17

Being a PhD student, I just finished supervising my first masters thesis. Now I have to assess and grade it. While I have a grading template provided by the university (listing the different aspects of the thesis and points out of X for each aspects), I am finding it hard to estimate how good the single achievements are compared to, well, the general level of theses. If I give full marks now because everything has been done satisfactory and well, and the next thesis that I supervise is even better - I cannot give that one better marks, so it would feel like I misgraded the first thesis.

I imagine that once you have amassed a certain amount of supervised theses, you can estimate the general level of one thesis much better and probably feel more competent to grade fairly and correctly.

But in my situation, having no previous experience in grading, how can I make sure as best as possible that my grading is neither too lenient nor too harsh?

11
  • 7
    Do you have acceess to previous theses and their grades? It could give you at least a reference to compare against.
    – Louic
    Jan 5 at 9:33
  • 9
    Are you the only supervisor, or is there a more experienced member of staff involved (at least formally). If so, that would be the person to talk to. Jan 5 at 11:22
  • 2
    @PieterNaaijkens there is a second supervisor (a professor) that definitely has more experience, but if I rely on her too much, that would kind of make having to seperate supervisors futile, wouldn't it?
    – Sursula
    Jan 5 at 11:40
  • 26
    @Sursula I'm actually surprised that you are formally involved in grading a thesis. In my experience, that is not something PhD students are allowed to do. The rules at your institution are quite unusual in that regard. You don't "rely too much" on the professor if you ask for her help with grading. That's literally part of her duties.
    – Roland
    Jan 5 at 11:56
  • 9
    The first time I supervised and graded a Master thesis (I was also a PhD student at the time), the other supervisor and I both graded the thesis independently and then compared grades. It's a good way to learn while maintaining a level of independent judgement. In the end, both our grades were in the same 'band' (say, an A in the US system or a First in the UK), and diverged by only a few points.
    – Alaimo
    Jan 5 at 16:53
6

there is a second supervisor (a professor) that definitely has more experience, but if I rely on her too much, that would kind of make having to seperate supervisors futile, wouldn't it?

For many aspects of the jobs of educator and supervisor, you learn by doing. So, an appropriate course of action here would be to

  1. First, try to write an evaluation to the best of your ability following the guideline you were provided, and advice from people with more experience on the question.
  2. Secondly, you submit your work for feedback to a more experienced coworker. Here, the second supervisor seems appropriate, but another senior faculty you trust could be also helpful (they may not know the details of the thesis you are supervising, but they know the classic "beginner's mistakes").
  3. Finally, modify your evaluation according to their opinion. If you think you had to do a major revision of your evaluation, you can go back to point 2. for another round.

I was feeling uncomfortable at first when doing this kind of task, as things can be hard to evaluate but you get used to it through experience. For example, one rubric for the evaluation of student research projects in my institution is the quality of writing of their thesis out of 20 points. But the first thesis of even the best students is seldom a flawless work, so how harsh should I be? I could know the answer only by comparing my evaluation to other faculty members.

20

Grading should not be on a curve.

If I give full marks now because everything has been done satisfactory and well, and the next thesis that I supervise is even better - I cannot give that one better marks, so it would feel like I misgraded the first thesis.

This is incorrect. Full marks does not indicate that there is no room for improvement.

Use the grading template. If it's not detailed enough, unfortunately people unfamiliar with your degree program cannot help you. Ask a person who is familiar, such as the degree coordinator.

how can you guarantee fair grading?

For small sample sizes, you cannot.

7
  • 3
    A bit of a straw man argument there. OP never proposed grading on a curve, but only made the (uncontroversial, I would have thought) observation that, in a criterion-referenced grading system, the rank order of how well student submissions fulfill a criterion is a useful tool for assigning scores on that criterion. One can even do it across multiple cohorts: if student A from 2022 fulfills criterion X better than student B from 2015, but not as well as student C from 2017, then one knows to give A an X score higher than B's but lower than C's, much faster than assigning a score in a vacuum. Jan 5 at 21:57
  • 6
    @DanielHatton Anything that involves ranking students should be controversial. Jan 5 at 22:41
  • 3
    @AnonymousPhysicist So you are against grading altogether, as grading will always rank students? While I somewhat understand this viewpoint, it doesn't help with my problem.
    – Sursula
    Jan 6 at 8:13
  • 2
    @Sursula Just that small subsets cannot be expected to form nice predictable ranks. Grading is necessary still, but constricting the grading process to an outcome that produces ideal ranking would be doing a lot of labor for little fruit. Maybe a student gets a higher mark than the ideal ranking would imply, but that is an artifact of the subset and not a big problem. Grading someone more severely because there might come a better one next year would be a mistake. Or less severely on the inverse basis. Give the grade, worry about ranking when you have a decent set. Jan 6 at 10:08
  • 2
    @ AnonymousPhysicist the approach @DanielHatton describes isn't exactly ranking students. It's ranking specific aspects of their work, and only getting to that point by comparison to more-or-less objective criteria
    – Chris H
    Jan 6 at 11:09
13

You seem to have the misconception that grading is, somehow, competitive and that there is some linear scale on which you need to place student work in relation to other work.

I suggest that you work to give that up.

Every piece of work is unique and they don't compete with one another. Treat each as such.

Note that grading scales in some (not all) places are intentionally quite broad. Letter grades in the US, for example are quite broad categories. Strict numeric grades are a bit different, but it is probably incorrect to think of them as having true precision. What, after all, is the essential difference between 87 and 88? You are making judgments after all, not measuring.

And, is is true that over the years there is some likely "drift" in the meaning of grades. It is often lamented, but things seem to generally work out in any case. IIRC, Einstein was denied a teaching certificate upon graduation from (the predecessor of) ETHZ.

8

Talk to an experienced supervisor in your department

I would assume you at least have a second marker to moderate your inexperience. They will balance out your lack of calibration to some extent, but you can - and arguably should - go to your more experienced colleagues and seek their opinion and guidance. They may not be willing, or able, to read the whole thing but they should be able to spare five or ten minutes to help you calibrate your marks if you put the effort into explaining why you chose the mark you did.

2

This is the problem that grade descriptors are designed to help with; you may be fortunate enough, by judicious Googling, to be able to find a set of grade descriptors issued by the relevant programme team/university/accrediting body. One example is here.

2
  • 1
    I would have thought this should not be a question of being fortunate or hunting something down, but rather something the university should 100% be providing to students and staff.
    – dbmag9
    Jan 5 at 17:28
  • 2
    @dbmag9 "should" yes; "is" rarely. Jan 5 at 17:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.