Teaching: in praise of rubrics

I spent last Friday grading for my five-week summer class. It took about nine hours*, which wasn't that bad, considering that the main graded work consisted of papers.

I like making students write. It lets me see their thought processes, and helps me differentiate between the students who can repeat what they've heard and the students who think for themselves - something that I especially want to see from an upper-level gen ed class like this one. So I assign papers. But I usually end up regretting that later, and wish I could convince myself that multiple-choice exams were adequate for a class like this.

When I started out teaching, it took me a loooooong time to grade papers. I would read each paper once, and then put them aside for a bit, and then read them again, this time with a red pen in hand. And then I would start sorting them into stacks based on how good they were. And THEN I would finally decide what grades to assign. It wasn't very efficient, and although I think it was fair within each class, the students may have had a difficult time figuring out exactly what I wanted.

About nine years ago, I was introduced to the concept of grading rubrics. The basic idea behind a rubric is to set up some kind of consistent grading scheme ahead of time, with grades or point values or something assigned to various aspects of the assignment. When I first started using them, the goal was to allow several faculty members to be consistent in grading their lab sections. But the rubrics we used were pretty complicated - there were long lists of things that we would give a few points to, and then we would add up all the points at the end. I found it hard to assign points to so many different things, and it was hard to make the assigned points agree with the grade I wanted to assign to the work.

My approach to rubrics has changed through time, mostly influenced by some of the examples I've gotten from people at SERC. For this set of papers, I only had two parts to my grading criteria: content, and style. (I separated them because I had learned that some students do critical thinking but haven't mastered the use of spellcheck, and others write beautifully but don't go very deep into the subject. I wanted to be able to acknowledge both combinations.) For each criterion, I described the typical characteristics of papers that would receive various grades. As an example, here is the rubric for grading the content of the final papers:

Content (both presentation and paper)

4........Meets all content requirements**. Great explanations, and logical and well-defended arguments. Great analysis of past actions (or inactions) to mitigate hazards. (Equivalent grade: A)

3........Meets most of content requirements. Good explanations and arguments. Good discussion of past attempts to mitigate hazards, but analysis could be better. (Equivalent grade: B)

2........Some problems with content. Location not clear, hazard needs better explanation, discussion of past actions are incomplete, or logic is difficult to follow. (Equivalent grade: C)

1........Major problems with content. Major factual errors or flaws in logic, or missing required content. (Equivalent grade: D

0........Assignment not completed.

I based the rubric on past experiences with this assignment - I had a long list of expectations for the paper, and I knew which expectations were typically difficult to meet. In this case, I used numbers rather than letters to make it easier to combine the two grading criteria in unequal proportions (75% content and 25% style). For my writing class, which has only one set of criteria, I only use letter grades.

Probably the most useful thing about these rubrics for students was that they were in the syllabus (which was available online, if students lost their original). The rubric went beyond the description of the assignment to tell students how I valued the different things that I told them to do. And I found that, for the most part, students did what they were asked - more so than in previous years.

And for me? Well, as long as the rubric reflects my real expectations for the paper, it helps me assign grades quickly. I know the criteria, and as soon as I've read the paper once, I can tell where it lies on the scale. I don't need to read all the papers, and then think about the grading criteria, and then read them again. I can forget what was in the first three papers I read when I'm on number 23, and I don't need to worry that my grading is getting easier or harsher depending on how much caffeine I've ingested.

There are downsides to rubrics. The biggest one, for me, is when a paper does something unexpected. In many cases, the unexpected things are fantastic, and I can grade them accordingly. But sometimes papers are great in some ways but lacking something important. Rubrics can commit me to criteria that don't fit every paper perfectly. So they're difficult for new assignments. (I got around that problem for my first two papers, which were new assignments this year, by making the criteria fairly generic. That meant that I took longer to grade the papers - I read each one and made notes to myself about it, and then went back to the papers and decided how the papers fit my grading scale. But I didn't read them all multiple times, at least.)

But for assignments that I have used many times before? Well, this year, at least, they kept me sane.

*That's grading time for final papers. There were other papers earlier in the term.

**There was a separate page explaining the requirements for the assignment.

More like this

I remember a discussion back in 1994 - during a quiet interlude in physics class at school - in which the teacher mentioned that the American educational system is very different from ours in that it makes much use of multiple-choice questions in examinations and very little use of questions answered in the form of an essay or paragraph. I'm convinced that our way is better, but I've met people educated under the American system who are convinced of the opposite.

I just rummaged through a drawer and fished out a copy of the First Year geology exam that I did in 1996 (I didn't take geology further than first year) to see what the essay questions were. For your interest, and for comparison with the expectations placed on First Year students in other countries, here they are:

__EITHER__ (a) Describe what is implied by (a) the relative geological time scale and (b) the absolute geological time scale. Outline in a general way how the relative time scale is calibrated in numerical years.
__OR__ (b) Regional metaphoric rocks cover vast areas of the Earth's crust. Discuss the mineralogical and textural features that may be used to distinguish between these various regional metamorphics and to determine their geological history.

__EITHER__ (a) Global climate change is one of the main causes of changes in global sea level. Discuss the possible causes of changes in global climate, giving evidence of major changes in the past.
__OR__ (b) Discuss the processes of plate tectonics and show how rock deformation, igneous activity and metamorphism occur as parts of those processes.

As I've commented here before, the key is to consider where you want students to get to before you start teaching- more specifically, what you want students to be able to do or demonstrate. Rubrics are the key to looking as objectively as possible at higher order thinking skills. Multiple choice and the like are necessary to evaluate what material is known, but to find out how the students are using what they know, you need to have more complex tasks. Because we have our own preferences, outlining evaluation criteria in a general way, making those criteria clear to learners, then sticking to them and applying them honestly, allows teachers to "grade" consistently and fairly. Good post.