Students are different in their abilities when they come to the classroom (see the curve above). However, we can still hear to this day talks among faculty of teaching on a curve in the academic halls … and this is a problem!
Teaching on a curve is the belief that students’ grades need to fall on a normal (bell) curve for each assignment; otherwise, something is wrong with the assignment (too hard or too easy). There is a huge caveat to this statement, a pitfall not easily avoided when one goes down the curving grades path the wrong way: This system assumes that teachers failed their students and that students, on average, do not learn in the classroom.
What I mean is quite straightforward, if as instructors we insist on our students’ grades to follow a bell curve every time, we basically admit that few students will perform poorly and few students will perform highly on every test. I don’t for you, but to me, this sounds like admitting teaching and learning doesn’t matter. More on this later. I admit that, in this scheme, some students will change categories along the semester between poor, average, and high performer. However, insisting on having the class grades falling on a bell curve is admitting that we will always have fewer students at the extremes. Seeing the class average increase over the course of the semester can be a sign of learning in the classroom, but it does not shake off these awful standards that every classroom is doomed to have poorly performing students and only a few high performing students!
Now, let me criticize my own opening statement when I assume that every classroom will start with students of different initial abilities … is that even true? Maybe, if your classroom is very large (>300 students at least I’d say). Otherwise, there will be too much variation in the small sample of students choosing your class and that assumption will not be valid anymore. Indeed, where I teach I have 40 kids per classroom, it is not unheard of to have better and worst groups of students between semester, in fact, it is very likely!
What can we do then? Let’s focus on assessing students’ learning and progression in our classroom. Forget about class grades’ distribution … they only inform us about the students’ relative mastery of course concepts compared to other students in the course. Pretty much worthless if you as me. Actually, I’ve grown to stop looking at the class grades’ distribution at all. It does not matter. It is not informative in any meaningful way for me or my students. Instead, we could look at students’ grades plotted over the term for each student, one at a time. This way we would see if students progress in their mastery of course concepts in an absolute way. This is simply done by looking at students’ grades evolution over the semester not compared to the rest of the class but compared to their previous grades. Are my students doing better than they were doing before? This is the real question we need to ask ourselves and act upon its answer:
Maybe students did not engage in proper study techniques, let’s teach them how.
Maybe students did not progress because the assignments were not consistently assessing students’ learning or were too hard, let’s change and improve them.
Maybe all our students progressed because we were stellar teachers this time around, let’s congratulate them and us …
Or maybe our assessments were too easy, let’s re-evaluate them.
But wait, you’re telling us that looking at distribution curves is bad and yet your blog article title mentions that you sometimes curve your exams up?! Yes, and I think my word choice is what is confusing here. Curving up sounds like grading on a curve; however, I don’t curve up by looking at the students’ grades’ distribution. Confusing, I know. Let me explain it with better words from better people. I should say that I calibrate my exams as I recognize that I can and often do make mistakes when designing assessments. This is especially true in those trying times with online teaching and my crusade against proctoring exams forcing me to redesign them every semester (my exams are taken with an open book policy … a story for another blog post). The process of calibration is simply adjusting students’ grades based on the class performance on the exam. I look at each question afterward and assess if I need to give a point back (poor wording, not achieving what I wanted, etc.) or even boost everybody’s grades because the entire exam was a bloodbath. This is not unusual in my classes as I rarely (almost never in fact) ask knowledge- or comprehension-based questions (out of high school they already showed they can do rote memorization). I almost exclusively ask understanding level questions as I like to say (application- and analyzing-based questions) and I prepare my students accordingly (see the table below that I share with them repeatedly along the semester starting on day 1).
For Your Information: I took this word of “calibrating” grades from an excellent post from James Flynn back in 2013 (opens in a new window).
[plain text URL: https://thevarsity.ca/2013/12/02/instructors-argue-that-often-used-grade-calibration-is-not-the-same-as-prohibited-bell-curving/]
A slightly modified table from the awesome article by Alison Crowe, Clarissa Dirks, and Mary Pat Wenderoth in CBE-Life Sciences Education (Biology in Bloom: Implementing Bloom's Taxonomy to Enhance Student Learning in Biology, opens in a new window).
[plain text URL: https://www.lifescied.org/doi/full/10.1187/cbe.08-05-0024]
In conclusion, do not look at your class grades’ distribution, focus on assessing students’ learning and understanding. Focus also on students’ progression by providing them with challenging yet doable assignments. Nobody said it would be easy, it is not. What is done then is the real work, not simply relying on artifacts based on poor statistic understanding to do the teaching in our place. And that’s why I usually calibrate (curve up) my exams up while not teaching on a curve.