More on KISS and Tell Grading
I’ve now implemented my ideas about a modified version of Specifications-based grading and Standards-Based grading, as I discussed in an earlier blog posting.
I also have a much better understanding of the economics of it. This blog posting is a note about how it works.
A reminder of the problem
First a reminder of the basic issue: I believe that students tend to seize up from traditional grading mechanisms that rate them from Excellent to Poor. At the University of Lethbridge, this method causes students to become intellectually defensive and conservative to an extend I consider an academic offence: they are so scared of bad grades that they would rather do what they think you want than what they think is interesting, correct, or advances knowledge—an approach that I find difficult to distinguish from plagiarism in terms of its effect on the advancement of knowledge.
There is at least some research that shows that students do better work when they are graded pass/fail.
At the same time, however, a system that fails completely to recognise qualitative exceptional work is also not doing anybody favours: it is important for formative reasons if nothing else for students to know when they are done extremely well or less well. And there is something dissatisfying—and also probably pedagogically dangerous—about a grading system that over-rewards careless, poor, inappropriate, or improperly digested work.
I have over the last few years begun using pass/fail mechanisms, bonus marks, and a distinction between summative and formative grading to try and de-stress things for my students. This has lead to great success—my attendance is way up and we just discovered that students are writing (on average) more than 2x as much as required for each year level as a result. But it has also suffered from a structural danger: it is too easy for people to game the system and get much better grades than their summative evaluation suggests they should have. I.e. too much pass/fail, especially if “pass” is set at 50% and bonus marks are available for doing additional work ends up rewarding quantity far too heavily over quality.
Standards-Based and Specification-based grading.
Specifications and Standards-based grading systems help answer some of these questions. Both of them address the issue of pass/fail being too easy (and, more importantly, failing to point out poor work to students) by setting the “standard” for passing at something other than 50%. I.e. in both you set out a list of requirements and students pass when they show mastery of those requirements—perhaps at a C or higher level, for example.
Specifications grading goes farther by setting out minimum amounts of work for a “decent” grade and providing specific additional work students who want a higher grade can do. So in this system, a student doesn’t pass unless they show mastery of the minimum skills required for the course; and they don’t get an excellent grade if they don’t do additional set exercises to this same minimum standard.
What neither of these systems seem to answer, however, is the problem of qualitative excellence. Because they are pass-fail systems, there is no way of recognising qualitative excellence in individual exercises. In Specifications grading, you can get an excellent score, but you do this by doing more work rather than work to a higher standard: as long as you hit the minimum standard in everything you do, your “excellence” is determined by how much you do.
How I’m addressing this.
My big idea is to address this issue of qualitative excellence by adding badges into the mix. In my system, work is assessed as in Specifications and Standards-based grading on the basis of mastery: if you show you can do the exercise, you get a pass or 100%; if you don’t, you either get 0% or take another shot at earning your pass (minus a small penalty).
But unlike Specifications or Standards-based grading, I also reserve a part of the course grade for badges or tokens that recognise excellence in any particular assignment. This semester, for example, excellence in a formative assignment will earn a badge worth 2.5% of the course grade.
An example
Here’s an example of what I mean:
Formative exercises
- First essay (15%)
- Second essay (15%)
- Mid-term exam (15%)
Summative exercises
- Final essay (20%)
- Final Exam (25%)
Badges
- Excellence Badge (3.4%)
In this course, each formative exercise is worth 15% and if you do the exercise to an acceptable level, that’s what you get on a pass/fail basis.
But you also can do a little bit better than 15% if you do any of these to an “excellent” standard. In that case you get both the 15% for doing the work competently and another 3.4% (i.e. the badge) for doing it to such a high standard.
The economics behind this
What I like about this system is that it combines the best of traditional grading with the best of new forms like specifications grading. I.e. in practice, each assignment is actually being graded F/C+/A, allowing me to distinguish between qualitatively excellent work and simply appropriate material. But this is not obvious to the student since I am placing the A part under a different category.
