Grading scheme which helps figuring out what is a right/wrong answer without generating frustrations

Question

I am in charge of the grading for an undergraduate class, and I am trying to determine the grading scheme which would best help students figure out what is a correct/incorrect answer without generating too much frustration.

My current position is to be demanding and give very little partial credit. For instance if the question asks to give a definition and the student gives an example, I would give no partial credit even if the example is a good one.

Strict grading is meant to make it clear when the student gives an inappropriate answer (e.g answering a different question from the one being asked, not providing enough of an argument to support an answer,...), and conversely when the student is answering correctly. For every graded piece, I write careful solutions and comments on how to answer properly. Strict grading is also meant to help the student figure out when she or he really needs to go look at the solutions and comments. Somehow, I fear that if students get partial credits more easily, they are more likely to believe that their answer is after all not that bad. Then they may be more likely not to check the solutions/comments, although the answer might be seriously flawed (e.g. they gave the right True/False answer but did not provide a correct justification of it).

In the class I am teaching, final grades are purely relative that is whether one gets an A or B solely depends on one's position in the grade distribution. So if they worry about final grades, student should not care too much about whether grading is strict, or whether partial credits are awarded, but rather whether the grading scheme is applied homogeneously between students. On this last point (making grading homogenous) I put a lot of effort (e.g. if no partial credit is awarded to a student answering "False" to a question without providing a proper counter-example, I make sure the same applies for every student answering "False" without providing a proper counter-example). I explained during class and repeated in all comments note that giving 0 to everyone or 5 to everyone was innocuous in terms of the final grade. I put the same effort in trying to make the purpose of strict grading clear.

However, I realize that there is more to getting a zero on a question than how it affects your final grade. My impression is that some students get frustrated for not getting partial credits. Some students expect to get partial credit as soon as they write something sensible even if it does not answer the question. I want to make clear that I don't blame them for that at all. I understand that this has to do with the grading culture they have been exposed to, and that my grading might be at odd with this culture. What worries me is that such frustration may eventually have detrimental effects on their learning : instead of thinking

" Oh, I got a zero, given what the instructor explained about his grading scheme, that means I am probably not answering the question right. I know that other people also got zero for the same kind of mistakes, so I should not worry about the way it will affect my grade, but I should definitely go check the solutions and comments to figure out what is wrong with my answer, and eventually go to office hours if there is still some confusion"

some of them react in a much different way. For many potential reasons, getting a zero makes them frustrated about the question/test/instructor, to the point that they are less likely to try to figure out what went wrong with their answer. More generally, not receiving partial credit may create discouragement on their behalf which would definitely negatively affect their learning.

Now, I could easily (I think?) solve the frustration problem while maintaining a fair grading procedure by uniformly giving (much?) more partial credit. However, not knowing about the average students' psychology in the class, it is hard for me to figure out which effect will dominate:

1. Would the decrease in grading clarity overweight the decrease in frustration and eventually decrease the number of students who thoroughly compare their answer to the solutions/comments (and hopefully learn how to give better answers)?
2. Or is the frustration-effect so high that giving more partial credit will have a positive overall effect on their learning?

Because this will differ from one class/university culture/topic to another, I am not primarily interested in guesses about whether 1 or 2 would hold.

What I would be more interested in is getting your insight on grading policies which would potential get me the best of both world, that is making it clear when a question is not answered properly and motivate the student to understand why, while avoiding any kind of detrimental frustration effect.

Thanks in advance for your help,

Answer 1

This recent review on grading gives evidence that when students receive both a grade and comments, they do not absorb the comments and become discouraged rather than motivated: Schinske, J., & Tanner, K. (2014). Teaching More by Grading Less (or Differently). CBE-Life Sciences Education, 13(2), 159-166. Link: http://www.lifescied.org/content/13/2/159.full.pdf

If motivating students to answer questions more carefully is your primary goal (and it is a good one), I would suggest considering a "redo" policy rather than stern grading. And skip the thorough comments and develop a simple system. That will make the time spent similar on your end, while motivating students to improve because they get another chance to earn all the points.

Answer 2

A professor in the math department has a really nice HW policy: there are 2 deadlines for every HW (a week apart), and everything is graded twice with the highest grade taken. The grading itself is very rigorous, and there little partial credit. Because the grading is harsh, it is very efficient and assignments are turned back in 48 hours. This subtly encourages a lot of good behavior:

• students start at least a week early
• students need to be clear to get better feedback on mistakes (instead of trying to get partial credit through ambiguity)
• it helps to have typed work
• there is complete transparency with the grade
• it promotes a healthier attitude about mistakes
• student correct their misunderstandings instead of mounting a legal defense for partial credit
• never have to deal with late HW excuses/ extension requests

Since originally answering this question we have implemented something similar in a for a computer science class.

• Very harsh automated grading that gave very specific feedback about program mistakes, with no limit for the number of submissions
• 25% of assignments could be "resubmitted" on the last day of class with no penalty

Answer 3

I think the "best of both worlds" is the world of strict grading that you are currently in:

My current position is to be demanding and give very little partial credit. For instance if the question asks to give a definition and the student gives an example, I would give no partial credit even if the example is a good one.

Sounds okay to me. Either the student doesn't know that such an answer is inappropriate, or knows that it is but chooses to give it anyway. This is bad in either case.

Some students expect to get partial credit as soon as they write something sensible even if it does not answer the question. I want to make clear that I don't blame them for that at all. I understand that this has to do with the grading culture they have been exposed to, and that my grading might be at odd with this culture.

It's always good to understand students rather than blaming them. But this should not lead you to compromise the standards of your discipline.

For many potential reasons, getting a zero makes them frustrated about the question/test/instructor, to the point that they are less likely to try to figure out what went wrong with their answer. More generally, not receiving partial credit may create discouragement on their behalf which would definitely negatively affect their learning.

They'll get used to it. Trying to shield students from the experience of frustration is just passing the buck to the next instructor and ultimately making the problem worse.

Moreover, unwarranted partial credit itself negatively affects learning by causing students to focus on tricks to get points on problems they don't understand, rather than focusing on understanding problems and writing correct solutions.

In my opinion, the only downsides to strict grading of the kind you describe are:

1. you have to make your expectations clear,
2. you have to grade carefully, and
3. you have to make the grading scheme clear.

But it sounds like you understand this and are doing it already. So I think you're fine and you don't need to compromise.

If you really want to give more partial credit, you could build it into your exam questions: for example, offer some smaller number of points for an example instead of a definition/proof. This way you don't encourage bad habits.

Answer 4

A few concerns that jump out to me:

• don't neglect the semiotics of letter grades! "A" means "you put in the right amount of work to succeed in this class." "B" means "you might get a grade you're happy with, but with a little more work you could get an A." I don't think it's wise (no citation, sorry!) to let students flounder in the dark wondering if the lauding comments mean "try harder" or "about right," which of course differs based on what the students are shooting for.
• Be very, very careful with equitable comments. Say you expect a student to do poorly (how privilege-sensitive are you?). So your comments may say "you've come along way" which translates to a B- for a student for whom you had low expectations or an A+ and is ready to be a researcher with a glowing rec letter coming from you for the best.
• Don't let students think this is a competition. Curve grading works in large classes where the students' individualism aside, statistics applies. It works in smaller classes too, but the students may not realize if there's more than 7 or so students it's not competition-based.
• You have a bigger class right? If it really is fewer than 7 or so students, there's a chance you'll have enough outliers to bring your planned curve into question. Or say at 7 students it's not independently, and they, fairly, collaborate a lot since they're really enjoying the material as well as each other's company.

tl;dr: give them meaningful letter grades, about 1/3-2/3 of a grade lower than they'll probably get, assure that the grades are approximations but will be boosted by the curve, and follow through with that. Source: I had two professors do this in college in very different settings (small, elite math weeder v. honors lit intro) and it worked out.