different viewpoints
Reinhold Friedrich Burger
A recent exchange on a cs course newsgroup. Some editing was done
for anonymity, but the last line, in quotes, is the student's
complete and literal response. And no, I'm not the TA.
-------------------------------------------------------------------------
Student (posting from school account):
If anyone knows how to do questions 2 & 3 of the assignment,
please email me at <student's yahoo account>.
TA (to students in general):
Careful! Please read the assignment guidelines before you answer.
(The guidelines deal with the usual things, in particular excessive
collaboration.)
Student's indignant response:
"I'm not asking for the answers, just how to do it."
-------------------------------------------------------------------------
Simon Law
I suspect that the student has difficulty with the English
language and just isn't being precise enough. It is possible to infer
that the student wants to be pointed in the correct direction and given
a swift, er... push.
Simon
Jason Grove
Hi
The Third Degree (grad student newspaper) welcomes submissions of amusing
quotes from students/ TAs/ professors - we would like to start a regular
column. All submissions will remain anonymous!
email them to me at mailto:edi...@thirddegree.org
cheers!
Jason
"Reinhold Friedrich Burger" <rfbu...@scg.math.uwaterloo.ca> wrote in
message
news:Pine.SOL.4.44.030312...@scg.math.uwaterloo.ca...
David Evans
That's certainly a reasonable interpretation. Someone better versed in
English slang might have said, "Dude, I'm just looking to compare equations
for delta. Mine is completely boned!"
--
David Evans (NeXTMail/MIME OK) dfe...@bbcr.uwaterloo.ca
Ph.D. Candidate, Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/
University of Waterloo "Default is the value selected by the composer
Ontario, Canada overridden by your command." - Roland TR-707 Manual
Graham AW Duke
At least (s)he is honest about it. I don't know how it is nowadays, but
when I was doing my degree (not long ago at all) I'd guess that at most
20 people in a class of 120 would actually DO their assignments...the
rest would collaberate more than a reasonable amount. Of course I don't
see this as being a problem when you need to pass exams to pass a course
and when the standards are reasonably high. Of course now that passing
grades are becomming much easier to achieve, I could see how this might
constitute a real problem.
Chris Redmond
Speaking as someone who never passed a programming course, I
don't exactly know how it's supposed to work . . . but isn't
it usual for the instructor or the textbook to explain "how
to do it", and then for the student to go and do it? If
after reading the textbook and hearing the instructor, the
student still doesn't know "how to do it", what is the approved
course of action, other than asking somebody else to teach
what's missing?
CAR
Julie Lavoie
> > Student's indignant response:
> > "I'm not asking for the answers, just how to do it."
> > -------------------------------------------------------------------------
>
> I suspect that the student has difficulty with the English
> language and just isn't being precise enough. It is possible to infer
> that the student wants to be pointed in the correct direction and given
> a swift, er... push.
That's seems rather likely. OTOH, I've been somewhat stunned this term by
the amount of cheating that seems to go on. I had two classmates ask me
if I could just send them a copy of my Unix assignments for cs241, with
one who asked outright for all 5 of them. Another duo reported that
they finished a difficult assignment due the next day by calling one of
their fathers, who has a Masters in math, and having him go over every
single question of the assignment.
Has it always been this bad?
-Julie
Anton
in fall'97 when I was taking cs134. a student posted to
the cs134 newsgroup (from a hotmail account supposedly)
with the following comment: (my paraphrase from memory)
"i have the code to assignment 5, if you would like to acquire
it please email me, we'll agree on a deal"
no idea what happened next... (that assignment was probably the hardest
cs134 coding assignment we've had all term.. ahh memories)
Anton -- CompSci / SoftEng @ UWaterloo
------------------------------------------
Everybody seems to think I'm lazy;
I don't mind, I think they're crazy.
Running everywhere with such a speed;
Till they find, that there's no need.
Simon Law
Sometimes I give out portions of my code as samples. However,
my coding style is rather, uhm, unique; and I always CC the prof. at the
exact same time, and warn the students that the prof. has been CCed.
For some odd reason, they never[1] actually read the sample code
but implement some broken crufty solution instead.
> Has it always been this bad?
I would suspect so.
Simon
[1] Never say never, actually. I remember a high school final where we
had an open book exam. Some guy had a binder full of course notes
which happened to have my solution to an implementation of Simpson's
rule integration. Boy was the prof. angry at him...
Doug Payne
> That's seems rather likely. OTOH, I've been somewhat stunned this term by
> the amount of cheating that seems to go on.[...]
>
> Has it always been this bad?
I'm old enough to remember when student ID cards didn't have pictures on them.
Several senior students were busted in at least one exam I was writing when
they were recognized as not belonging. They had quite a profitable business
going, writing final exams for other students for money. The next year,
pictures were required on student ID cards.
Michael Froh
was actually that it was (optionally) a group project. That particular
student had done the assignment entirely on their own, and was asking if
anyone wanted to "be their partner", perhaps with some sort of
remuneration involved. There later ensued some debate on the morality and
academic integrity of such an arrangement (suggesting that in any division
of labour, someone will be carrying more of the burden, so it's just a
question of where one draws the line. If the "new partner" were
responsible for stapling the assignment and bringing it to the drop-box,
who is to say that they didn't do any work?)
At the time, I just shook my head, logged off my dialup connection, and
went back into TurboPascal, to single-handedly do the coding for my group.
Hmm... I wonder how many people would believe me if I tried to convince
them that my work on that assignment was the cause of my partner later
going on to win the Alumni Gold Medal at graduation.
Regards,
Michael
Stephen Forrest
> At the time, I just shook my head, logged off my dialup connection, and
> went back into TurboPascal, to single-handedly do the coding for my group.
> Hmm... I wonder how many people would believe me if I tried to convince
> them that my work on that assignment was the cause of my partner later
> going on to win the Alumni Gold Medal at graduation.
In what sense? That all of his assignments and work in university
together were the "cause" of result, or that your taking on this burden
saved him from the clutches of CS?
Steve
David Evans
Chris Redmond <cred...@watserv1.uwaterloo.ca> wrote:
>>> > "I'm not asking for the answers, just how to do it."
>
>Speaking as someone who never passed a programming course, I
>don't exactly know how it's supposed to work . . . but isn't
>it usual for the instructor or the textbook to explain "how
>to do it", and then for the student to go and do it?
In introductory courses assignments may have some of this flavour. However,
that aspect becomes less and less dominant, as students are expected to have
a better grasp of the skill of turning the solution to a problem into code
that the computer can execute. The main point of many upper-year assignments
is for students to figure out "how to do it."
>If
>after reading the textbook and hearing the instructor, the
>student still doesn't know "how to do it", what is the approved
>course of action, other than asking somebody else to teach
>what's missing?
>
Going to the TA to ask for a hint is fairly common. In such occasions I
would ask the student fairly specific questions where the answers would lead
towards an understanding of "how to do it."
Graham AW Duke
>
> Hmm... I seem to remember that as well. The interesting point in that case
> was actually that it was (optionally) a group project. That particular
> student had done the assignment entirely on their own, and was asking if
> anyone wanted to "be their partner", perhaps with some sort of
> remuneration involved. There later ensued some debate on the morality and
> academic integrity of such an arrangement (suggesting that in any division
> of labour, someone will be carrying more of the burden, so it's just a
> question of where one draws the line. If the "new partner" were
> responsible for stapling the assignment and bringing it to the drop-box,
> who is to say that they didn't do any work?)
Indeed, but when a CS partner becomes a commodity it is not dissimilar
from when a sexual partner becomes a commodity. The results may appear
identical on the surface, but are in fact completely hollow.
Prabhakar Ragde
> Speaking as someone who never passed a programming course, I
> don't exactly know how it's supposed to work . . . but isn't
> it usual for the instructor or the textbook to explain "how
> to do it", and then for the student to go and do it? If
> after reading the textbook and hearing the instructor, the
> student still doesn't know "how to do it", what is the approved
> course of action, other than asking somebody else to teach
> what's missing?
Programming is not quite as creative as writing a short story or
painting, but it is not just a matter of applying a simple technique
in a straightforward fashion. There are decisions to be made,
low-level details to be fleshed out, and your first try is unlikely to
work as you expect it to. It is more or less true that a programming
assignment can be done given sufficient time. Some students do not
wish to put in that time, and other students are weak enough at it
or do not budget their time, and so they try to save time by copying
someone else's work. While this would be obvious plagiarism in the
case of a short story or painting, it is not so obvious in
programming, because correct submissions will resemble each other to a
considerable extent.
After the first course or two, programming is more or less taken for
granted, and students are expected to be able to take a software
concept and implement it. Some assignments may not involve programming
at all; they are done on paper and involve design and analysis. When
mathematical techniques are involved, or problem-solving skills are
exercised, it is quite possible that some students will not know how
to do a question after some work. It is not clear, at this point,
whether they just need to put in more work, or whether they are
missing some knowledge or understanding necessary to complete the
question.
It is my impression that copying assignments is more or less rampant
in computer science. We face the dilemma that some skills cannot be
learned without considerable practice. If we do not give credit for
that practice, students will not do it, and then they do poorly on the
final exam, which they typically must pass in order to pass the
course. But giving any credit for the practice encourages students to
earn that credit by cheating, which is a lot easier on an assignment
than on a final exam. It also encourages reasoning of the sort, "This
is worth only 5%, but it's taking me hours, it shouldn't take this
long, I've put in my time and earned that 5%, I'm just going to copy
the rest."
We also have the problem that if a large number of students fail the
final exam, it is difficult to make the case that they should not get
the credit because they lack the skills needed. Even if the exam is
fair enough that we can reach that conclusion, it is impossible to
tell whether the instructor has been ineffective, or whether the
students simply did not put in the effort to learn. The quality of
instruction in computer science, the stability of the curriculum, and
the level of preparation from high school are all quite variable. So
we adjust the marks, fail the worst cases, and send a fair number of
poorly-prepared students on to the next course, where the cycle
repeats.
Since the primary measurement of instructor effectiveness, apart from
the failure rate, is the statistics from the anonymous student
evaluations taken in about the second-last week of classes, any
attempt by an individual instructor to alter this dynamic is doomed to
failure. Nor do I think there is much institutional will to tackle the
problem.
Is the problem worse than before? There's always been copying. But we
are passing through (hopefully through) a cultural phase where the
dominant metaphors on television and radio are the remake, the
pastiche, the collage, the sample, the remix. Do a Google search for
anything and see how many hits you get with chunks of identical prose
in them. It is understandable, though perhaps not excusable, that
students may undervalue individual effort. --PR
--
Prabhakar Ragde plr...@uwaterloo.ca
Professor, School of Computer Science DC 1314, (519)888-4567,x4660
Faculty of Mathematics Waterloo, Ontario CANADA N2L 3G1
University of Waterloo http://plg.uwaterloo.ca/~plragde
Joshua Chud
> It is my impression that copying assignments is more or less rampant
> in computer science. We face the dilemma that some skills cannot be
> learned without considerable practice. If we do not give credit for
> that practice, students will not do it, and then they do poorly on the
> final exam, which they typically must pass in order to pass the
> course. But giving any credit for the practice encourages students to
> earn that credit by cheating, which is a lot easier on an assignment
> than on a final exam. It also encourages reasoning of the sort, "This
> is worth only 5%, but it's taking me hours, it shouldn't take this
> long, I've put in my time and earned that 5%, I'm just going to copy
> the rest."
>
> We also have the problem that if a large number of students fail the
> final exam, it is difficult to make the case that they should not get
> the credit because they lack the skills needed. Even if the exam is
> fair enough that we can reach that conclusion, it is impossible to
> tell whether the instructor has been ineffective, or whether the
> students simply did not put in the effort to learn. The quality of
> instruction in computer science, the stability of the curriculum, and
> the level of preparation from high school are all quite variable. So
> we adjust the marks, fail the worst cases, and send a fair number of
> poorly-prepared students on to the next course, where the cycle
> repeats.
Isn't the obvious solution to this problem to make the assignments worth
fewer marks (or none) and actually fail the people who fail the exams?
Even if the reason for failure has more to do with the quality of the
instruction, why are we pushing people who don't know the material into
more challenging courses?
> Since the primary measurement of instructor effectiveness, apart from
> the failure rate, is the statistics from the anonymous student
> evaluations taken in about the second-last week of classes, any
> attempt by an individual instructor to alter this dynamic is doomed to
> failure. Nor do I think there is much institutional will to tackle the
> problem.
Why not?
-Josh
Prabhakar Ragde
> > So
> > we adjust the marks, fail the worst cases, and send a fair number of
> > poorly-prepared students on to the next course, where the cycle
> > repeats.
>
> Isn't the obvious solution to this problem to make the assignments worth
> fewer marks (or none) and actually fail the people who fail the exams?
> Even if the reason for failure has more to do with the quality of the
> instruction, why are we pushing people who don't know the material into
> more challenging courses?
Students do not like being treated as children, but in this respect
they often act like children: if the assignments are worth nothing,
they will not do them. Even quite self-aware and responsible students
will not do them. Assignments are worth less in engineering, but they
have a cohort system where classes stay together as they proceed
through a fixed course sequence, so it is hard to compare the
situations.
As for failing students who do not pass exams: you were recently a
student in a course I taught where a large number of students failed
the midterm. That was supposed to be a wakeup call, the value of which
I chose not to undercut by adjusting the marks (though I suggested
that adjustments would be made after the final exam). The result was a
large number of late withdrawals (WD), which caused anxious queries
from the CS advisors and from the people responsible for staffing the
course, as it might have meant an extra section required in the winter
or spring term. In addition, there was a general collapse in the marks
on subsequent assignments, and an overall atmosphere of bitterness and
negation.
Now, if I believed that every other instructor was willing to go
through that in order to ensure that standards were met, we might
collectively be able to do it. But I do not believe that. The midterm
average in this term's offering of that class was 3% higher than in
the class you took. But the instructors heeded my advice in advance,
which was to adjust the marks right away.
> > Nor do I think there is much institutional will to tackle the
> > problem.
>
> Why not?
Because it would mean admitting that there was a problem somewhere --
in the admissions process, in curriculum, in teaching, in evaluation,
maybe in all of those places, and it would mean getting people to give
up some of their autonomy in the classroom and do a lot of extra work
of the sort which is generally not rewarded. It's easier to trumpet
the achievements of our programming teams. --PR
Karen Amanda Yeats
difficult.
The assignments I would consider the best are the ones which are simply
too hard for even a good student to do by themselves. I think these are
the ones I learned most on. Now I say this as a pure math student so I'm
not talking about quite the same type of assignment that has been the
subject of this thread, and I was always sure to credit anyone or any book
which helped me.
At the same time it's clear to me that there are widespread problems with
real plagiarism in CS. Having quite stringent rules on excessive
colaberation, as I understand it, is done in large part so that real
plagiarism will never turn out looking grey. Through some combination of
prevailing culture and response to the former much possible colaberation
becomes quite grey in its own right. So assignments like I described
above simply can't be given.
This is what is done and plagiarism remains common. What is to be done?
In my experience people do keep a good eye out for cheaters, but there is
only so much to be done in that respect. In my opinion reasonable
colaberation has already been clamped down on as much as can be done
without strongly cutting into the educational side of things.
The engineering technique of not having assignments worth anything is, I
think, a very bad idea. As already noted the fact that assignments simply
don't get done is a problem, and while I'll agree that good mature folks
should do their assignments anyway, it is very understandable that under
competing pressures the things which aren't due don't get done. Notice
that engineering students have quite stringent course load requirements,
they are required to take enough courses to have a heavy work load every
normal term they are on campus, which among other effects exacerbates the
problem of not doing assignments which aren't worth anything. A different
problem is that without marked assignments students don't get any feedback
until the midterm. This is a real problem particularly, but not only,
when any substantial fraction of marks go to style issues or anything else
somewhat intangible. (Now I do base the above engineering comments on
mech eng which may not be typical.)
What happens if you simply start being stricter about removing the
possibilities for cheating. It's somewhat hard to see what to do about
assignments in that respect, but it's questionable whether it is worthwile
in any case. For instance a friend took a course where any midterm not
written in pen would not be remarked for any reason, and must be handed
back in the same class it was received in order to be remarked. Now
certainly there is a lot of mark grubbing when it comes to remarking exams,
but genuine problems occur far to often and the responsible student may
well want to check on some things before handing something back for
remarking. It really doesn't take much before techniques to avoid
cheating start to be irritating, insulting, and ultimately detrimental to
learning.
So what is the answer to the problem? I don't know. I really have
thought about it and it seems to me an impossible problem, once cheating
has established itself techniques in response only limit learning or
further destroy academic culture.
Karen
Rob Ewaschuk
> cred...@watserv1.uwaterloo.ca (Chris Redmond) writes:
>
> > Speaking as someone who never passed a programming course, I
> > don't exactly know how it's supposed to work . . . but isn't
> > it usual for the instructor or the textbook to explain "how
> > to do it", and then for the student to go and do it? If
> > after reading the textbook and hearing the instructor, the
> > student still doesn't know "how to do it", what is the approved
> > course of action, other than asking somebody else to teach
> > what's missing?
>
> Programming is not quite as creative as writing a short story or
> painting, but it is not just a matter of applying a simple technique
> in a straightforward fashion. There are decisions to be made,
> low-level details to be fleshed out, and your first try is unlikely to
> work as you expect it to. It is more or less true that a programming
> assignment can be done given sufficient time. Some students do not
> wish to put in that time, and other students are weak enough at it
> or do not budget their time, and so they try to save time by copying
> someone else's work. While this would be obvious plagiarism in the
> case of a short story or painting, it is not so obvious in
> programming, because correct submissions will resemble each other to a
> considerable extent.
My understanding of the situation (having been a tutor for CS130, and
having many tutor-contacts still) is that detecting cheating is not really
the problem is not finding candidate cheating cases, but pursuing them.
The available software that they have was not used, to the best of my
knowledge, because there was no available time/motivation to pursue the
cases reported.
If the problem was really determining cheating, I can imagine relatively
simple systems that made cheating more of a PITA than not cheating. Like
asking students to voluntarily submit their code every little while as
they work on it. Two solutions that appeared very similar could then have
the progression examined. It's a bit invasive, but it didn't take me long
to think of that idea and it does seem like other things along those lines
are available.
> After the first course or two, programming is more or less taken for
> granted, and students are expected to be able to take a software
> concept and implement it. Some assignments may not involve programming
> at all; they are done on paper and involve design and analysis. When
> mathematical techniques are involved, or problem-solving skills are
> exercised, it is quite possible that some students will not know how
> to do a question after some work. It is not clear, at this point,
> whether they just need to put in more work, or whether they are
> missing some knowledge or understanding necessary to complete the
> question.
Because of my habits on assignments, when I hit a wall on a conceptual
question it tends to be at 02:00. TAs are not available then, and the
assignments are frequently due the next day. Fault me for that if you
want, but I think that's a common enough trait amongst both profs and
students that, right or wrong, it should be considered as part of the
reality of the landscape. Now, at that point, I can turn to my friends or
roommates, and risk an excessive collaboration, or I can give up.
Having been hauled in for excessive collaboration once, I wouldn't want it
to happen again. Despite being found 'innocent', it left a bad taste in
my mouth. I think that my paranoia about this has been detrimental to my
learning.
Perhaps the right solution is to require collaboration information
question-by-question, and for the markers to cross-examine assignments
that claim collaboration. They could then, at their discretion, give a
mark of zero. The students could appeal this mark directly with the TA,
who could then get the student to work through a similar question to
regain the marks.
The risk of getting -100% on an entire assignment for hitting a wall on a
specific question is silly, and overdone. The paranoia I have about even
talking about assignments with my peers may be ill-founded, but I don't
think it is.
The current collaboration regime (I love that word and all its
contemporary connotations) has been severely detrimental to both my actual
learning, and my feelings about CS.
> It is my impression that copying assignments is more or less rampant
> in computer science. We face the dilemma that some skills cannot be
Yes. But those who do not, and *would* not, are being punished.
> learned without considerable practice. If we do not give credit for
> that practice, students will not do it, and then they do poorly on the
> final exam, which they typically must pass in order to pass the
> course. But giving any credit for the practice encourages students to
> earn that credit by cheating, which is a lot easier on an assignment
> than on a final exam. It also encourages reasoning of the sort, "This
> is worth only 5%, but it's taking me hours, it shouldn't take this
> long, I've put in my time and earned that 5%, I'm just going to copy
> the rest."
When I took CS341, one term before Josh, Therese made the assignments
worth marks, but if you didn't do them (IIRC, this only included the
written assignments -- the programming assignments were mandatory.) then
they would be filled in with what you got on the final exam. I don't
think I did a single written assignment, but I did ensure that I
*could* do them. I might have the programming/written thing backwards,
but I seem to recall doign programming assignments.
It's worth noting that there is significant grunt-work to go from "I know
and understand the solutions to each of these questions" to "I have
written this out in a form that I will submit", and generally I find that
this grunt work doesn't have much bang-for-buck. I have settled into a
habit of picking off the easy questions, or the questions that I think are
worth more marks than they should be, and taking a glance at the other
questions.
> We also have the problem that if a large number of students fail the
> final exam, it is difficult to make the case that they should not get
> the credit because they lack the skills needed. Even if the exam is
> fair enough that we can reach that conclusion, it is impossible to
> tell whether the instructor has been ineffective, or whether the
> students simply did not put in the effort to learn. The quality of
> instruction in computer science, the stability of the curriculum, and
> the level of preparation from high school are all quite variable. So
> we adjust the marks, fail the worst cases, and send a fair number of
> poorly-prepared students on to the next course, where the cycle
> repeats.
This cycle starts in CS1xx, where there are non-CS students that "have to
pass" those courses. This leads to an attitude of "everybody must pass"
that you find in some courses like STAT231 and CS354, but not in CS360 or
CS251. Perhaps the pass-fail division should be higher in these courses
for students planning on continuing in CS. I have long advocated an
'advanced' series of first year CS courses, which I think would have a
similar effect.
> Since the primary measurement of instructor effectiveness, apart from
> the failure rate, is the statistics from the anonymous student
> evaluations taken in about the second-last week of classes, any
> attempt by an individual instructor to alter this dynamic is doomed to
> failure. Nor do I think there is much institutional will to tackle the
> problem.
I don't follow the 'doomed to failure' bit.
> Is the problem worse than before? There's always been copying. But we
> are passing through (hopefully through) a cultural phase where the
> dominant metaphors on television and radio are the remake, the
> pastiche, the collage, the sample, the remix. Do a Google search for
> anything and see how many hits you get with chunks of identical prose
> in them. It is understandable, though perhaps not excusable, that
> students may undervalue individual effort. --PR
I think the problem is bad enough that I find it hard to believe that it
couldn't be better.
-Rob
--
: Rob Ewaschuk : University of Waterloo : Computer Science (4A) :
: www.studentforce.ca : Currently on co-op at www.decisionsoft.co.uk :
Prabhakar Ragde
> My understanding of the situation (having been a tutor for CS130, and
> having many tutor-contacts still) is that detecting cheating is not really
> the problem is not finding candidate cheating cases, but pursuing them.
>
> The available software that they have was not used, to the best of my
> knowledge, because there was no available time/motivation to pursue the
> cases reported.
I remember one report to the effect that, when they used the
cheating-detection software (which detects structural similarities,
and can't be fooled by simply renaming variables) on a CS 130
assignment, it flagged 30% of the assignments. Pursuing cases takes an
immense amount of time and emotional energy. The best we can do is a
sort of reverse lottery: try cheating, and much of the time you win a
little, but occasionally you lose a lot.
> Perhaps the right solution is to require collaboration information
> question-by-question, and for the markers to cross-examine assignments
> that claim collaboration. They could then, at their discretion, give a
> mark of zero. The students could appeal this mark directly with the TA,
> who could then get the student to work through a similar question to
> regain the marks.
TAs are graduate students who have been promised the work as part of
their stipends, and are supposed to be limited to a small number of
hours a week. (We also have full-time undergraduate tutors hired as
co-op students, but it is hard to fill those positions.) If professors
are not motivated to pursue plagiarism cases, TAs are even less
motivated.
> The current collaboration regime (I love that word and all its
> contemporary connotations) has been severely detrimental to both my actual
> learning, and my feelings about CS.
Do you have a suggestion that does not involve either "invasive
procedures", as you put it, or time-consuming examination of
individual cases?
> [...] those who do not, and *would* not, are being punished.
Only if you believe that the true reward is the numerical grade and
not the knowledge and skills gained by doing it yourself. Now, I
understand that in the "real world", that number does open doors. But
this line of reasoning seems to argue for taking the easiest courses
you can, all the way along.
> When I took CS341, one term before Josh, Therese made the assignments
> worth marks, but if you didn't do them (IIRC, this only included the
> written assignments -- the programming assignments were mandatory.) then
> they would be filled in with what you got on the final exam. I don't
> think I did a single written assignment, but I did ensure that I
> *could* do them. I might have the programming/written thing backwards,
> but I seem to recall doign programming assignments.
I spoke to Therese and Dan (the other instructor that term) about this
policy before settling on my marking scheme. They told me that the
failure rate was higher than usual, and biased towards those who had
not submitted written work during the term. In other words, they had
bought peace during the term, but at the cost of letting more students
dig their own graves. I decided against doing this, though in
retrospect, like every other decision I took that term, it seemed to
have been the wrong one.
> It's worth noting that there is significant grunt-work to go from "I know
> and understand the solutions to each of these questions" to "I have
> written this out in a form that I will submit", and generally I find that
> this grunt work doesn't have much bang-for-buck. I have settled into a
> habit of picking off the easy questions, or the questions that I think are
> worth more marks than they should be, and taking a glance at the other
> questions.
I think it is very easy to delude oneself into thinking one knows the
answer, only to realize, in the process of formalizing it, that
crucial details have been missed. I have this happen to me repeatedly
in writing up technical results for publication: this is the reason
why refereed technical papers are the gold standard when it comes to
assessing a researcher. That being said, we should try to avoid
excessive grunt work. But we are likely not to agree on the definition
of exactly what that is.
> > Since the primary measurement of instructor effectiveness, apart from
> > the failure rate, is the statistics from the anonymous student
> > evaluations taken in about the second-last week of classes, any
> > attempt by an individual instructor to alter this dynamic is doomed to
> > failure. Nor do I think there is much institutional will to tackle the
> > problem.
>
> I don't follow the 'doomed to failure' bit.
Perhaps not "doomed"; that's a bit fatalistic. But an isolated effort
to only grant credit for having achieved a proper standard will be
seen as unfair relative to other courses the student has taken (in
which, perhaps, they were lobbed too many easy pitches) and the fault
ascribed to the instructor. --PR
Gregory Lipscomb
>student in a course I taught where a large number of students failed
>the midterm. That was supposed to be a wakeup call, the value of which
>I chose not to undercut by adjusting the marks (though I suggested
>that adjustments would be made after the final exam). The result was a
>large number of late withdrawals (WD), which caused anxious queries
>from the CS advisors and from the people responsible for staffing the
>course, as it might have meant an extra section required in the winter
>or spring term.
they didn't add an extra section this summer. I know this because the
only available section available conflicts with the only section of FR152.
My attempts to further my french education thus being squashed.
J'etais calme vexe : (
--
Greg Lipscomb
Computer Science
http://www.student.math.uwaterloo.ca/~glipscom
"The danger already exists that mathematicians have made a covenant with
the devil to darken the spirit and confine man in the bonds of Hell."
- St. Augustine
Matthew Skala
Prabhakar Ragde <plr...@plg2.math> wrote:
>co-op students, but it is hard to fill those positions.) If professors
>are not motivated to pursue plagiarism cases, TAs are even less
>motivated.
I always pursue plagiarism cases because as I see it, if plagiarism goes
unpunished then everyone in the system is hurt, including me. My
perception of where the problem is may be a little skewed because I have
tended to TA for more theory-intensive courses, where the work handed in
tends to be written in some approximation of English rather than in a
programming language. But my experience has been that the problem is more
often in the nature of reference abuse rather than students copying each
others' finished work.
I've seen, for instance, several essays that consisted almost entirely of
quotes from works on the references list, run into the text along with a
few original sentences so that (except for glaring style differences) one
might think all the text was original. I'm sure the students who "wrote"
those essays thought that it was a perfectly legitimate way to write
essays as long as they listed the bibliographic data of the works they
copied from. If they knew they were doing something against the rules,
then they would have tried to conceal it somehow instead of handing me all
the evidence in the form of the references list. I also once saw an
assignment with an answer copied word-for-word from the answer guide for
the course textbook, with a note added at the end saying "Note: I got this
answer from the textbook answer guide, page so-and-so". I really wanted
to give someone a smack upside the head when I saw that one.
Maybe there's simply a lot of unauthorized collaboration and code-copying
going on that has escaped my notice. It is true that every once in a
while I'll come across two assignments that are formatted similarly and
I'll wonder if those two people were studying together. But some amount
of collaboration or shared studying is allowed, I think most cases of
cheating are obvious as soon as they're examined, and so if I look at a
pair of assignments a second time and I think it's plausible that they
were done legitimately, I'll let them go. The actual cheating cases are
easy to recognize and easy to prove.
This thread started with a student asking "How should I do this question?"
and someone suggesting that that was obviously, and laughably, a request
for inappropriate collaboration. I don't think it is obviously a request
for inappropriate collaboration. I think it could well be just a slightly
clumsy way of asking for perfectly acceptable collaboration. I will often
have students come to me in my office hour and ask exactly that, "How
should I do this problem?" and I'll typically help them find the
appropriate section in the textbook. I think that's my job, and it's not
clear to me that if they get the same kind of help from a classmate, that
that's any worse than getting it from me. If I answered the question "How
should I do this?" by handing them an answer to copy onto their
assignment, then I wouldn't be doing my job properly, and if a classmate
gave them that kind of answer, they would clearly be cheating; but I don't
think asking the question "How should I do this?" necessarily indicates
that anything bad is going on.
--
Matthew Skala, CS PhD student, University of Waterloo
msk...@math.uwaterloo.ca <-- school
msk...@ansuz.sooke.bc.ca <-- home
http://ansuz.sooke.bc.ca/ http://www.edifyingfellowship.org/
Anton
Let Them Cheat!
it seems to me that cheating by students is as old as schools.
i'm guessing that if 2000+ years of human ingenuity haven't
given us a solution to the "cheating students" problem, nothing ever will.
This will end up being an endless cat-and-mouse game that will drain
resources from the real goal of the school -- to educate those who
are willing to learn.
a case could be made for the fact that everywhere outside of school
collaboration is encouraged (every single job I am applying to lists
"team work" or "team player" as a requirement). Furthermore
I value resourcefulness higher then "book-smarts" and I don't think I'm
alone on this one. In other words, I may not know the answer to something
but i sure as hell know where to get one! (find a book, or a smart
roommate). I hate exams for this reason. I don't like to memorize
rather useless facts, because "in the real world" if I don't remember
something like Java syntax, sorting algorithm or runtime of Divide and
Conquer, i can *easily* look these things up. Further on, when a person
is working in the real world, the company will encourage him to ask for
input from his coworkers when the person is stuck on a problem. If I
don'tknow how to do something at work, i ask someone.. there's no shame
in not knowing in the workplace, unlike at a school
I must agree that for the most part throughout my UW Math degree the exams
required more thinking and analytical skills then memorization. but there
was a fair amount of rather useless (i think) memorization still.
for those who will think that cheating devalues a degree, i don't buy that
argument. You can only help those who help themselves -- or in this case
educate those who are willing to educate themselves. As such, there will
be students who will be dilligent in their assignment, and those who will
cheat. The frauds will be exposed at one point or another.
i believe that if we were to ignore cheating we would live in a happier
world
Anton -- CompSci / SoftEng @ UWaterloo
------------------------------------------
Everybody seems to think I'm lazy;
I don't mind, I think they're crazy.
Running everywhere with such a speed;
Till they find, that there's no need.
Prabhakar Ragde
> How about this solution:
> Let Them Cheat!
So you'd be fine with 100% exams?
> a case could be made for the fact that everywhere outside of school
> collaboration is encouraged (every single job I am applying to lists
> "team work" or "team player" as a requirement). Furthermore
> I value resourcefulness higher then "book-smarts" and I don't think I'm
> alone on this one. In other words, I may not know the answer to something
> but i sure as hell know where to get one! (find a book, or a smart
> roommate).
Yeah, fine, but someone has to write those books or be those smart
roommates. Sure, it's artificial to have you work alone, but all of
school is an artificial environment to develop a broad range of
learning skills in you. You're solving questions on assignments that
people have already solved. What do you do in a situation where your
employer expects you to solve a problem, it's proprietary, and the
people who know most about it are working for your competitors?
> for those who will think that cheating devalues a degree, i don't buy that
> argument. You can only help those who help themselves -- or in this case
> educate those who are willing to educate themselves. As such, there will
> be students who will be dilligent in their assignment, and those who will
> cheat. The frauds will be exposed at one point or another.
Ideally, perhaps. You and I know that frauds can get a very long way
and can die of old age quite comfortably wealthy. --PR
Karen Amanda Yeats
virtues of friends and references. But as you note this is in part at
least a different sort of problem: the problem of how to pass on the
culture of what is and isn't acceptible instead of the problem of catching
people who are deliberately breaking the (plagiarism in this case) rules.
Karen
Jonathan Lee
Prabhakar Ragde <plr...@plg2.math> wrote:
>policy before settling on my marking scheme. They told me that the
>failure rate was higher than usual, and biased towards those who had
>not submitted written work during the term. In other words, they had
>bought peace during the term, but at the cost of letting more students
>dig their own graves. I decided against doing this, though in
>retrospect, like every other decision I took that term, it seemed to
>have been the wrong one.
This is purely out of curiosity, but what did the second programming
assignment originally entail? I wasn't quite sure what to think when
you claimed that something related to Floyd-Warshall would make for
an overly distracting assignment.
Jon Lee
Prabhakar Ragde
> This is purely out of curiosity, but what did the second programming
> assignment originally entail? I wasn't quite sure what to think when
> you claimed that something related to Floyd-Warshall would make for
> an overly distracting assignment.
[There were supposed to be two programming and five written
assignments, but after the weak performance on the first programming
assignment and the midterm, we cancelled the second one.]
I don't think we ever formalized a second programming assignment. Due
to circumstances, we were behind in prep throughout the term --
normally I would have had lecture modules and assignments ready before
the term began. I think I was being ironic when referring to
Floyd-Warshall, which after all is five lines of pseudocode (and not
many more lines of actual working code). --PR
Prabhakar Ragde
> I think most cases of
> cheating are obvious as soon as they're examined, and so if I look at a
> pair of assignments a second time and I think it's plausible that they
> were done legitimately, I'll let them go. The actual cheating cases are
> easy to recognize and easy to prove.
It is my experience that there are only two types of plagiarism I can
catch in a theoretical course. One is when two people have the same
bizarre incorrect solution, and the other is when I find people using
phrases from a known solution. --PR
Julie Lavoie
> a case could be made for the fact that everywhere outside of school
> collaboration is encouraged (every single job I am applying to lists
> "team work" or "team player" as a requirement).
> for those who will think that cheating devalues a degree, i don't buy that
> argument. You can only help those who help themselves -- or in this case
> educate those who are willing to educate themselves. As such, there will
> be students who will be dilligent in their assignment, and those who will
> cheat. The frauds will be exposed at one point or another.
Part of the problem is that school has two purposes: one is to teach you
stuff, another is to certify to the world at large (employers, grad
schools, etc) that you know that stuff, and quantify the degree to which
you know it. If its sole purpose were to teach, then I agree with you
that it would make more sense to encourage collaboration. I think that
"in the real world" (TM) situations when you are encouraged to collaborate
with your colleagues are more common than the kind of situation Prabhakar
describes.
However, because it is also school's job to quantify how much you
know to the outside world, I believe that cheating _does_ devalue the
value of a degree. You can argue that marks don't necessarily reflect how
much you know, and school should not serve this purpose (I would agree
with you), but they _are_ used by employers, scholarship committees and
other universities to evaluate how much you are worth and what goodies
you get. I take the advanced math classes, where a lot of the scholarship
holders for Math end up --- incidents like having classmates ask me to
copy all my assignments really make me wonder how much of the marks that
got them here with scholarships were really theirs.
-Julie
Howard Cheng
> I always pursue plagiarism cases because as I see it, if plagiarism goes
> unpunished then everyone in the system is hurt, including me. My
> perception of where the problem is may be a little skewed because I have
> tended to TA for more theory-intensive courses, where the work handed in
> tends to be written in some approximation of English rather than in a
> programming language. But my experience has been that the problem is more
> often in the nature of reference abuse rather than students copying each
> others' finished work.
I have also seen my share of examples, and all my experience comes
from theory-intensive courses...I would have to say that most of them
are not "plagiarism" but rather "excessive collaboration". e.g. A
group of people would work together and come up with the same wrong
answer. It is not easy to spot this when they make a common mistake
(is it because everyone worked together, or the same idea got passed
around, or is it just a mistake people are expected to make?), but it
is very obvious when the people in a group make the same strange
error. When I call them in they admit collaborating on the assignment
99% of the time. When I taught with Therese our policy was to call
the students in and give only warnings if they can explain what they
wrote. If they clearly have no understanding of what they wrote then
we punish them (or maybe warn them heavily if the offense was minor).
I have only seen straight copying from some reference without
acknowledgement a few times.
Howard
---
Howard Cheng e-mail: hchc...@scg.math.uwaterloo.ca
University of Waterloo URL : http://www.scg.uwaterloo.ca/~hchcheng/
Computer Science Graduate Student (PhD)
"Having first claimed that it did not shoot the victim, and then that
everyone does it, and then that the victim would have died anyway,
Microsoft now argues that the victim is unharmed." - gov't anti-trust lawyers
Howard Cheng
> msk...@softbase.math.uwaterloo.ca (Matthew Skala) writes:
>
> > I think most cases of
> > cheating are obvious as soon as they're examined, and so if I look at a
> > pair of assignments a second time and I think it's plausible that they
> > were done legitimately, I'll let them go. The actual cheating cases are
> > easy to recognize and easy to prove.
>
> It is my experience that there are only two types of plagiarism I can
> catch in a theoretical course. One is when two people have the same
> bizarre incorrect solution, and the other is when I find people using
> phrases from a known solution. --PR
And note that this may also provide an answer to those who believe
cheating or collaboration should be allowed, in the following sense.
If you are smart enough to collaborate with someone and get the right
answer, you probably actually understand the solution and it is very
difficult for us to catch. If you copy your ideas from someone and
actually understand the ideas, you can probably write up the solution
appropriately and no one can tell. But if you just copy word for word
without understanding, then you would probably get caught. Even
if you believe in facilitating learning through collaboration, you
can't really support collaboration without learning anything, right?
Howard Cheng
> Part of the problem is that school has two purposes: one is to teach you
> stuff, another is to certify to the world at large (employers, grad
> schools, etc) that you know that stuff, and quantify the degree to which
> you know it. If its sole purpose were to teach, then I agree with you
> that it would make more sense to encourage collaboration. I think that
> "in the real world" (TM) situations when you are encouraged to collaborate
> with your colleagues are more common than the kind of situation Prabhakar
> describes.
I would say that collaboration should be encouraged only so that everyone
learns from the experience. But in practice collaboration means a few
stronger people giving their ideas to others, who will use the ideas
without much, if any, understanding.
David Evans
Howard Cheng <hchc...@scg.math.uwaterloo.ca> wrote:
>If you are smart enough to collaborate with someone and get the right
>answer, you probably actually understand the solution and it is very
>difficult for us to catch. If you copy your ideas from someone and
>actually understand the ideas, you can probably write up the solution
>appropriately and no one can tell. But if you just copy word for word
>without understanding, then you would probably get caught.
There is also the possibility that you collaborate with someone and are
convinced (or do the convincing) that the bizarre, incorrect solution is
actually right.
Howard Cheng
> In article <Pine.SOL.4.44.030314...@scg.math.uwaterloo.ca>,
> Howard Cheng <hchc...@scg.math.uwaterloo.ca> wrote:
> >If you are smart enough to collaborate with someone and get the right
> >answer, you probably actually understand the solution and it is very
> >difficult for us to catch. If you copy your ideas from someone and
> >actually understand the ideas, you can probably write up the solution
> >appropriately and no one can tell. But if you just copy word for word
> >without understanding, then you would probably get caught.
>
> There is also the possibility that you collaborate with someone and are
> convinced (or do the convincing) that the bizarre, incorrect solution is
> actually right.
But in that case they would be punished...either for not getting it
right, or collaborating and getting it wrong, or both.
Howard
---
Howard Cheng e-mail: hchc...@scg.math.uwaterloo.ca
University of Waterloo URL : http://www.scg.uwaterloo.ca/~hchcheng/
Computer Science Graduate Student (PhD)
Perhaps the greatest paradox of all is that there are paradoxes in
mathematics. - E. Kasner & J. Newman
Prabhakar Ragde
> Part of the problem is that school has two purposes: one is to teach you
> stuff, another is to certify to the world at large (employers, grad
> schools, etc) that you know that stuff, and quantify the degree to which
> you know it. If its sole purpose were to teach, then I agree with you
> that it would make more sense to encourage collaboration. I think that
> "in the real world" (TM) situations when you are encouraged to collaborate
> with your colleagues are more common than the kind of situation Prabhakar
> describes.
I don't believe I ever said anything about how common collaboration
was in the real world, and I also don't believe that collaboration is
all right when the sole purpose of education is to teach -- because I
more or less act as if that is the case. Collaboration is a way of
adding together individual talents to achieve something more or faster
than individuals can by themselves, but those individual talents have
to be developed somehow. We have groupwork in the curriculum at
several points to foster necessary skills such as communication and
consensus, but this in no way obviates the need for the development of
individual skills. --PR
Rick De Visser
'collaborate' because they don't have the capacity to engineer a solution on
their own.
I say the system works - encourage assignment with marks, if they don't get
the theory with or without collaboration, etc, they will shoot themselves
(in the foot)in the exam.
Obvious gross plagiarism can and should be dealt with.
Labs often have in-class assignments which will narrow the field.
"Prabhakar Ragde" <plr...@plg2.math> wrote in message
news:x6k7f1a...@plg2.math...
David Evans
Howard Cheng <hchc...@scg.math.uwaterloo.ca> wrote:
>On Fri, 14 Mar 2003, David Evans wrote:
>
>> There is also the possibility that you collaborate with someone and are
>> convinced (or do the convincing) that the bizarre, incorrect solution is
>> actually right.
>
>But in that case they would be punished...either for not getting it
>right, or collaborating and getting it wrong, or both.
>
Well, obviously they should be punished for not getting it right since
they didn't get it right.
Mike
Fri, 14 Mar 2003 23:32:54 +0000 (UTC)
>
> Well, obviously they should be punished for not getting it right since
> they didn't get it right.
Now now, in today's society you cannot punish individuals for such things.
Rather, they should not be rewarded for not having attained the so-called
"correct" answer.
Mike
Matthew Nichols
any of the university's business whether or not students collaborate on
assignments.
Now, whether this collaboration can be done in a newsgroup hosted by the
university is another issue. I'm simply objecting to the stance of some
deparments/schools (specifically computer science) that students cannot
freely discuss the assignment amongst themselves in whatever manner they
choose.
Certainly *once a student hands in something* that is not his or her own
work, then they should receive no marks, since a student's grade in a
course is supposed to reflect the work he or she has personally done. I
also feel that the student who originally did the work (who "let" others
copy it) should not be penalized in any way. This is assuming, of course,
that it is clear who did the original work and who did the copying.
Otherwise you enter into an entirely more complex discussion.
A university education is supposed to be about the free exchange of ideas.
What someone chooses to do with those ideas should not be the
responsibility of those who conveyed the ideas. And such exchanges of
ideas should not be prohibited.
./matt
On Wed, 12 Mar 2003, Reinhold Friedrich Burger wrote:
>
> A recent exchange on a cs course newsgroup. Some editing was done
> for anonymity, but the last line, in quotes, is the student's
> complete and literal response. And no, I'm not the TA.
>
>
> -------------------------------------------------------------------------
> Student (posting from school account):
> If anyone knows how to do questions 2 & 3 of the assignment,
> please email me at <student's yahoo account>.
>
>
> TA (to students in general):
> Careful! Please read the assignment guidelines before you answer.
>
> (The guidelines deal with the usual things, in particular excessive
> collaboration.)
>
>
> Student's indignant response:
> "I'm not asking for the answers, just how to do it."
> -------------------------------------------------------------------------
>
>
Prabhakar Ragde
> I probably have a "unique" view on this topic, but I don't feel that it's
> any of the university's business whether or not students collaborate on
> assignments.
You are seeking a credential from the university that is for
individual achievement and widely interpreted as such. The university
has the right to set the rules under which you are granted that
credential, as long as they do so in a consistent fashion. You have
the right to refuse to earn that credential, but you don't have the
right to not play by their rules and then demand it.
> Certainly *once a student hands in something* that is not his or her own
> work, then they should receive no marks, since a student's grade in a
> course is supposed to reflect the work he or she has personally
> done.
This contradicts what you said above. If students are to receive zero
for work which is not their own, then the university has to be
concerned about identifying these cases. I am sympathetic to the idea
that if someone hands in work and writes at the bottom, "I copied this
from so-and-so," they are merely guilty of idiocy and should get
zero. But if they do not admit copying, they are guilty of academic
fraud and deserve a greater penalty.
> I also feel that the student who originally did the work (who "let" others
> copy it) should not be penalized in any way.
They are penalized for subverting the process of assessing individual
achievement.
> A university education is supposed to be about the free exchange of ideas.
> What someone chooses to do with those ideas should not be the
> responsibility of those who conveyed the ideas. And such exchanges of
> ideas should not be prohibited.
You are being idealistic. In a setting where the university simply
held lectures and did nothing else, this would work fine. In the
current setting, where students are earning credentials, this begs for
100% of the marks to be earned through exams. --PR
Rob Ewaschuk
You've mentioned 100% exams a couple of times, as though they are a
non-starter.
An idea I've had in the back of my head for a while is to allow students
to choose from a small set of marking schemes at the *beginning* of the
term. You could offer a standard marking scheme (25% midterm, 65% final,
10% assignments or whatever), and a 100% final scheme. A student could
choose a 100% final, at which point their assignments would be marked to
help them out. You could even require that they submit (and pass) some %
of their assignments in order to stay on their chosen marking scheme if
you were really worried about them. You could also allow 100% marking
scheme only on the permission of the instructor.
Students that chose 100% marking schemes would be free to collaborate as
much or as little as they wanted on their assignments. They would only be
penalized if they were found to "subvert the process of assessing
individual achievement" (I don't think I buy that, but it's not an
argument I feel strongly enough to pursue.)
Students that had historically chosen 100% marking schemes and failed
would not be allowed to choose them again.
There's obviously some infrastructure required there if you want to
formally track things, but I don't think it's any more odious than the
current scheme, and you'd effectively eliminate trying to identify
excessive collaboration in some segment of your population.
Had I had this option, I'm fairly certain I would have taken it and done
better (numerically and academically) in several courses.
It obviously doesn't fit the mold for some courses like OS, and CS241, but
it fits well for most non-CS math courses and many CS courses.
Rob Ewaschuk
To followup on my own post, I'd even support a 100% assignments scheme, on
penalty of death for any collaboration. If people think they can go it
alone, so be it. Although perhaps the assignments would need to be more
strenuous for this group. The idea would be to help out those people who
simply don't perform well under the stress of exams. (Admittedly, any
suggestion that would require maintaining parallel streams of assignments
is probably a non-starter for most course staff.)
Prabhakar Ragde
> You've mentioned 100% exams a couple of times, as though they are a
> non-starter.
I think they are, for two reasons you identified: students can't be
trusted to get the work done during the term that they need to do to
pass the final (even with the current scheme, we have lots of
last-minute cramming), and some people don't do well on exams. In
addition, the set of problems one can set on an exam is a small
fraction of the set of problems one wants students to be able to solve
upon leaving the course. If it were logistically possible, I'd give
exams which looked like assignments and took about as long.
> An idea I've had in the back of my head for a while is to allow students
> to choose from a small set of marking schemes at the *beginning* of the
> term. You could offer a standard marking scheme (25% midterm, 65% final,
> 10% assignments or whatever), and a 100% final scheme.
If you can convince me of the merits of any arrangement, I'm willing
to try it with my fall classes (assuming you can also convince any
co-instructors I may have). Different marking schemes are just a
matter of spreadsheet configuration, assuming you aren't suggesting
different sets of assignments. But why should anyone opt for a 100%
final? Currently, they can get some fraction of their marks by getting
away with cheating, which appears pretty easy. With the new scheme,
they don't have this option. In other words, I don't think this
will solve the problem. I'm not even sure why you would have opted for
this in past courses, except if you feel that the assignments were
makework that just got in the way of real learning (which might be a
defensible argument for some courses). --PR
Prabhakar Ragde
> To followup on my own post, I'd even support a 100% assignments scheme, on
> penalty of death for any collaboration.
But you've argued that the current penalty is excessive and inhibits
learning. It's clear we can't catch all or even most collaboration,
and making a case is a matter of interpreting ambiguous situations.
So, to summarize:
100% final: more foolproof, but inhumane, and widens gap between
testing instrument and desired goals
current scheme (roughly 75% exams, 25% assignments): appears to encourage
copying on assignments
100% assignments: really encourages copying on assignments
Rob Ewaschuk
> Rob Ewaschuk <raew...@fenchel.math.uwaterloo.ca> writes:
>
> > To followup on my own post, I'd even support a 100% assignments scheme, on
> > penalty of death for any collaboration.
>
> But you've argued that the current penalty is excessive and inhibits
> learning. It's clear we can't catch all or even most collaboration,
> and making a case is a matter of interpreting ambiguous situations.
That's because the current system is *non-optional*. I am perfectly
comfortable collaborating on assignments. I know when I'm learning and
when I'm not. I would not copy an assignment just to get the marks. Yet
I am required (coerced?) into being paranoid about any collaboration I do.
If people choose 100% assignments, then they choose the penalty that goes
along with it. As it is, I have no choice but to be involved in a system
where I'm paranoid to do some things that will help me learn and (since
when I give up on doing it by myself and seek help, I lose the ability to
write up my new understanding and get marks for it.)
> So, to summarize:
>
> 100% final: more foolproof, but inhumane, and widens gap between
> testing instrument and desired goals
Perfectly humane if done by choice. And I'm not sure it widens said gap.
It certainly does at a first glance, but then I sincerely believe I lose
significant marks on my assignments because I err on the side of paranoid
caution. If I collaborated to the fullest point that each prof found
acceptable, I'm sure I'd do much better on my assignments. I'm certain I
err on the side of less than what profs find acceptable, because under the
current regime I *must* be certain.
> current scheme (roughly 75% exams, 25% assignments): appears to encourage
> copying on assignments
> 100% assignments: really encourages copying on assignments
Think so? First, how many people do you think would actually take this
route? If they were on their own assignment stream, it wouldn't be hard
to cross-check a lot of them. With a severe enough penalty I think people
would be deterred. You could also be required to write the final (at a
weight of zero) to ensure you knew some of it in a less cheatable
environment.
Another factor is that if you allow people to choose, they just might take
some responsibility for their choices. Perhaps I'm being too idealistic.
Rob Ewaschuk
> Rob Ewaschuk <raew...@fenchel.math.uwaterloo.ca> writes:
>
> > You've mentioned 100% exams a couple of times, as though they are a
> > non-starter.
>
> I think they are, for two reasons you identified: students can't be
> trusted to get the work done during the term that they need to do to
> pass the final (even with the current scheme, we have lots of
> last-minute cramming), and some people don't do well on exams. In
Which students? All students? Do all students need to get the work done?
Would the cramming really get any worse?
> addition, the set of problems one can set on an exam is a small
> fraction of the set of problems one wants students to be able to solve
> upon leaving the course. If it were logistically possible, I'd give
> exams which looked like assignments and took about as long.
Yup. Which is solved by still requiring some work on assignments, even
though it's not worth any marks.
> If you can convince me of the merits of any arrangement, I'm willing
> to try it with my fall classes (assuming you can also convince any
> co-instructors I may have). Different marking schemes are just a
> matter of spreadsheet configuration, assuming you aren't suggesting
> different sets of assignments. But why should anyone opt for a 100%
> final? Currently, they can get some fraction of their marks by getting
> away with cheating, which appears pretty easy. With the new scheme,
This thought pattern is perturbing. I think it's showing that you've
institutionalized the exact adversarialism that frustrates me so much. I
would not cheat. I don't care how easy or hard it is. I play by the
rules, and many other students will too. I even play by the rules when I
think they're dumb, but I try to get verbose about it when I can. :-)
I would choose 100% final because I don't like doing assignments. I don't
find them to be a good use of my time, either academically or marks-wise.
I tend to pick off the easy questions to get the marks for my time, and
look at the hard questions and think about them to get the learning for my
time. I seldom get more than 60% on my assignment mark, and I generally
do much better than that on my finals. I learn from lectures, not
textbooks or assignments.
They aren't worth my time academically because to learn the most from
them, I would have to collaborate so much that I wouldn't take the risk of
handing them in. Zero marks is unacceptable to me, though I often
consider it. In playing the game 'the system' forces me to play, I end up
with lower marks and less learning than I feel I ought to get.
Occasionally, yes, an assignment question is much harder than it appears.
Like doing a nice clean recursive proof to Brook's Theorem, which ended up
being unfinished after a lot of effort and acceptable collaboration. I
think these are few. I think they are also probably common enough that a
prof can predict them, and require them even for 100%-final students.
> they don't have this option. In other words, I don't think this
> will solve the problem. I'm not even sure why you would have opted for
We might be trying to solve different problems, but it seems to solve my
problem, which is the 'playing the system' that I end up doing.
> this in past courses, except if you feel that the assignments were
> makework that just got in the way of real learning (which might be a
> defensible argument for some courses). --PR
I think I've clarified that point above. I'd be happy to expound more,
though. :-)
Joshua Chud
> If you can convince me of the merits of any arrangement, I'm willing
> to try it with my fall classes (assuming you can also convince any
> co-instructors I may have). Different marking schemes are just a
> matter of spreadsheet configuration, assuming you aren't suggesting
> different sets of assignments. But why should anyone opt for a 100%
> final? Currently, they can get some fraction of their marks by getting
I would opt thus in almost all of my classes. Just as some people do very
poorly on exams, some people do very well and I'm one of those. I also
tend to do poorly (relatively) on assignments. Think is has something to
do with my inability to do good work unless I'm under severe time
pressure. The same kind of thing that helps me to stay calm and work
quickly on exams.
> away with cheating, which appears pretty easy. With the new scheme,
> they don't have this option. In other words, I don't think this
> will solve the problem. I'm not even sure why you would have opted for
> this in past courses, except if you feel that the assignments were
> makework that just got in the way of real learning (which might be a
> defensible argument for some courses). --PR
It might even be defensible for some of yours. I'm pretty sure (I'd have
to go back and check) that both times I've had you as an instructor I've
done far better on my exmas than on the assignments. (Maybe this has to do
with leniency on exam marking schemes?)
-Josh
Howard Cheng
> Perfectly humane if done by choice. And I'm not sure it widens said gap.
> It certainly does at a first glance, but then I sincerely believe I lose
> significant marks on my assignments because I err on the side of paranoid
> caution. If I collaborated to the fullest point that each prof found
> acceptable, I'm sure I'd do much better on my assignments. I'm certain I
> err on the side of less than what profs find acceptable, because under the
> current regime I *must* be certain.
I am not completely convinced when people use this argument. Many
people who work by themselves can do well on their own. Many people
who collaborate just end up with sharing wrong ideas and solutions.
This is at least true for theory type courses.
Howard
---
Howard Cheng e-mail: hchc...@scg.math.uwaterloo.ca
University of Waterloo URL : http://www.scg.uwaterloo.ca/~hchcheng/
Computer Science Graduate Student (PhD)
The C Programming Language -- A language which combines the flexibility of
assembly language with the power of assembly language.
Joshua Chud
> Occasionally, yes, an assignment question is much harder than it appears.
> Like doing a nice clean recursive proof to Brook's Theorem, which ended up
> being unfinished after a lot of effort and acceptable collaboration. I
That's "inductive" not "recursive", dumbass :-). Go directly to MATH135,
do not pass go, do not collect $3.14.
And it was only clean for the first five minutes.
-Josh
Rob Ewaschuk
Recursive..Inductive. Been on a work term for too long. :-)
$3.14 CAD or USD? :-)
-Rob
--
: Rob Ewaschuk : University of Waterloo : Computer Science (4A) :
: www.studentforce.ca : Currently on co-op at www.decisionsoft.co.uk :
Rob Ewaschuk
> On Tue, 18 Mar 2003, Rob Ewaschuk wrote:
>
> I am not completely convinced when people use this argument. Many
> people who work by themselves can do well on their own. Many people
> who collaborate just end up with sharing wrong ideas and solutions.
> This is at least true for theory type courses.
Hmm. All I was arguing was that a 100% final is not obviously a less
accurate testing mechanism than assignments where I don't do them and a
60% final.
You seem to be arguing about whether I can decide when I'm learning
through collaboration or not. That's a fair question, but since it all
comes out on the exam I'd be daft to collaborate when I wasn't learning.
I can do fine on my own, I just don't enjoy it, and I don't stick to it.
Marks just aren't enough of an incentive for me. When learning becomes
the incentive I'd rather learn a little less and enjoy it a lot more
(since I'll continue it if I'm enjoying it, and I'll stop if I'm not.)
Prabhakar Ragde
> > I think they are, for two reasons you identified: students can't be
> > trusted to get the work done during the term that they need to do to
> > pass the final (even with the current scheme, we have lots of
> > last-minute cramming), and some people don't do well on exams. In
>
> Which students? All students? Do all students need to get the work done?
> Would the cramming really get any worse?
Not all students need to get the work done. Most do. A smaller number
will realize on their own what they need to do and get it done during
the term instead of in the last few weeks. So if you remove term work
incentives, yes, there will be more cramming.
> >If it were logistically possible, I'd give
> > exams which looked like assignments and took about as long.
>
> Yup. Which is solved by still requiring some work on assignments, even
> though it's not worth any marks.
So what is the penalty if they don't do the work? If there's a
penalty, then you are basically assigning marks for assignment work,
you're just moving the zero point.
> > But why should anyone opt for a 100%
> > final? Currently, they can get some fraction of their marks by getting
> > away with cheating, which appears pretty easy. With the new scheme,
>
> This thought pattern is perturbing. I think it's showing that you've
> institutionalized the exact adversarialism that frustrates me so much. I
> would not cheat. I don't care how easy or hard it is. I play by the
> rules, and many other students will too. I even play by the rules when I
> think they're dumb, but I try to get verbose about it when I can. :-)
Rob, I didn't institutionalize anything. I'm just telling you my
working model based on experience with about twenty years' worth of
students, not all of whom are like you. Of course you play by the
rules. You do a lot of things around here because you think it's
right, not because you're being forced. But I can't plan for the ideal
student. I have to plan for the typical student, and in some cases for
the worst student. The typical student is decent, but if they see a
loophole being exploited by others, they'll go for it. Remember the
first piece of trash on the ground argument?
> I would choose 100% final because I don't like doing assignments. I don't
> find them to be a good use of my time, either academically or marks-wise.
> I tend to pick off the easy questions to get the marks for my time, and
> look at the hard questions and think about them to get the learning for my
> time. I seldom get more than 60% on my assignment mark, and I generally
> do much better than that on my finals. I learn from lectures, not
> textbooks or assignments.
Dare I suggest to you that on one of my exams you might not do as well
as you think? I tend to ask problem-solving questions of the type that
require practice. If you haven't done the assignments, you've
understood the problems solved in class, but you haven't done any on
your own. You may be blessed with a high degree of self-awareness
regarding your limitations and how to overcome them, but again I don't
think that's typical.
>> 100% assignments: really encourages copying on assignments
>
>Think so? First, how many people do you think would actually take this
>route?
Lots. People loathe exams. Assignments can be done on one's own time,
in one's own room... and, as I've said, cheating is hard to detect.
>If they were on their own assignment stream, it wouldn't be hard
>to cross-check a lot of them. With a severe enough penalty I think people
>would be deterred.
As it is, each case we prosecute takes up way too much time and
effort. It would only get worse if the penalty was more severe. Have
you ever done marking for a course? I think if you had, you wouldn't
be saying "it wouldn't be hard" about any of this.
>Another factor is that if you allow people to choose, they just might take
>some responsibility for their choices. Perhaps I'm being too idealistic.
It's certainly no problem for me to offer a 100% final option. I'll
call it the Rob Ewaschuk option, so they don't blame me when they fail
the course. --PR
Prabhakar Ragde
> It might even be defensible for some of yours. I'm pretty sure (I'd have
> to go back and check) that both times I've had you as an instructor I've
> done far better on my exmas than on the assignments. (Maybe this has to do
> with leniency on exam marking schemes?)
I don't have the 251 marks on this machine, but yes, this was true for
you for 341. It wasn't leniency. You aren't motivated to do the
assignments, and you get through on raw ability. You would have done
even better, I suspect, had you bothered with the assignments. --PR
Joshua Chud
> Joshua Chud <jec...@uwaterloo.ca> writes:
>
> > It might even be defensible for some of yours. I'm pretty sure (I'd have
> > to go back and check) that both times I've had you as an instructor I've
> > done far better on my exmas than on the assignments. (Maybe this has to do
> > with leniency on exam marking schemes?)
>
> I don't have the 251 marks on this machine, but yes, this was true for
> you for 341. It wasn't leniency. You aren't motivated to do the
> assignments, and you get through on raw ability. You would have done
> even better, I suspect, had you bothered with the assignments. --PR
Perhaps, although I think I'd probably have done optimally well on the
exam by ignoring the assignments and reading some Knuth. (I wouldn't have
done as well in the course though, since the assignments were worth so
much. (Or were they? I never did quite figure out how the final marks in
that course were generated. Seemed kinda random.)
-Josh
Joshua Chud
> Rob Ewaschuk <raew...@fenchel.math.uwaterloo.ca> writes:
>
> Dare I suggest to you that on one of my exams you might not do as well
> as you think? I tend to ask problem-solving questions of the type that
> require practice. If you haven't done the assignments, you've
Interestingly enough, I've never actually found this to be true on exams
that you've set. In both of the CS courses in which I attended your
lectures, I found that the exams were relatively easy to get by with a
good deal of memorization - not of equations, definition and formulas, but
of entire assignment solutions. Your exam questions tend to be very
similar, at least in spirit to the mid-difficulty questions on your
assignments. (Usually requiring only slight tweaking of a solution
memorized from the published solutions or from lecture examples.)
> >> 100% assignments: really encourages copying on assignments
> >
> >Think so? First, how many people do you think would actually take this
> >route?
>
> Lots. People loathe exams. Assignments can be done on one's own time,
> in one's own room... and, as I've said, cheating is hard to detect.
Yeah, but some people loathe assignments. Then there's those pesky
take-home exams - worst of both worlds. (For me at least - all the
difficulty of an assignment with none of the time-pressure that I thrive
on during exams.)
-Josh
Rob Ewaschuk
> Rob Ewaschuk <raew...@fenchel.math.uwaterloo.ca> writes:
>
> > Which students? All students? Do all students need to get the work done?
> > Would the cramming really get any worse?
>
> Not all students need to get the work done. Most do. A smaller number
> will realize on their own what they need to do and get it done during
> the term instead of in the last few weeks. So if you remove term work
> incentives, yes, there will be more cramming.
I'm not sure about this, but I have no arguments, so I'll ignore it until
I do.
You didn't answer one of my questions, though. I was not initially asking
which students needed to their work, but which students couldn't be
trusted to do the work they needed to do. This is {all
students} \ {students that don't need to do work} \ {students that can be
trusted to do their work.} I don't know how large that is.
> So what is the penalty if they don't do the work? If there's a
> penalty, then you are basically assigning marks for assignment work,
> you're just moving the zero point.
I can think of three penalty schemes:
- Failure. Probably excessive and inappropriate.
- Dropping them out of their chosen marking scheme, so their assignments
now count. (They can game this by collaborating fully on early
assignments, then missing some critical assignment so that the earlier
assignments which were collaborated-on to a degree that would otherwise
have been unacceptable are now counted. However, if this was ever
suspected, it's certainly a case of academic dishonesty.)
- The loss of the marks for that assignment. This makes the scheme
equivalent to "assignments as pass/fail, but you can collaborate as much
as you want.)
> > > But why should anyone opt for a 100%
> > > final? Currently, they can get some fraction of their marks by getting
> > > away with cheating, which appears pretty easy. With the new scheme,
> >
> > This thought pattern is perturbing. I think it's showing that you've
> > institutionalized the exact adversarialism that frustrates me so much. I
> > would not cheat. I don't care how easy or hard it is. I play by the
> > rules, and many other students will too. I even play by the rules when I
> > think they're dumb, but I try to get verbose about it when I can. :-)
>
> Rob, I didn't institutionalize anything. I'm just telling you my
> working model based on experience with about twenty years' worth of
> students, not all of whom are like you. Of course you play by the
> rules. You do a lot of things around here because you think it's
> right, not because you're being forced. But I can't plan for the ideal
> student. I have to plan for the typical student, and in some cases for
> the worst student. The typical student is decent, but if they see a
> loophole being exploited by others, they'll go for it. Remember the
> first piece of trash on the ground argument?
No. But you asked "Why should anyone?" when there are clear reasons.
(Perhaps it should be scratched up to newsgroup communication fuzziness.)
> > I would choose 100% final because I don't like doing assignments. I don't
> > find them to be a good use of my time, either academically or marks-wise.
> > I tend to pick off the easy questions to get the marks for my time, and
> > look at the hard questions and think about them to get the learning for my
> > time. I seldom get more than 60% on my assignment mark, and I generally
> > do much better than that on my finals. I learn from lectures, not
> > textbooks or assignments.
>
> Dare I suggest to you that on one of my exams you might not do as well
> as you think? I tend to ask problem-solving questions of the type that
You can dare. You might be right. I think most profs would say the same
thing (that they try to ask these sorts of problem-solving questions, but
by reputation I'd be more likely to take your word for it than some
others.)
> require practice. If you haven't done the assignments, you've
> understood the problems solved in class, but you haven't done any on
> your own. You may be blessed with a high degree of self-awareness
> regarding your limitations and how to overcome them, but again I don't
> think that's typical.
I think I have a stronger-than-average ability to generalize problem
solving, and a weaker-than-average ability to memorize anything. This
leads me to do better-than-average on exactly these types of exams.
> >Think so? First, how many people do you think would actually take this
> >route?
>
> Lots. People loathe exams. Assignments can be done on one's own time,
> in one's own room... and, as I've said, cheating is hard to detect.
Yes. However, I have suggested that they still be required to write and
pass the final exam. I'll hug on to that suggestion here.
> >If they were on their own assignment stream, it wouldn't be hard
> >to cross-check a lot of them. With a severe enough penalty I think people
> >would be deterred.
>
> As it is, each case we prosecute takes up way too much time and
> effort. It would only get worse if the penalty was more severe. Have
> you ever done marking for a course? I think if you had, you wouldn't
> be saying "it wouldn't be hard" about any of this.
I have indeed -- I was a tutor for CS130 in F00. I was definitely
unconvinced. I was quickly told it was infeasible, but never convinced.
I was definitely deflated about it, but it seemed like institutional
inertia more than true lack of time. (Incidentally, that was the least
time consuming co-op job I've had. The course staff were excellent and I
enjoyed working with them, but they seemed either unallowed or unable to
delegate work.)
> >Another factor is that if you allow people to choose, they just might take
> >some responsibility for their choices. Perhaps I'm being too idealistic.
>
> It's certainly no problem for me to offer a 100% final option. I'll
> call it the Rob Ewaschuk option, so they don't blame me when they fail
> the course. --PR
As long as my name gets another hit on google, I'm happy. If you're
serious, though, I'd be happy to hammer out some policy details here or in
email. I'll even explain it to your class if you like. :-)
Isaac Morland
> Recursive..Inductive. Been on a work term for too long. :-)
Actually, recursion and induction are just two words for the same thing,
with slightly different connotations. So I don't view confusing them as
a serious issue.
Howard Cheng
> I have indeed -- I was a tutor for CS130 in F00. I was definitely
> unconvinced. I was quickly told it was infeasible, but never convinced.
> I was definitely deflated about it, but it seemed like institutional
> inertia more than true lack of time. (Incidentally, that was the least
> time consuming co-op job I've had. The course staff were excellent and I
> enjoyed working with them, but they seemed either unallowed or unable to
> delegate work.)
Note that most graduate TAs are only supposed to work for 5 hours a
week. It is not a full-time (?) job that you had. I once spent two
days marking one question (about 150 students) to see if I could catch
any cheating cases. I found one definite case, and many that are not
so easy to tell (and I didn't have the time nor the inclination to
call them all in and see if they understood what they are writing).
If cheating was clear-cut, then maybe it's not as bad...one reason
why it is infeasible is that there are so many fuzzy cases.
Howard
---
Howard Cheng e-mail: hchc...@scg.math.uwaterloo.ca
University of Waterloo URL : http://www.scg.uwaterloo.ca/~hchcheng/
Computer Science Graduate Student (PhD)
Everything should be made as simple as possible, but not simpler.
- Albert Einstein
Prabhakar Ragde
> In both of the CS courses in which I attended your
> lectures, I found that the exams were relatively easy to get by with a
> good deal of memorization - not of equations, definition and formulas, but
> of entire assignment solutions. Your exam questions tend to be very
> similar, at least in spirit to the mid-difficulty questions on your
> assignments. (Usually requiring only slight tweaking of a solution
> memorized from the published solutions or from lecture examples.)
For 251 I can see this -- first of all, I wasn't the only instructor,
and second of all, the course is more about taxonomy, organization,
and modular components. For 341 I would agree with the "similar in
spirit" part but not with "slight tweaking of memorized solution". If
this were the case, why did people do so badly on the exam,
considering many of them were carrying in photoreduced solution sets
on their "cheat sheets"?
> Yeah, but some people loathe assignments. Then there's those pesky
> take-home exams - worst of both worlds. (For me at least - all the
> difficulty of an assignment with none of the time-pressure that I thrive
> on during exams.)
You and I are polar opposites, it seems -- I have enjoyed every
takehome exam I've sat for, and I've set one for my current graduate
Joshua Chud
> Joshua Chud <jec...@uwaterloo.ca> writes:
>
> > In both of the CS courses in which I attended your
> > lectures, I found that the exams were relatively easy to get by with a
> > good deal of memorization - not of equations, definition and formulas, but
> > of entire assignment solutions. Your exam questions tend to be very
> > similar, at least in spirit to the mid-difficulty questions on your
> > assignments. (Usually requiring only slight tweaking of a solution
> > memorized from the published solutions or from lecture examples.)
>
> For 251 I can see this -- first of all, I wasn't the only instructor,
> and second of all, the course is more about taxonomy, organization,
> and modular components. For 341 I would agree with the "similar in
> spirit" part but not with "slight tweaking of memorized solution". If
> this were the case, why did people do so badly on the exam,
> considering many of them were carrying in photoreduced solution sets
> on their "cheat sheets"?
Good question - I have no idea. That's pretty much all I had on mine, and
I used it extensively during the exam (the first time I've been allowed
one and actually used it).
> > Yeah, but some people loathe assignments. Then there's those pesky
> > take-home exams - worst of both worlds. (For me at least - all the
> > difficulty of an assignment with none of the time-pressure that I thrive
> > on during exams.)
>
> You and I are polar opposites, it seems -- I have enjoyed every
> takehome exam I've sat for, and I've set one for my current graduate
> course. --PR
Maybe not quite polar - I've only sat for one (ignoring a couple in high
school), Jeff Shallit's CS462 final last term. I _did_ quite enjoy it, I
just didn't do very well, which I partly/mostly attribute to a lack of the
immediacy of the motivation experienced on timed exams.
-Josh
Prabhakar Ragde
> I was not initially asking
> which students needed to their work, but which students couldn't be
> trusted to do the work they needed to do. This is {all
> students} \ {students that don't need to do work} \ {students that can be
> trusted to do their work.} I don't know how large that is.
It is not this simple, because the extent to which students avoid work
that they are not directly rewarded for with marks and the extent to
which they suffer (for simplicity, on their final grade) by not doing
that work are both variable. So those aren't really sets we're talking
about, but rather a penalty function for each student. If you want to
know the number of students pushed below a passing grade by this
function, the one data point we have (from Therese and Dan's 341
class) suggests 5-10%, though there are any number of alternate
interpretations.
> > Dare I suggest to you that on one of my exams you might not do as well
> > as you think? I tend to ask problem-solving questions of the type that
>
> You can dare. You might be right. I think most profs would say the same
> thing (that they try to ask these sorts of problem-solving questions, but
> by reputation I'd be more likely to take your word for it than some
> others.)
Point taken -- also, I tend to ask fewer of these questions than I
would like, because I know most people won't be able to do well on
them.
> I have indeed -- I was a tutor for CS130 in F00. I was definitely
> unconvinced. I was quickly told it was infeasible, but never convinced.
> I was definitely deflated about it, but it seemed like institutional
> inertia more than true lack of time.
It could also be fear of what one might find out. I repeat the
statistic I heard, that on the one CS 130 assignment subject to
similarity-detection software, 30% of the class was flagged. You can't
prosecute that many cases, even if they were cheating. --PR
Prabhakar Ragde
> > It's certainly no problem for me to offer a 100% final option. I'll
> > call it the Rob Ewaschuk option, so they don't blame me when they fail
> > the course. --PR
>
> As long as my name gets another hit on google, I'm happy. If you're
> serious, though, I'd be happy to hammer out some policy details here or in
> email. I'll even explain it to your class if you like. :-)
You're on. Might as well hammer out the policy details in public, some
newsgroup readers have good ideas. But I'll move this thread to
uw.cs.ugrad. I think I set followups properly on this message. --PR
Matthew Nichols
> > any of the university's business whether or not students collaborate on
> > assignments.
>
> You are seeking a credential from the university that is for
> individual achievement and widely interpreted as such. The university
> has the right to set the rules under which you are granted that
> credential, as long as they do so in a consistent fashion. You have
> the right to refuse to earn that credential, but you don't have the
> right to not play by their rules and then demand it.
Wouldn't you agree, though, that there are limits to the university's
rules? An extreme example: Rule #1 - Only male students may earn degrees.
Would you support this rule under the same pretext (i.e. the university
can set any rules they wish as long as they are consistent)?
My real point is that restricting private conversations between students
oversteps the bounds of the university's authority.
For example (this is a little contrived, I admit), let's say you were
grocery shopping and two students in your class were walking in front of
you, discussing the latest assignment, and you overheard one of them give
the answer to one of the assignment questions to the other student. It
seems that university policy allows you, given this alone, to give both of
them -100% on the assignment.
> > Certainly *once a student hands in something* that is not his or her own
> > work, then they should receive no marks, since a student's grade in a
> > course is supposed to reflect the work he or she has personally
> > done.
>
> This contradicts what you said above. If students are to receive zero
> for work which is not their own, then the university has to be
> concerned about identifying these cases.
I don't feel it does. I'm drawing the line between students' conversations
versus their actually handing in work that is not their own. "Identifying
the cases" should be limited to the actual work submitted by the student.
./matt
Howard Cheng
> For example (this is a little contrived, I admit), let's say you were
> grocery shopping and two students in your class were walking in front of
> you, discussing the latest assignment, and you overheard one of them give
> the answer to one of the assignment questions to the other student. It
> seems that university policy allows you, given this alone, to give both of
> them -100% on the assignment.
This depends on the course policy. For CS360 it is common to say that
students are allowed to discuss solutions without taking any notes,
and must go away and write up solutions in their own words. Of
course, there is no way we can really check this.
> > > Certainly *once a student hands in something* that is not his or her own
> > > work, then they should receive no marks, since a student's grade in a
> > > course is supposed to reflect the work he or she has personally
> > > done.
> >
> > This contradicts what you said above. If students are to receive zero
> > for work which is not their own, then the university has to be
> > concerned about identifying these cases.
>
> I don't feel it does. I'm drawing the line between students' conversations
> versus their actually handing in work that is not their own. "Identifying
> the cases" should be limited to the actual work submitted by the student.
I may be wrong, but I think the "don't talk to your classmates" policy
is stated to minimize the "fuzzy cases" where it is not clear if the
students cheated or not. If there is enough resources, the best
approach, I think, is to allow you to do whatever you want, but the
assignment must be done individually and orally presented to the
instructor or the TA so there is no doubt that the students actually
understand the material (e.g. ask "why do you think this is a good way
to do it?") And generally the students learn more because of the
immediate feedback (and from my experience, many students don't even
bother picking up their assignments to see what comments they got).
Or maybe we can call students in randomly (or at least threaten them
to do so). But for a large class this can be a nightmare to organize
(schedule conflicts, etc.).
Howard
---
Howard Cheng e-mail: hchc...@scg.math.uwaterloo.ca
University of Waterloo URL : http://www.scg.uwaterloo.ca/~hchcheng/
Computer Science Graduate Student (PhD)
A proof tells us where to concentrate our doubts.
- Morris Kline
Prabhakar Ragde
> Wouldn't you agree, though, that there are limits to the university's
> rules? An extreme example: Rule #1 - Only male students may earn degrees.
> Would you support this rule under the same pretext (i.e. the university
> can set any rules they wish as long as they are consistent)?
No, this is not consistent, in that two different students are treated
differently without justification.
> My real point is that restricting private conversations between students
> oversteps the bounds of the university's authority.
If you insist, for ideological reasons, that the university not
differentiate between someone who does an assignment on their own and
hands it in, and someone who takes answers from another source, copies
it out, and hands it in, then you deny the possibility of giving
credit for assignments. The only hope for preserving that, if you
believe it beneficial, is to allow the university some leeway in
dictating the circumstances under which the assignment is to be done. --PR
Jesse Helmer
> It is my impression that copying assignments is more or less rampant
> in computer science. We face the dilemma that some skills cannot be
> learned without considerable practice. If we do not give credit for
> that practice, students will not do it, and then they do poorly on the
> final exam, which they typically must pass in order to pass the
> course.
A quick note: Senate's reaction to dramatic increases in cheating was
to require all instructors of all classes to append a written
description of what cheating is, etc, to the syllabus. Such a
description is currently appended to syllabi in the Faculty of Arts. I
voted against the motion.
Now, on to the more interesting discussion.
If marks are the only or primary incentive for students, then the
pressure to cheat is great. As marks become more and more important to
students (for whatever reason), the motive to cheat increases.
Reliance on marks as incentive to do work will encourage cheating.
Given the overall university system, it is very difficult (nearly
impossible, perhaps?) to do away with marks altogether. So let's
assume that marks are here to stay and that they will continue to be
an incentive to do work, to some variable extent and degree.
Marks-as-incentive could be minimized by the use of other incentives.
The first one that jumps out at me is -- long pause -- learning.
Marks will be primary for students concerned with the external (get a
scholarship, get a degree and get a job, get a good co-op job, and get
respect from one's peers, parents, and betters). Learning will be
primary for students concerned with the internal (self-improvement as
an end unto itself).
Of course, most students are concerned with both the external and the
internal. For now, consider only those who are primarily concerned
with the external (assuming that internally-focused students aren't
cheaters). If marks are the key to external recognition, then marks
are very important to these students -- perhaps so important that they
will cheat to obtain them.
So, for some students (who knows how many?), there is a strong
incentive to cheat. What, then, are the disincentives? Since these
students are focused on the external, it would seem that the most
important marks are final marks (not the mark on a particular
assignment or mid-term). So assignment marks are only important in
terms of their affect on the final mark. The best disincentives,
therefore, would seem to be those that significantly affect the final
mark. -100% for cheating is a very strong disincentive, but how does
it appear on a student's transcript? To go further than aggregate
marks, does it prevent someone from obtaining a degree?
It seems to me that marks-as-incentive to do work and to cheat are
here to stay. Moreover, we don't have a great deal of influence over
incoming students (other than deciding to admit them or not), and less
over society-at-large. We do have control over *some* of the
disincentives. I don't know what they all are, but I suspect that they
aren't strong enough because they aren't visible to external parties.
By making marks so important, we've created this problem for
ourselves. If the incentive to cheat increases, shouldn't the
disincentive to cheat increase as well?
Well, I've rambled on long enough.
One more thing: The incentive to cheat has increased greatly because a
university education is more and more often obtained, rather than
earned. It's obtained and used, like a backstage pass, to gain access
to previously unattainable places.
Regards,
Jesse
Anna Lubiw
Matthew Nichols <mjni...@maddison.math.uwaterloo.ca> wrote:
>I probably have a "unique" view on this topic, but I don't feel that it's
>any of the university's business whether or not students collaborate on
>assignments.
On the contrary, the university should foster collabortion
and the free exchange of ideas. It should be a place where
students learn from other students as well as from us "experts".
>Now, whether this collaboration can be done in a newsgroup hosted by the
>university is another issue. I'm simply objecting to the stance of some
>deparments/schools (specifically computer science) that students cannot
>freely discuss the assignment amongst themselves in whatever manner they
>choose.
Computer science has no such policy, to the best of my knowledge.
I don't impose such rules in my courses.
I think there is far too much emphasis on marks
here and at other universities I've seen. Marks as a scarce
resource. How to compete for marks. How to allocate marks
fairly. How to ensure that no one gets marks they shouldn't.
How to deal with those who want marks so badly they'll
cheat to get them.
Do we need to do this?
Personally, I like to spend a minimum amount of time trying
to make sure that marks are allocated fairly,
and then to spend the bulk of my time and energy on
people who care about ideas and about learning.
I would love to do what Bengt Aspvall did once in a course I
took from him at Cornell: in the first lecture, he
told us we'd all get A's; we could do the assignments if
we wanted to, and we could hand them in if we wanted his comments.
Anna Lubiw
Joshua Chud
> I would love to do what Bengt Aspvall did once in a course I
> took from him at Cornell: in the first lecture, he
> told us we'd all get A's; we could do the assignments if
> we wanted to, and we could hand them in if we wanted his comments.
Why don't you do that, then?
-Josh
Prabhakar Ragde
> My real point is that restricting private conversations between students
> oversteps the bounds of the university's authority.
Do you believe that the university should not restrict private
conversations between students during midterms and exams? --PR
Anna Lubiw
Because I'd rather sit in my office thinking about algorithms
and firing off the occasional wild comment to
uw.general than engage in subversive political action
for a hopeless cause.
Anna Lubiw
David Evans
Anna Lubiw <alu...@beta.math.uwaterloo.ca> wrote:
>Because I'd rather sit in my office thinking about algorithms
>and firing off the occasional wild comment to
>uw.general than engage in subversive political action
>for a hopeless cause.
>
You think that taking on all these tasks at once might be a bit odious?
--
David Evans (NeXTMail/MIME OK) dfe...@bbcr.uwaterloo.ca
Ph.D. Candidate, Computer/Synth Junkie http://bbcr.uwaterloo.ca/~dfevans/
University of Waterloo "Default is the value selected by the composer
Ontario, Canada overridden by your command." - Roland TR-707 Manual
Matthew Skala
Prabhakar Ragde <plr...@plg2.math> wrote:
>Matthew Nichols <mjni...@maddison.math.uwaterloo.ca> writes:
>
>> My real point is that restricting private conversations between students
>> oversteps the bounds of the university's authority.
>
>Do you believe that the university should not restrict private
>conversations between students during midterms and exams? --PR
I think that's different from the situation he described, in that midterms
and exams take place on University property. I myself think that
unauthorized collaboration is equally bad no matter where (physically)
it's conducted, but his claim seems to be that that makes a big
difference.
--
Matthew Skala, CS PhD student, University of Waterloo
msk...@math.uwaterloo.ca <-- school
msk...@ansuz.sooke.bc.ca <-- home
http://ansuz.sooke.bc.ca/ http://www.edifyingfellowship.org/
Matthew Nichols
>
> > My real point is that restricting private conversations between students
> > oversteps the bounds of the university's authority.
>
> Do you believe that the university should not restrict private
> conversations between students during midterms and exams? --PR
> --
> Prabhakar Ragde plr...@uwaterloo.ca
My argument would be that it's a minor restriction on students' freedoms
to not be able to talk for a few 2- or 3-hour periods during the term. On
the other hand, it's a major restriction to not allow them to discuss
assignments for the entire term (~ 3 months).
It's all about the balance of infringing on freedoms versus the ability of
the university to accurately assess students' abilities.
I also found Anna's comments about the overemphasis on marks to be very
refreshing. Ideally, we could change the system such that, rather than
manipulating it to maximize their marks, students would be focussed on
self-improvement and maximization of their learning.
I recall the phenomenon discusses in my psychology class whereby, once you
provide an external reward for an individuals' activities, then that
reward is what they become focussed on, rather than on the intrisic
rewards of the activities themselves. So, by making marks such an
important "reward" for individuals starting at such an early age, we are
simply reinforcing the notion that the rewards of learning are high marks
rather than self-improvement.
./matt
Joshua Chud
> I also found Anna's comments about the overemphasis on marks to be very
> refreshing. Ideally, we could change the system such that, rather than
> manipulating it to maximize their marks, students would be focussed on
> self-improvement and maximization of their learning.
Sure. The obvious extension of this line of thought is to move towards a
system where the only grading done in the program is through comprehensive
exams administered when a student feels they are ready to take them.
If you think the attrition rate is bad now, what do you think this would
do to it?
-Josh
Mike
Thu, 20 Mar 2003 10:38:05 -0500:
> On Thu, 20 Mar 2003, Matthew Nichols wrote:
>
>> I also found Anna's comments about the overemphasis on marks to be very
>> refreshing. Ideally, we could change the system such that, rather than
>> manipulating it to maximize their marks, students would be focussed on
>> self-improvement and maximization of their learning.
>
> Sure. The obvious extension of this line of thought is to move towards a
> system where the only grading done in the program is through comprehensive
> exams administered when a student feels they are ready to take them.
The obvious ultimate extension of this line of thought would be to eliminate
marks altogether.
Mike (staying out of the rest of it)
Joshua Chud
> Thus spake Joshua Chud <jec...@uwaterloo.ca>,
> >
> > Sure. The obvious extension of this line of thought is to move towards a
> > system where the only grading done in the program is through comprehensive
> > exams administered when a student feels they are ready to take them.
>
> The obvious ultimate extension of this line of thought would be to eliminate
> marks altogether.
That's what I meant - a set of comprehensive exams administered on a
pass/fail basis to determine if you'd earned your degree.
Unless you're proposing:
a) Not giving out degrees at all - just letting people come and hang out,
listen to lectures, etc. or,
b) Granting degrees to anyone who's registered as a student for four
years.
-Josh
Ray Butterworth
...
>I recall the phenomenon discusses in my psychology class whereby, once you
>provide an external reward for an individuals' activities, then that
>reward is what they become focussed on, rather than on the intrisic
>rewards of the activities themselves. So, by making marks such an
>important "reward" for individuals starting at such an early age, we are
>simply reinforcing the notion that the rewards of learning are high marks
>rather than self-improvement.
It's not only rewards; it happens with any method of measuring success.
Once there is a simple, easy to determine, objective measure
of something (regardless of the validity of that measure),
then optimizing that measure becomes people's goal,
independent of what that goal was supposed to measure.
And once such measures become known, they take on far more
importance and significance than they deserve.
In the case of the education system this is obvious,
high marks are the goal, with education being something
that tends to happen as a side effect.
But it happens in many other aspects of life too.
e.g. in sports such as fencing where judges were replaced
by electronic scoring, the goal of skewering one's opponent
was replaced by the goal of doing whatever it takes to
depress the sword-tip button in order to trigger the scoring
machine. This changed the style of the game quite considerably
(e.g. rather than traditional poking, one can now flick the
blade against the opponent in a way that causes the tip to
whip over depressing the button. In real sword play such an
attack would have been just silly, but now it is technique.)
Prabhakar Ragde
> I also found Anna's comments about the overemphasis on marks to be very
> refreshing. Ideally, we could change the system such that, rather than
> manipulating it to maximize their marks, students would be focussed on
> self-improvement and maximization of their learning.
Many students have that attitude now; certainly I did as a student,
and I continue to seek intrinsic rewards rather than extrinsic
ones. But as a professor, it seems to me that the system expects me to
assess marks in a meaningful fashion. I would like nothing better than
for this requirement to go away, but I can't unilaterally decide to
ignore it. Barring a complete change in the system, or at least some
feature that allows an asterisk to go next to a mark I assess with a
note about its lack of basis (and permission from the university to
operate in this fashion), it seems to me that it is my job to try to
provide a meaningful assessment while trying to limit the ways in
which the methods of measurement either permit manipulation or
motivate students to behave in a manner detrimental to their
education. If you want to talk principles, fine, both you and I can
agree that students should be free to do what they want and marks too
easily become the reward. But I am interested right now in practical
suggestions that I can implement. --PR
--
Prabhakar Ragde plr...@uwaterloo.ca
Alex Lopez-Ortiz
Ray Butterworth <rbutte...@mfcf.math.uwaterloo.ca> wrote:
>
>But it happens in many other aspects of life too.
>e.g. in sports such as fencing where judges were replaced
>by electronic scoring, the goal of skewering one's opponent
>was replaced by the goal of doing whatever it takes to
>depress the sword-tip button in order to trigger the scoring
>machine. This changed the style of the game quite considerably
>(e.g. rather than traditional poking, one can now flick the
>blade against the opponent in a way that causes the tip to
>whip over depressing the button. In real sword play such an
>attack would have been just silly, but now it is technique.)
I wonder.... if one were to point out this to a fencer who uses
this technique, would he say "touche'" ?
--
Alex Lopez-Ortiz alop...@uwaterloo.ca
http://db.uwaterloo.ca/~alopez-o Assistant Professor
School of Computer Science University of Waterloo
Mike
> On Thu, 20 Mar 2003, Mike wrote:
>
>> Thus spake Joshua Chud <jec...@uwaterloo.ca>,
>> >
>> > Sure. The obvious extension of this line of thought is to move towards a
>> > system where the only grading done in the program is through comprehensive
>> > exams administered when a student feels they are ready to take them.
>>
>> The obvious ultimate extension of this line of thought would be to eliminate
>> marks altogether.
>
> That's what I meant - a set of comprehensive exams administered on a
> pass/fail basis to determine if you'd earned your degree.
That makes sense, yes, a decent-enough compromise. Of course, under the
current system (at least at many universities, I don't know about the
specifics here) one can challenge for credit in a course, and I would expect
that one is generally given an exam. I doubt that universities would grant
degrees based solely (or even largely) on the basis of challenged-for
credits, however.
> Unless you're proposing:
I wasn't really proposing either, just throwing the idea out. I'm making a
mental note to engage brain before posting now; I didn't read between the
lines of your original post.
> b) Granting degrees to anyone who's registered as a student for four
> years.
Perish the thought! Some would argue that's where the current system is
heading though, I'm sure.
Mike
--
Our policy is, when in doubt, do the right thing.
- Roy L. Ash, ex-president, Litton Industries
Prabhakar Ragde
> > That's what I meant - a set of comprehensive exams administered on a
> > pass/fail basis to determine if you'd earned your degree.
>
> That makes sense, yes, a decent-enough compromise.
This is the traditional system used at Oxford and Cambridge. The
credit system evolved as a more humane and flexible alternative. --PR
--
Prabhakar Ragde plr...@uwaterloo.ca
David Evans
Prabhakar Ragde <plr...@plg2.math> wrote:
>Mike <mike.pa...@unb.ca> writes:
>
>> > That's what I meant - a set of comprehensive exams administered on a
>> > pass/fail basis to determine if you'd earned your degree.
>>
>> That makes sense, yes, a decent-enough compromise.
>
>This is the traditional system used at Oxford and Cambridge. The
>credit system evolved as a more humane and flexible alternative. --PR
Connected with this issue is the trend within some departments to eliminate
PhD breadth comprehensive exams.
Tomas Vinar
> On Thu, 20 Mar 2003, Mike wrote:
>
> > Thus spake Joshua Chud <jec...@uwaterloo.ca>,
> > >
> > > Sure. The obvious extension of this line of thought is to move towards a
> > > system where the only grading done in the program is through comprehensive
> > > exams administered when a student feels they are ready to take them.
> >
> > The obvious ultimate extension of this line of thought would be to eliminate
> > marks altogether.
>
> That's what I meant - a set of comprehensive exams administered on a
> pass/fail basis to determine if you'd earned your degree.
Well, I have been a subject to three such exams during my undergraduate
studies. (This is on top of all other obligatory final exams for all
courses which I have taken.) They were not much fun though reviewing
material in bigger context was useful.
Tomas
JA Grove
On 20 Mar 2003, Prabhakar Ragde wrote:
> Mike <mike.pa...@unb.ca> writes:
>
> > > That's what I meant - a set of comprehensive exams administered on a
> > > pass/fail basis to determine if you'd earned your degree.
> >
> > That makes sense, yes, a decent-enough compromise.
>
> This is the traditional system used at Oxford and Cambridge. The
> credit system evolved as a more humane and flexible alternative. --PR
Having been through the Oxford system, I can tell you that sitting all
your exams in two weeks at the end of the third year is not a very
pleasant experience. I might also point out that they are not pass/ fail;
in fact, the class of degree that goes on your certificate is determined
largely by your position - it is essentially a competitive system.
There are also a few other issues that arise from a purely finals-based
grading system. The most debated of these, in Oxford at any rate, is that
women in Oxford are consistently being awarded fewer firsts (and thirds in
fact). The system also puts people under enormous stress. I remember
seeing an article once showing that the number of attempted suicides at
Oxbridge are above the average for the UK.
One thing I think you could borrow is the idea of the viva voce (oral)
examination. You can be called for this at any point after your written examinations and it can be used to
over-rule the results of those written exams. Many departments use it to decide where the people on the
class boundaries should be placed. Although I never had one for this reason, I know it is difficult to cheat
or bluff your way through if you are being interviewed by several experts.
Prabhakar Ragde
> Having been through the Oxford system, I can tell you that sitting all
> your exams in two weeks at the end of the third year is not a very
> pleasant experience. I might also point out that they are not pass/ fail;
> in fact, the class of degree that goes on your certificate is determined
> largely by your position - it is essentially a competitive system.
But there are only four or five classes, right? Here we used to have
15 classes (A+, A, etc.) in some Faculties, but because of SISP/Quest,
everything has gone to 100 classes (100%, 99%, etc.).
> One thing I think you could borrow is the idea of the viva voce
> (oral) examination. You can be called for this at any point after
> your written examinations and it can be used to over-rule the
> results of those written exams. Many departments use it to decide
> where the people on the class boundaries should be placed. Although
> I never had one for this reason, I know it is difficult to cheat or
> bluff your way through if you are being interviewed by several
> experts.
There is something to be said for this, but objections are likely to
be resource-based. --PR
JA Grove
On 21 Mar 2003, Prabhakar Ragde wrote:
> JA Grove <jag...@engmail.uwaterloo.ca> writes:
>
> > Having been through the Oxford system, I can tell you that sitting all
> > your exams in two weeks at the end of the third year is not a very
> > pleasant experience. I might also point out that they are not pass/ fail;
> > in fact, the class of degree that goes on your certificate is determined
> > largely by your position - it is essentially a competitive system.
>
> But there are only four or five classes, right? Here we used to have
> 15 classes (A+, A, etc.) in some Faculties, but because of SISP/Quest,
> everything has gone to 100 classes (100%, 99%, etc.).
>
That's right - 1st, 2.1, 2.2, 3rd, pass, fail. Pass and fail don't really
get used though - it is pretty hard to be bad enough to get them and not
get thrown out somewhere along the way!
jason
Prabhakar Ragde
> That's right - 1st, 2.1, 2.2, 3rd, pass, fail. Pass and fail don't really
> get used though - it is pretty hard to be bad enough to get them and not
> get thrown out somewhere along the way!
So how is one thrown out earlier? Are there other assessment
opportunities along the way to discourage people who would not pass
the comprehensives? --PR
JA Grove
On 21 Mar 2003, Prabhakar Ragde wrote:
> JA Grove <jag...@engmail.uwaterloo.ca> writes:
>
> > That's right - 1st, 2.1, 2.2, 3rd, pass, fail. Pass and fail don't really
> > get used though - it is pretty hard to be bad enough to get them and not
> > get thrown out somewhere along the way!
>
> So how is one thrown out earlier? Are there other assessment
> opportunities along the way to discourage people who would not pass
> the comprehensives? --PR
There are exams at the end of the first year that you are required to
pass, if you fail there is an opportunity to retake during the vacation.
Fail that and the university will throw you out. In engineering, they lose
about 15% of the year this way, although I believe that is about the
highest of any department.
A lot of teaching is done in tutorials - one professor teaching two or
three undergraduates for an hour. I had two of these every week. If you
are not working they will spot it and there are college disciplinary
procedures which they can use to force you to work. The main mechanism is
to give someone "penal collections." Collections are beginning-of-term
examinations, set by your tutors and administered by the college. Someone
on penal collections is required to get a mark set by the college in
advance of the exam; if you fail, they can throw you out. They are
actually used relatively often, but few people actually get thrown out this
way. It sounds harsh, but you get plenty of warning and have to be doing
very little work in the first place for the college to use it.
As you said before, it is all very heavy on resources. The college system
means that Oxford gets two grants for every student off the government and
the university is pretty well endowed, so it can afford to do this.
jason
Prabhakar Ragde
> As you said before, it is all very heavy on resources. The college system
> means that Oxford gets two grants for every student off the government and
> the university is pretty well endowed, so it can afford to do this.
Thanks -- although little of this sort of approach can be implemented
here, it does give us a wider perspective on ways in which we can
tweak what we do. --PR
Isaac Morland
> JA Grove <jag...@engmail.uwaterloo.ca> writes:
>
>
>>Having been through the Oxford system, I can tell you that sitting all
>>your exams in two weeks at the end of the third year is not a very
>>pleasant experience. I might also point out that they are not pass/ fail;
>>in fact, the class of degree that goes on your certificate is determined
>>largely by your position - it is essentially a competitive system.
>>
>
> But there are only four or five classes, right? Here we used to have
> 15 classes (A+, A, etc.) in some Faculties, but because of SISP/Quest,
> everything has gone to 100 classes (100%, 99%, etc.).
This was not because of SISP/Quest. Those systems are perfectly able to
handle multiple grading schemes, and in fact they do so for historical
data. It was felt that while everything else was being re-arranged
would be a good time to make the grading system consistent across campus.
On a related point, I should point out that the reason for 100 levels is
not that there really is a big difference between an 83 student (for
example) and an 82 student. The reason is to avoid severe sampling
error, i.e. it is the same reason that one does not (if one knows what
one is doing, which many people do not) round intermediate results of
real-number calculations. Concrete example: imagine two students, one
who constantly gets 79s and another who typically gets 80s. Obviously,
they are almost equivalent in terms of their average marks. Under a
letter-grade system, the first student gets B+s and a B+ average. The
second gets A-s and an A- average. Or if you eliminate + and - marks,
Bs and As respectively. So a system with large quanta artificially
magnifies some small differences into large ones, and collapses some
larger differences into small ones (consider the above with one student
at the bottom of a letter and another at the top).
So the only statistically defensible way to grade a course, if something
more sophisticated than pass/fail is to be used, is to grade each
component of the course on some reasonable scale (say /10 or /20 for an
essay, or possibly marks out of 5 for various aspects, or whatever
according to the circumstance), and then compute an appropriate weighted
average without rounding. However, I do admit that rounding to the
nearest 1% is a reasonable concession to convenience in mark entry,
reporting, and comprehension.
Prabhakar Ragde
> On a related point, I should point out that the reason for 100 levels is
> not that there really is a big difference between an 83 student (for
> example) and an 82 student. The reason is to avoid severe sampling
> error, i.e. it is the same reason that one does not (if one knows what
> one is doing, which many people do not) round intermediate results of
> real-number calculations. Concrete example: imagine two students, one
> who constantly gets 79s and another who typically gets 80s. Obviously,
> they are almost equivalent in terms of their average marks. Under a
> letter-grade system, the first student gets B+s and a B+ average. The
> second gets A-s and an A- average.
This is only a problem if one mechanically rounds numerical marks,
which is not what one should do in computing letter grades. I have
assessed marks in an environment where letter grades were used, and
the way we assessed them was to look at the numerical marks and try to
determine natural boundaries. We then looked at the work of students
close to the boundaries and adjusted if necessary.
I also don't give numerical grades on essays in CS 492 when I teach it,
here.
> So the only statistically defensible way to grade a course [...]
I would hesitate to call anything we do statistically defensible,
because I don't believe that the measurement errors are random. --PR
Rob Ewaschuk
Jesse Helmer tried to post this via google.com, but it didn't make it
here. Presumably it managed to get trapped in that *other* uw.general.
From: Jesse Helmer (je...@helmer.ca)
Subject: Re: different viewpoints
Newsgroups: uw.general
Date: 2003-03-18 23:00:06 PST
Prabhakar Ragde <plr...@plg2.math> wrote in message
news:<x67kb2c...@plg2.math>...
> It is my impression that copying assignments is more or less rampant
> in computer science. We face the dilemma that some skills cannot be
> learned without considerable practice. If we do not give credit for
> that practice, students will not do it, and then they do poorly on the
> final exam, which they typically must pass in order to pass the
> course.
A quick note: Senate's reaction to dramatic increases in cheating was
to require all instructors of all classes to append a written
description of what cheating is, etc, to the syllabus. Such a
description is currently appended to syllabi in the Faculty of Arts. I
voted against the motion.
Now, on to the more interesting discussion.
If marks are the only or primary incentive for students, then the
pressure to cheat is great. As marks become more and more important to
students (for whatever reason), the motive to cheat increases.
Reliance on marks as incentive to do work will encourage cheating.
Given the overall university system, it is very difficult (nearly
impossible, perhaps?) to do away with marks altogether. So let's
assume that marks are here to stay and that they will continue to be
an incentive to do work, to some variable extent and degree.
Marks-as-incentive could be minimized by the use of other incentives.
The first one that jumps out at me is -- long pause -- learning.
Marks will be primary for students concerned with the external (get a
scholarship, get a degree and get a job, get a good co-op job, and get
respect from one's peers, parents, and betters). Learning will be
primary for students concerned with the internal (self-improvement as
an end unto itself).
Of course, most students are concerned with both the external and the
internal. For now, consider only those who are primarily concerned
with the external (assuming that internally-focused students aren't
cheaters). If marks are the key to external recognition, then marks
are very important to these students -- perhaps so important that they
will cheat to obtain them.
So, for some students (who knows how many?), there is a strong
incentive to cheat. What, then, are the disincentives? Since these
students are focused on the external, it would seem that the most
important marks are final marks (not the mark on a particular
assignment or mid-term). So assignment marks are only important in
terms of their affect on the final mark. The best disincentives,
therefore, would seem to be those that significantly affect the final
mark. -100% for cheating is a very strong disincentive, but how does
it appear on a student's transcript? To go further than aggregate
marks, does it prevent someone from obtaining a degree?
It seems to me that marks-as-incentive to do work and to cheat are
here to stay. Moreover, we don't have a great deal of influence over
incoming students (other than deciding to admit them or not), and less
over society-at-large. We do have control over *some* of the
disincentives. I don't know what they all are, but I suspect that they
aren't strong enough because they aren't visible to external parties.
By making marks so important, we've created this problem for
ourselves. If the incentive to cheat increases, shouldn't the
disincentive to cheat increase as well?
Well, I've rambled on long enough.
One more thing: The incentive to cheat has increased greatly because a
university education is more and more often obtained, rather than
earned. It's obtained and used, like a backstage pass, to gain access
to previously unattainable places.
Regards,
Jesse
--
: Rob Ewaschuk : University of Waterloo : Computer Science (4A) :
: www.studentforce.ca : Currently on co-op at www.decisionsoft.co.uk :
Joshua Chud
wrote:
> So, for some students (who knows how many?), there is a strong
> incentive to cheat. What, then, are the disincentives? Since these
> students are focused on the external, it would seem that the most
> important marks are final marks (not the mark on a particular
> assignment or mid-term). So assignment marks are only important in
> terms of their affect on the final mark. The best disincentives,
> therefore, would seem to be those that significantly affect the final
> mark. -100% for cheating is a very strong disincentive, but how does
> it appear on a student's transcript? To go further than aggregate
> marks, does it prevent someone from obtaining a degree?
I'm pretty sure that more that one or two of those -100%s on file will
cause a student to get put on probation or perhaps even booted from the
school? Someone who knows (Prabhakar?) want to back me up on this?
-Josh
Isaac Morland
A typical penalty for a second offence would be suspension for a term,
which at the least slows down the student's progress towards their
degree. I believe suspensions also appear on the transcript, which
simple mark penalties and disciplinary probation do not.
K Cheung
The motivation to learn is only one of the many variables.
McCabe and Trevino and many other authors have written articles
on this subject.
Now let me throw in my two cents. I think whether or not one takes a
course to learn or to get an extra line printed on the transcript, one
still has the choice to cheat. Consider the following situation.
Suppose a student named Bob thinks he knows everything in the course
already. But because of course requirements, he has to take that course.
He thinks that the assignments are a waste of his invaluable time.
But he doesn't think that he deserves to fail the course because he
knows he can pass if he spends time doing the assignments. So he
decides to copy the assignments, not because he can't do them,
but because he *knows* he can do them but he has something else
better to do. If you were Bob, would you have done the same thing?
Ultimately, I think to cheat or not to cheat is a moral choice.
Students who believe that cheating is morally wrong will tend not
cheat even if they are put in situations in which they will
certainly not be caught cheating (e.g. unproctored exams).
Students who do not believe cheating is wrong will be more inclined
to cheat when there is gain and they can get away with it or the
punishment for getting caught is lenient.
So how does one combat cheating? Many authors have offered many
suggestions. One thing that some authors claim to have worked very
well is to create a culture (or community) of honour among the
students by having an honour code. The honour code specifies
what is acceptable and unacceptable behaviour and students
make a pledge to the code. Violations are reported by peers.
In other words, the students keep one another in check. It's a
communal effort. There is more information and a comparative study
in McCabe, Trevino, Butterfield, Academic Integrity in Honor Code
and Non-Honor Code Environments. The Journal of Higher Eduation.
70:211-234 (1999).
One thing that many authors seem to agree on is that there should
be uniformity in the institution when it comes to cheating. I know
that U of Waterloo has the same set of policies. But I am not
sure if all the faculties enforce them in the same way or to the
same extent.
Kevin.
--
Kevin.
-- http://www.grad.math.uwaterloo.ca/~kkhcheun --
Prabhakar Ragde
> Consider the following situation.
> Suppose a student named Bob thinks he knows everything in the course
> already.
Is this the same Bob that someone else was going to go to to get the
answers?
> One thing that some authors claim to have worked very
> well is to create a culture (or community) of honour among the
> students by having an honour code.
I have in idle moments suggested an honour code here -- on the grounds
that students would be far harsher on transgressors than we faculty
are.
> There is more information and a comparative study
> in McCabe, Trevino, Butterfield, Academic Integrity in Honor Code
> and Non-Honor Code Environments. The Journal of Higher Eduation.
> 70:211-234 (1999).
A reference? To a printed journal? Can you do that?
--PR, who put references in his first child's birth announcement
Rob Ewaschuk
> was going to go to to get
Isn't english grand? It took my brain three passes to make sure that you
said what I knew you were trying to say.
> I have in idle moments suggested an honour code here -- on the grounds
> that students would be far harsher on transgressors than we faculty
> are.
I've thought about that. After I tutored, I spent a fair bit of time
considering the option of becoming a 'tattle-tale', but the pain of it in
UWs culture would have been quite bad. I think even many profs would look
down on it. People dislike the messenger more than the cheater in society
at large, and that carries through to UW in particular.
> A reference? To a printed journal? Can you do that?
Not in *my* backyard.
-Rob