[Monte314]

[HIDE="Spoiler"]
To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.

[/HIDE]
I have to write a Midterm Examination for my statistics class. My class is supposed to be the Pons Asinorum(1) that weeds people out of the MBA Program. This makes me a "gatekeeper" of sorts, and I don't like it.

I teach because I love to teach... I don't want to have to evaluate peoples' performance. I certainly don't want to be the end of someone's dream, but sometimes that is how it goes. Doing poorly can be worse for students than just having to switch to a different program, because in graduate school, a "C" gets you dismissed from the University.

I hate writing exams. I hate giving exams. I hate grading exams. In fact, I hate giving grades at all.

I think of myself as my students' servant; I am there for them, not the other way around. My primary responsibility is doing my level best to make *every* student succeed in the course. I can't always make that happen, and I regard this as a personal failure.

I grade on performance relative to other students taking the course at that particular time. I do not have an a priori fixed grading scale; since I am a senior faculty member, I don't have to adhere to any particular grading distribution.

However, I do take grading seriously (as a necessary "evil"), and I try to do it in a principled way. For example, I don't give good grades for "trying hard".

----------------

If you were teaching, what would be your fundamental grading philosophy?


(1)
To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.

[bmallerd]

If only all professors had your wisdom and dedication...

As an undergraduate, I have strong opinions on how things should be graded. Although I doubt my views are shared by the majority of my peers.

I would develop an exam in a fashion similar to how the Celsius scale was developed, establish boundaries and then make even sized divisions along the way. A 100% means the complete mastery of all materials presented in class. This does not mean every tiny factoid must be memorized and regurgitated. Given that exams are time limited, it would only be fair to give questions that are at the limit of the presented material. Questions that require significant synthetic thought should be given during assignments rather than exams.

e.g from CS. After having covered different searching algorithms in class. A fair question would be to ask whether or not binary search is better than a randomized search on average and why. An unfair question would be to ask the student to come up with a brand new searching algorithm that performs in O(lg n).

On the other end of the spectrum, a bare pass of 50% should indicate at least rudimentary knowledge. If I had taught a class on probability, then I would expect a passing student to at least be able to explain the probabilities associated with rolling dice. I.e why is P(3 and 3) = 1/36 when P(2 and 5) = 1/18. To pass a student that doesn't understand the basic definitions of a field seems unfair.

The test would be designed to spread students over this spectrum of bare pass to complete mastery. Effort marks would play no role in this measurement.

[Monte314]

I don't give in-class exams anymore, for several reasons. For one, I think time-pressure makes the playing field uneven, making it difficult for some students to demonstrate what they know. Secondly, I teach advanced classes, and each problem takes some time to do. My customary practice is to cancel class the week of the Midterm so I'm not "double-dipping" on students' time.

So, all my exams are take homes, unlimited time, open book, open notes, open internet, use any tools you want (e.g., EXCEL, MATLAB, Python). Generally, I hand out the Midterm the first day of class so people can be working on it. The final exam is always an individual term project, generally assigned the second week of class. In mid-level classes, I choose the project; in advanced classes, the students choose their project. I have found that serious students get a lot more out of a project they choose themselves, because they care about the outcome, so put a lot more into it. Any project that demonstrates mastery of the course material is acceptable. All projects are presented by students to the class at the last class meeting.

[Warrior]

	Originally Posted by Monte314 To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.
	So, all my exams are take homes, unlimited time, open book, open notes, open internet, use any tools you want (e.g., EXCEL, MATLAB, Python).

Ouch. When I was a student, those were the worst. When one of my professors started with that statement, it was usually followed by "because none of those things are going to help you". We didn't get those kinds of tests very often, but when we did, they usually took many hours to complete. I don't recall any that were unlimited time. Most had to be turned in within a week and you usually needed about all of that time. Sounds like your philosophy might be a little different, though.

I never liked getting a grade relative to the other students in the class. I preferred to know what was expected out of the class and be graded on a scale that represented how much I demonstrated I knew relative to that objective. I didn't enjoy being graded on a curve. Although I usually did well on the curve, I never quite knew if I was learning what was needed for more advanced classes.

[Mogura]

Perhaps grades should reflect what you've learned and not what you know.

A diagnostic exam at the start of the course to establish a baseline for each student. And then subsequent exams would be compared with the baseline to track "progress".

It's a bit personalized and out-of-the box, so it may be too time-consuming for practicality. But that was the thought that popped into my head when I saw this thread.

Whatever the approach, you want to emphasize learning, and de-emphasize "the grade". When the grade becomes the "be-all, end-all", then learning is abandoned...

[BostonIan]

I'd make the final exam the entire grade. Lessons and lectures, mandatory tests and voluntary assignments along the way, but no score enters until the last test.

The final would be long, lasting days or weeks. 70% high-density concepts, the rest intended as an opportunity to display mastery of the subject. The final day would be an off-your-ass, 0-10 value extra credit for a problem so hard that they're not intended to get correct, submit all scratchwork. Since the scoring would be arbitrary, I'd reward and penalize based on whatever intuition I'd happen to be following at the moment.

Since we learn from mistakes, any fumbling approaching an end is beneficial, not something to be frightened of, no pressure until the final month. The system doesn't arbitrarily penalize those of temperaments that happen not to fit a linear, hand-fed structure, nor reward busy-work. The students approach the end individually, in the way best suited to them. All that matters is the result, at the end, after all the classes are done.

I might also experiment with tossing the planned lecture format. Rather, give the students a Big List of "stuff you need to know" on the first day, let them pore over it, then field the questions.

[Monte314]

@Mogura: Interesting approach. This is how we track small children, and it is helpful in deciding what they need to do next (as opposed to the one-size-fits-all standard approach to education). Of course, this kind of individualized instruction would be expensive.

@Bostonian: this is how the "real world" does it. It's hard to argue with "reality".

[Akzis]

Originally Posted by Mogura To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.

Perhaps grades should reflect what you've learned and not what you know. A diagnostic exam at the start of the course to establish a baseline for each student. And then subsequent exams would be compared with the baseline to track "progress". It's a bit personalized and out-of-the box, so it may be too time-consuming for practicality. But that was the thought that popped into my head when I saw this thread. Whatever the approach, you want to emphasize learning, and de-emphasize "the grade". When the grade becomes the "be-all, end-all", then learning is abandoned...

Easily abused.

I, too, find grading on a curve annoying. What you know doesn't change based on your peers. I think there should be an objective standard to what you grade against, but that's just me.

[lioness]

I am studying to be a teacher and worked as a TA directly assisting in teaching classes with an ESL professor this past year. The grading philosophy question is one that's always unpleasant to discuss for me; I've discussed it with former teachers of mine and professors that I've met through being a TA.

I guess what I've learned is that grading philosophies very much depend on teaching style - who the teacher is, what the class is like, what purpose the class serves. The most important thing beyond this is that it be solid and dependable - students should know what to expect from day one and have periodic updates on their grades (including how it was determined).

Ultimately, however, grading is what it is. A grade. It is the students job to do well to earn it; the best thing to do is to maintain a positive class with clear expectations that provides the necessary access to the information that needs to be learned, and to be available to help.

I'm sure I'm not saying anything new, but I always love when I hear from a professor like you that cares enough to consider and reconsider his grading philosophy... thank you! If only more professors did the same.

[ACRV]

Thanks for the welcome.

Get the students up to the board. Have THEM work example problems and encourage comments. Give points for working problems. Even if you only do it once a week, your students will be more engaged. Give points for class attendance.

As for tests, tell your students the test questions will be 4 of 20 or so homework problems. Then change two of the numbers in the problems you give. Best way to get them to do the work AND checks to see if they are only doing the work by rote.

[kittychris07]

Being the hard-core sensing type that I am, what would be most helpful for me (when taking a midterm or final exam) is:

a) Have an in-class, closed book part that tests knowledge of concepts and important facts. I often need to have an in-class part to hold me accountible for learning concepts and big facts. While solving problems out of class helps with learning material, having book and notes available does not make me memorize the important information, and sometimes I can get by not understanding important concepts in the material.
b) Have an out-of-class, open book part with problems to solve (after I've nailed the concepts and basic facts, these can help me figure out how to apply the knowledge I've acquired and think through things). In class tests should not have trivial questions on them though, EVER. (As a business major, I've seen too occasions where the questions are about really random, obscure facts).

Problem-solving needs to be an important part of any class, and coupled with in-class tests, I learn a lot.

[XFire35]

What frustrates me about the way 'education' is measured in the British system, is that it's not a measure of what you know, it's a measure of what you recall (arguably the mark scheme). I'd like an exam where it gives a fairly broad question, e.g.: "What do you understand of genetics and protein synthesis?" or some other rather vague question which has large space to answer. Marks would be awarded for quality or writing, depth and scope and linearity within the paragraph. I think it's unreasonable to expect to work in a linear fashion in an exam when they do not in real life. So as long as it makes sense and follows within the paragraph, that would be the only concern.

I wouldn't use letters, I would use numbers. 0 would be what is expected, and increase is better - exceeding expectation (maximum of 3). A negative number is worse, not meeting expectation (minimum of -3). The overall class grade would be a mean of the values, and if a letter is required, C would be 0.

[vampyroteuthis]

I don't teach in a strictly quantitative field, so i don't know if this is at all useful, but my usual approach is to be extremely consistent with my grading rubric, almost pedantically so, to avoid the possibility/allegations of preferential or unfair treatment, and then I do everything I can to help my students meet these criteria. I refuse to grade on a curve. I try to offer as many opportunities as possible in the way that assignments are structured, as well as being available for feedback/consultations outside of class early on in the semester.

For example, when I set a class requirement of 2 short papers, a presentation, and a final, I tell them at the outset that they have the option of doing a rewrite on one short paper after they get feedback (with grade revision if that's warranted); presentations can be pure text or pure visual, as long as discussed in advance; and they have the option to submit an early draft of the final paper and get a brief paragraph of comments. All of these are contingent on them putting in some extra effort to make sure their work is up to scratch, so it's a way to accommodate effort without saying "here's a B for trying", as they're graded on the quality of the final work alone.

A colleague also just passed on a syllabus she heard of, where the professor specified that students had to fulfill requirements by points. She had a multiple-choice list of assignments, each with a point value. Basically you have to choose one 40-pt. assignment, two 20-pt., and two 10-pt. ones for a total of 100; the point value of each assignment reflects its weight in the final grade. Sounds complicated to design but it seems like it might better accommodate a variety of ways of learning.

[Gloed]

I'd grade them relative to all the students I'd have had. That said, I'd keep positively affirming them throughout the year. You can't expect someone to give their best, to want to grow, to get interested, without it. The grade-giving should pale in comparison to this. That way, one can decide for himself if he deems his grade enough to go through with it, because now he knows he gave his best, this is how he liked it and this is how he compares to others.

[Booko]

I taught Organic Chemistry, also famous as a "weed" class.

The thing is, there's Organic for chem majors, which is much easier than Organic for Pre-Meds and others. The Chem majors' class you spend the time learning a few mechanisms and how to apply them.

The other supposedly "easier" Organic class requires you to mindlessly learn a bunch of equations (80 per week easily) that you will promptly forget the moment you complete the class. It's supposed to be a Chemistry class, not an exercise in memorization. Pre-meds will have that later when they take Anatomy & Physiology.

If you will have to take Organic Chemistry and you haven't yet, I advise you to take the course the Chem majors take. It will be easier, you will learn something, and it's much more relaxed.

But I digress.

Organic Chemistry is a fairly objective academic subject, so I graded accordingly. It makes no sense to me to grade on a curve. You either know the material or you don't and the curriculum is well-established.

In addition, I graded on proper writing and spelling as well. That may seem odd for a Chemistry class, but proper spelling of chemical names is absolutely essential and organic chemicals have names that proves it was a field invented by Germans. :D Also, if you can't write your way out of a paper sack, no one is going to be able to decipher your lab notes, much less any journal articles you look to have published. Science is useless if you can't communicate your findings clearly.

Other subjects are not quite so well-established or as objective, so grading philosophy must be handled differently. DH taught English Lit, and there's much to be said there for grading on a curve. The entire grading process in such a subject is highly subjective.

I've also taught music theory. It's not entirely subjective, as there are things that are clearly "wrong" musically -- notation errors, parallel 5ths, insane harmonization, poor voice leading, that sort of thing. However, what makes a composition "good"? What makes a composition "creative"? That is highly subjective. I used to grade fairly objectively on the mechanical matters. Generally I gave homework that would demonstrate a mastery of the more objective matters.

The midterm and final consisted of: "Here's some paper and a pencil. Write a <fill in type of composition, such as "fugue">. You have 3 hours. The coffee is over there."

Now the finals I didn't exactly grade on a curve, but I did look at a curve to see if I was being too pessimistic or optimistic overall. I also wrote a lot of notes, so the student would realize the grade they ended up with was not something I just hauled out of a high colonic source.

For the midterm, because we had time to do so, the students would give their own opinions/grades, with no knowledge of whose composition they were looking at. The students' opinions were 50% of the grade and the other 50% was mine. Getting student involvement served a dual purpose -- they had to prove they could analyze music well enough to give an informed opinion and they could serve to balance me out where I might've been too subjective in my judgements of what is "good." It's great practice for them.

[Silence]

Booko, that's fascinating. With that style of teaching/grading, I can't imagine the students not wanting to participate as much as possible.

Monte, there's got to be a method of grading that will satisfy all protocols. I have to think on this a bit more before I answer.

[thod]

	I'd like an exam where it gives a fairly broad question, e.g.: "What do you understand of genetics and protein synthesis?" or some other rather vague question which has large space to answer. Marks would be awarded for quality or writing, depth and scope and linearity within the paragraph.

Clearly not a mathematician. The questions are objective and there is only one correct answer, no amount of well written prose will compensate for getting it wrong. The answers will consist of strings of Greek symbols, the only English text will be short notes explaining the reasoning behind a particular step. Instead of "tell us all you know about integration", one will be presented with a series of integrals. If you get them right, with the right method, you know about integration.

[BobG]

Originally Posted by Monte314 To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.

[HIDE="Spoiler"] To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts. [/HIDE] I have to write a Midterm Examination for my statistics class. My class is supposed to be the Pons Asinorum(1) that weeds people out of the MBA Program. This makes me a "gatekeeper" of sorts, and I don't like it. I teach because I love to teach... I don't want to have to evaluate peoples' performance. I certainly don't want to be the end of someone's dream, but sometimes that is how it goes. Doing poorly can be worse for students than just having to switch to a different program, because in graduate school, a "C" gets you dismissed from the University. I hate writing exams. I hate giving exams. I hate grading exams. In fact, I hate giving grades at all. I think of myself as my students' servant; I am there for them, not the other way around. My primary responsibility is doing my level best to make *every* student succeed in the course. I can't always make that happen, and I regard this as a personal failure. I grade on performance relative to other students taking the course at that particular time. I do not have an a priori fixed grading scale; since I am a senior faculty member, I don't have to adhere to any particular grading distribution. However, I do take grading seriously (as a necessary "evil"), and I try to do it in a principled way. For example, I don't give good grades for "trying hard". ---------------- If you were teaching, what would be your fundamental grading philosophy? (1) To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.

Two things:

1) I think grading on a curve is unfair. The best designed course should:

a) Determine what the student needs to know by the end of the course.
b) Figure out how you're going to test whether the student learned the material or not
c) Write you lesson plan and hopefully your lesson plan will focus on the things the student absolutely has to know

Grading on a curve:

a) Sweeps any mistakes the teacher made under the rug since his results always look 'average'. (i.e. - if only 10% of the class passes, the teacher should probably be trying to figure out why he's failing to convey the info necessary to pass the test).
b) Punishes at least some students even if both the teacher and all of the students did a good job on the course material. It may be unlikely, but it is possible that all of the students could get A's or B's on a good, objective test.
c) Can wind up punishing even students that receive a good grade if none of the students are learning the material very well. In fact, in the classes I've taken, the worst teachers tend to either grade on a curve or grade incredibly leniently, figuring students that get an 'A' won't complain until they're in a future class that assumes they know all of the material the 'lousy' teacher was supposed to cover in their class.

2) Being the gatekeeper sucks, but, if the gatekeeper courses are early in a program, they're really a lot more merciful than letting a student waste 2 or 3 years of their life and their money on a program they're never going to be capable of completing.

I remember a 'gatekeeper' course I took in electrical engineering. We started with 24 students and then, one week, I realized the entire back of the class was gone, as if they'd slowly been picked off by snipers. Seven of us completed the course, with 6 passing. This was the third time one of the passing students had taken the course. Having finally conquered his nemesis, he promptly changed his major. Even having finally passed, he realized there was no way he could go through 3 more years of that.

A lot of my philosophy on teaching is heavily influenced by the time I spent in the Air Force. While there's a lot of things one could say about the military, one thing I think they do do very well is teach people how to teach in a very efficient manner.

[taciturn]

Have you tried numbering the exams to be sure that as you grade, you aren't becoming harder or lighter on the students?

[Monte314]

Here is the grading *methodology* (not philosophy) I've come up with after doing this for 30+ years:

I grade all papers at one sitting. I grade everyone's problem number 1, then everyone's problem number 2, and so on. This minimizes the chances of inconsistency that might arise because I forget what I did on a paper I graded earlier. I do not look at student's names while grading papers.

1.) I begin by making a pass through every paper on the problem I'm grading, and determine the types of mistakes people are making; I sort the papers into "correctness" stacks based upon this initial assessment.

2.) I remove the papers that have full credit on that problem.

3.) I then go through the stack that is closest to correct, sorting it so that the "least incorrect" paper in that stack for that problem is on top. At this "detailed" stage, some papers might be moved to other stacks. This is the stage at which I make sure I haven't misunderstood the students' work.

4.) I proceed through the next most correct stack, and so on until all papers are in the right stack, and all stacks for that problem have been sorted "best" to "worst". At this stage, NO POINTS HAVE BEEN DEDUCTED yet.

5.) I then go back through each stack (starting with the most correct stack), and assign scores within each based upon the sort order in the stack. The highest score given in a "worse" stack is no higher than the lowest score given in a "better" stack.

In this way, all of the scores assigned on the problem have been determined in a manner that is consistent across the entire class. This is the key when you don't grade on the curve: the numeric point values you assign are much less important than doing things consistently, because it is not the final total that determines the grade, but the rank within the class.

6.) Once all problems have been graded, I total the scores for each paper. I sort the papers from Highest Score to Lowest Score. Based upon what I have seen on the papers, I assign a letter grade to the paper having the highest score (which is now on top of the pile), and a letter grade to the paper having the lowest score (which is now on the bottom of the pile). This is the point at which "instructor subjectivity" will have its greatest impact... as it should. Notice that this impact is spread across the entire class, rather than focused on any particular student.

7.) Finally, I assign letter grades to the rest of the papers according to their order in the pile, so that they fall between the "best" and "worst" grades. I never assign a grade in a way that violates the order of the final stack, and students having equal scores receive the same grade.

[Silence]

Originally Posted by Monte314 To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.

Here is the grading *methodology* (not philosophy) I've come up with after doing this for 30+ years: I grade all papers at one sitting. I grade everyone's problem number 1, then everyone's problem number 2, and so on. This minimizes the chances of inconsistency that might arise because I forget what I did on a paper I graded earlier. I do not look at student's names while grading papers. 1.) I begin by making a pass through every paper on the problem I'm grading, and determine the types of mistakes people are making; I sort the papers into "correctness" stacks based upon this initial assessment. 2.) I remove the papers that have full credit on that problem. 3.) I then go through the stack that is closest to correct, sorting it so that the "least incorrect" paper in that stack for that problem is on top. At this "detailed" stage, some papers might be moved to other stacks. This is the stage at which I make sure I haven't misunderstood the students' work. 4.) I proceed through the next most correct stack, and so on until all papers are in the right stack, and all stacks for that problem have been sorted "best" to "worst". At this stage, NO POINTS HAVE BEEN DEDUCTED yet. 5.) I then go back through each stack (starting with the most correct stack), and assign scores within each based upon the sort order in the stack. The highest score given in a "worse" stack is no higher than the lowest score given in a "better" stack. In this way, all of the scores assigned on the problem have been determined in a manner that is consistent across the entire class. This is the key when you don't grade on the curve: the numeric point values you assign are much less important than doing things consistently, because it is not the final total that determines the grade, but the rank within the class. 6.) Once all problems have been graded, I total the scores for each paper. I sort the papers from Highest Score to Lowest Score. Based upon what I have seen on the papers, I assign a letter grade to the paper having the highest score (which is now on top of the pile), and a letter grade to the paper having the lowest score (which is now on the bottom of the pile). This is the point at which "instructor subjectivity" will have its greatest impact... as it should. Notice that this impact is spread across the entire class, rather than focused on any particular student. 7.) Finally, I assign letter grades to the rest of the papers according to their order in the pile, so that they fall between the "best" and "worst" grades. I never assign a grade in a way that violates the order of the final stack, and students having equal scores receive the same grade.

To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.


That's a lot of work...and sounds like a process that has been developed over years.

I was trying to think of a grading style that might be a bit less cut-and-dried...but the subject is math. I don't think math allows for fuzziness.

The procedure you use sounds like it works. It also sounds very tough on the professor.

[Monte314]

At this point, it has become mechanical; no more difficult than any other method, and it is formulaic enough that I rarely make clerical or logistic errors.

[roninpro]

	Originally Posted by Silence To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.
	I was trying to think of a grading style that might be a bit less cut-and-dried...but the subject is math. I don't think math allows for fuzziness.

It depends on what you mean by fuzziness!

[BobG]

	Originally Posted by Monte314 To view links or images in this forum your post count must be 2 or greater. You currently have 0 posts.
	Here is the grading *methodology* (not philosophy) I've come up with after doing this for 30+ years:

I agree it's a methodology rather than a philosophy. While it's very good at ranking students, you made no mention of how your method meets an objective, or even what the objective of your grading is.

In other words, your courses have some objectives that a successful student should meet. Do the tests prove that the student met those objectives (if they don't, then the grades are irrelevant)? And does the grade reflect how well the students met those objectives?

In general, I've noticed math and/or science professors aren't very good at writing or grading tests. The harder it is to find someone that can understand the material being taught, the less a school cares about some of finer details of teaching. So, your method is probably better than most in your field.

In fact, in discussions on other forums, there's a lot of variety of opinion on what should be on a test. Some like to put a problem that tests material that was never covered in the course, but could be solved if a person took everything they learned in the course one step further (that's a good idea, but should probably be done on a lab, project, or homework assignment; not a test). Some like to create incredibly hard tests, but grade really easy. Like I said, testing in general seems to be something that's just all over the board with college professors.

See how the professors teaching management courses do it. Of course, that might be incredibly disillusioning if they're teaching TQM, yet have the same problems measuring performance that everyone else does.

[themuzicman]

Obviously testing and grading is going to vary according to the material given.

I do like the idea of a take home mid-term and final. But I would be a stickler for demonstrating proper application of the course material. If there are people who are simply not cut out for a particular area (some people just suck at math), or are unwilling to put in the work to grasp the material, then they should fail or at least fall from the better grades. I really would not care so much what percent of the class passes or does not pass (beyond what affects my employment, that is...)