How to determine course grades fairly (RF & RB)

Note: The material that follows is drawn from pp. 49–51 of Teaching and Learning STEM:
A Practical Guide

It’s headache time again.

The semester (or quarter or summer session) is over at last. You gave and graded your final exam and entered the grades in the spreadsheet, right next to the ones for the midterm exams and assignments and whatever else counts toward the final course grade. The spreadsheet instantly calculated the weighted-average numerical grade for each student, and you sorted the sheet to put the numbers in that column in descending order. You’re now looking glumly (no one likes grading) at a column of weighted-average numerical grades, with a student’s name next to each number. Your only remaining task is to put a letter—A, B, C, D, or F—next to each name, possibly (depending on where you teach) followed by plusses and minuses next to some of the letters. That may sound simple to a non-educator, but as all educators know, it’s anything but.

There are two ways to assign course grades: curving (or norm-referenced grading), in which the primary basis for a student’s letter grade is the ranking of her numerical (weighted-average) grade in the column, and absolute grading (criterion-referenced grading), in which the numerical grade itself is the primary basis.

Curving also comes in two flavors: (a) the top 10% of the weighted-average numerical grades get A’s, the next 25% get B’s, the next 35% get C’s the next 20% get D’s, and the bottom 10% get F’s (those percentages are just illustrative), and more commonly, (b) the numerical grades from the top of the column to the first moderately-sized gap between grades get A’s, those from that gap to the next one get B’s, and so on down to F’s. In contrast, the letter grades in absolute grading are determined entirely from the numerical grades. For example, the letter grades and their corresponding ranges of numerical grades might be A(90–100), B(80–89.9), C(70–79.9), D(60–69.9), and F(<60). If pluses and minuses are given, there would be a larger number of narrower ranges.

This blog post (more accurately, this series of linked posts) addresses three questions:

  1. Should I curve or not? (Spoiler alert: Not!)
  2. Suppose I use absolute grading and one of my students gets a weighted-average grade of 70 (which gets a C in the course) and another gets a 69.9 (which gets a D). I know the performances of those two students in the course are virtually identical. Do I still have to give them different course grades? (Spoiler #2: No!)
  3. Suppose I use absolute grading and I give a test on which most of the students get failing marks, bad enough to lower most of their course grades by one or two letters and to cause many of them to fail the course. Do I have to give them those grades? (Spoiler #3: No, especially if you decide the test wasn’t fair.)

OK, let’s look at the detailed responses to those questions. You can view them in any order you choose.

There are of course other important questions related to course grades, including these:

  • How much should I count midterm exam and quiz grades toward the final course grade?
  • Should I drop the lowest exam grade?
  • How much should I count the final exam grade?
  • How much should I count homework assignments? Should I count them less if students work on the homework in teams?
  • What else should count, and how much? Lab and project grades? Attendance? Class participation?
  • When and why should I give an incomplete?
  • What, if anything, should I do about seniors who failed my course and would graduate if they passed it?

We may devote blog posts to these questions in the future, but in the meanwhile you can find suggested answers on pp. 47–49 of Teaching and Learning STEM.


Please follow and like us:

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.