As we start up the new semester, this is an apt time to evaluate, and update or change, our grading schemes.
I don’t like giving grades. I wouldn’t assign grades if I didn’t have to, because grades typically are not a good measure of actual learning.
Over the least year, I’ve heard more about a new approach to assigning grades, that has a lot of appeal: “standards based grading,” in which students get grades based on how well they meet a detailed set of very clearly defined expectations. This is apparently a thing in K-12 education and now some university instructors are following suit.
I haven’t taught a course using standards-based grading – yet – but I’m really intrigued. I think there is the potential to remove a lot of the gamesmanship in college courses and increase the focus on learning actual knowledge and skills. It would make things more fair, more transparent, and provide students with motivation to accomplish rather than a motivation to survive.
At my university, we are expected to include our “Expected Learning Outcomes” prominently on our syllabi. And then our syllabi tell the students how they are graded. But in nearly all courses, including my own, the connection between expected learning outcomes and earned grade is, well, indirect.
Student get good grades by performing well on assignments and exams. Of course, we design these assignments and exams because they articulate with the information that we want students to learn.
But what if we gave our students As because they met a very detailed set of criteria that we set out in advance, then wouldn’t that be a win-win, as long as we can control what those criteria are?
How about we give a piece of paper to our students that tells them exactly what they need to get an A, B, C. D or F? And it can’t be just “answer exam questions well,” it needs to be far more specific than that. We can still administer exams, of course, but the items on the exams are specifically connected to the standards that are communicated in advance, so that the students know precisely what is expected of them before they take the exam.
This would clearly change not just how we assess performance but also how we design and teach our classes. What do you think?
(As it turns out, I’m not teaching any regular classes this semester, with time reassigned to research, mentorship and other projects. So I have several months to think about how I’m going to run my class next time. But I’m not planning to waste tenure, and I don’t mind experimenting.)
Here is related reading:
Beatty, Ian D. 2013. Standards-based grading in introductory university physics. Journal of the Scholarship of Teaching and Learning 13: 1-22.
I am a mathematician who does lots of inquiry-based instruction. I have dabbled with standards based grading concepts for a while now, and it does align better with thinking about actual learning. It takes a lot of time! The reassessments are a challenge to administrate. There are several others I know who are doing it, too. Most of what I have learned, I stole from HS teachers. But they have a very different schedule and pace.
For next term, I am trying out “Specifications Grading,” following a recent book by Linda Nilson. Maybe have a quick look at that. Many of the underlying motivations are the same (though Nilson’s book is often too negative about students for my taste), but I think specs grading might work better for a few kinds of college courses. At least, I hope so. My experiment begins next monday.
Here is what I write at the end of every syllabus. Most of my assignments address these skills and attitudes to learning. To become a good scientist you have to solve problems and find the answers yourself……
My ideal profile of the A student:
1) Thoughtfully reads all assigned readings (textbook, primary papers from the scientific literature, lab instructions)
2) Asks questions in class.
3) Looks beyond memorization to understanding unifying principles. Should be able to apply knowledge and concepts to novel problems
4) Stays on track with readings since course material is synergistic and past concepts are going to apply to future material
5) Pays attention to quantitative, data-rich figures and tables. Science is evidence based.
6) Every written assignment will have primary literature support. The more deep knowledge, the better the writing.
7) In summation: Engaged intellectually with the concepts of the course.
Andrea, those sounds like good traits we want to see in an A student,. If these were converted to a standards-based grading system, then I imagine it would look like this.
Asked a minimum of 10 questions in class over the semester: 100%. 5-10 questions: 50%. Less than 5: 20%
Applied knowledge and concepts to novel problems by completing an assignment in which this specific task was required.
Demonstrated ability to understand and interpret figures and tables by answering exam questions associated with this criterion.
… and so on.
But I think the standards are ideally tied to the course material more specifically. For an introductory cell/molec biology course, for example, I can imagine about 100 or so specific standards that need to be met, most of which deal with specific content, while some may have to do with reading, presentations, and so on.
At our university, Harper Adams, we have a 20 point marking scheme which sets out exactly how marks are given – students get to see this as well – so for example to get a mark between 13 and 15 out of 20 at MSc level, the student’s work must meet these criteria
“Assessment criteria set out in the assignment brief are satisfied to a high standard with no areas of weakness. Students display, as appropriate, clear evidence of evaluation, coherence, creativity, originality, autonomy and the ability to synthesize materials from other modules, from practical experience or from self guided study to complete the assessment task”
Standards based grading seems to have many good points but seems like a ton of work/grading. I am thinking about going with specifications-based grading (also mentioned by TJ), which seems to have many of the same goals as standard-based but involves less grading (or regrading).
Here’s a really useful post on specifications based grading:
http://chronicle.com/blognetwork/castingoutnines/2014/12/02/thinking-about-grading/
This is just the post that I was trying to relocate when writing this but I couldn’t think of where I saw it. Thanks! It indeed sounds like specifications-based grading sounds a lot more do-able.
I am trying this out this semester in an upper level Animal Behavior class. I am writing the syllabus and am finding that writing the syllabus is much more work than I expected. Also, I am discovering that it is necessary to know what every assignment is going to be if you are really going to specify what it takes to get an A, B, C, etc. I am also struggling with how to specify what the “creating” level of Anderson and Krathwohl (2000) would look like in different types of assignments (in ways that are understandable to my students without writing a whole page about it). I’ll probably write a few posts about this in the upcoming semester. Stay tuned.
I am very confused about what exactly is standards based assessment in the setting of a college and university. The idea of defining skills the students can achieve in a course is strange since aspects of problem solving, extracting information from many sources, hypothesizing “what is next”…and many more things that are involved in science and research don’t have building block instructions (like a math problem might in calculus or using a scientific instrument correctly).
I use rubrics for grading all writing assignments (that the students see before they do the assignment) that indicate what components go into a GREAT research report or review paper. Does that fall under standards based assessment?
I had more success with Standards Based Grading in Calculus and Linear Algebra, and much less in courses which are based around making arguments and testing out ideas. this is one of the reasons I am trying out Specifications Grading. Specs seems a better way to deal with situations where the natural assignments are complex projects with a professional nature (like finding a proof of a conjecture and writing a paper about it, or testing a potential definition).
Andrea, your rubric is naturally a set of specifications. It is probably aligned with a set of learning outcomes, like in standards based assessment, but less easily atomized than a “routine calc homework exercise.”