and if something should be graded then it
must be learned.
Some general rules of thumb:
A. Asking students to grade themselves or their peers is a cop
out, ie. while teachers really cannot teach students, they are
obligated to grade them.
B. Students should not be surprised by their grades, should
understand how the grades are arrived at, and should feel they
are being treated fairly.
C. A teacher should have enough information on a student's
knowledge and ability that the teacher can present a rationale
for the grade given the student.
Grading the Large Intro. Undergraduate Lecture - Or How I Taught 215.
I have taught a number of large introductory undergraduate lecture
courses. By large, I mean at least 100 students. By introductory, I mean
1st, 2nd or 3rd course in a subject which every student majoring in
computer science is expected to be proficient in. (some network grammarian
will get me for that preposition). By lecture, I mean that I only give
lectures, but I also have a team of TAs for recitation sections.
For this type of a course, I invariably give the students a syllabus which
contains the following:
a. A description of what the course is about, and what the student
is to know at the completion of the course.
b. A description of how grades will be determined, (a composite of
10 highest quizzes out of 15, 2 exams, a final, and 5 assignments).
c. A lecture by lecture outline of the course, giving the topic
and selected readings for each lecture from the first lecture
to the last.
d. The dates for all exams in the course and the material each
exam will cover. (I give two hourlies and one final.)
e. All assignments and when they are due, minimum of 5 assignments.
f. Office hours and locations for myself and my TAs.
Half of the first lecture is describing the contents of the syllabus and
answering questions on it. The second half starts on the material of the
course. The first quiz is given by the end of the second week and the
first assignment is due by the end of the third week.
I find that those students who are not motivated leave. I have had as
many as a quarter of the students transfer from my lecture to a less
demanding one after the first lecture. CS was in enough demand that they
were quickly replaced.
Those students who do stay realize that I am serious and that their peers
are serious about the course. This provides for an intensity of learning.
Since I do not grade on a curve, there is no competition, so each student
can strive for as high a grade as {s|""}he can get. I find that once the
grading is described then I can concentrate on the material. Also
constant, preferably positive, feedback to the students is a real motivating
factor. Quizzes are designed to be easy, both to grade and to do well on.
Generally, after the first exam, I can tell whether or not a student will
pass the course and ask those students who are failing or doing poorly to
come in and see me.
Final grading is never easy. First, I compute the grade exactly as
specified in the syllabus. Everything has a point value so this is a matter
of summing the point values of 10 highest quizzes, exams, final,
assignments, and then checking a chart for the grades. This is done via a
program. (parenthetical comment while this may sound objective, remember
the assignment of the point values is subjective) Second, my TA's and I
then consider each student's record as an individual. The question is, are
the grades received truly representative of this student's ability. It is
at this point that we make adjustments for individual scores which are
anomalies or if a student just cannot take exams, but has shown
outstanding ability on programming assignments, or a student who did
poorly at the start of the course but has made constant improvement till
the end of the course. Third, we assign grades based on step two. Fourth,
I personally consider each grade with what I know of the student. The
fifth, and final phase is resolving any problems that the student may have
with the grade given. This I view as my personal responsibility.
This process is not easy and for a class of 100 students takes at least
eight hours. That is still less than 5 minutes per student per grade of
my time. What happens is that students who are border line get more
consideration than students who are not. Without a large number of
quantized measures to look at the task is impossible.
I repeat this is how I handle large introductory courses. Seminars I
generally handle based on assignments and presentations alone, no exams.
Medium sized classes depend highly on the subject matter. For example a
software engineering course might be graded based on a mixture of group
and individual work. The important aspect is to be clear how the grade
will be arrived at and to know this before the course starts. Feedback,
both to the teacher and student are important.
Some Comments on Robert S. Giansiracusa's Alternatives
The alternatives are numbered and indented.
1. Keep guaranteed minimum grades (based on exams so far),
allowing some upward mobility but no downward (this is
what I actually did).
I think this is what I described in detail. I disagree about downward
mobility since I have had students who did very well at the beginning of a
course and do progressively worse as the course proceeds. Usually they
are the ones who have some background in the material and think the course
is just a repeat of what they already know.
2. Keep current grades as TENTATIVE only, subject to upward as
well as downward mobility (this is the usual default).
I would like to understand better the circumstances under which a student
whose grades would be lowered. If, for example, the student didn't
participate and participation was a factor, then participation should be
graded. In short, I disagree with this.
3. Give all Cs.
4. Give all As.
Only if the students deserve the grades. If the grades are given
regardless of ability then this is a cop out.
5. Announce all As given, then give REAL grades based on effort and
initiative demonstrated AFTER that announcement.
This is playing games with the students. It is a really cop out and the
teacher loses credibility.
6. Base EVERYTHING on a single final exam.
Lazy teacher!!
7. Base EVERYTHING on a project.
If there are intermediate points of feedback along the way this is not
bad. Without constant feedback its just another Lazy teacher!
8. Purely subjective evaluation.
All evaluation is subjective. If this means let's take a guess at the
end of the year then this is the worst scheme I can think of. Of course,
students cannot really complain, cause they don't know how they were
measured. Grossly unfair and extremely lazy teacher! Guess who sleeps
with the teacher.
Larry Welsch
{allegra|ihnp4}!houxu!welsch
bobgian@psuvax.UUCP (Robert S. Giansiracusa) (01/02/84)
Just a quickie clarification to my original posting: > 2. Keep current grades as TENTATIVE only, subject to upward as > well as downward mobility (this is the usual default). > > I would like to understand better the circumstances under which a student > whose grades would be lowered. If, for example, the student didn't > participate and participation was a factor, then participation should be > graded. In short, I disagree with this. What I mean by "this is the default" is that most people grade this way: whatever average a student has at some point can go up or down depending on future tests or assignments, etc. I agree with everything else Larry says! -- Bob Giansiracusa (Dept of Computer Science, Penn State Univ, 814-865-9507) Arpa: bobgian%psuvax1.bitnet@Berkeley Bitnet: bobgian@PSUVAX1.BITNET CSnet: bobgian@penn-state.csnet UUCP: allegra!psuvax!bobgian USnail: 333 Whitmore Lab, Penn State Univ, University Park, PA 16802