Equitable Grading

You have heard me reference this idea called "Equitable Grading" several times in FCCS, FOP, and now this course.  My journey towards embracing equitable grading is a fairly recent one, and I don't even remotely claim to have gotten it right yet.  I'm still adjusting years of previous notions about teaching/learning/grading with this newer concept (to me, maybe not to you). 

This all started with an article from Mindshift (part of KQED, a public media outlet in Northern California). The article was titled How Teachers Are Changing Grading Practices With an Eye on Equity. A colleague of mine found the original article, examined the the web site and bought the book (from the web site). He shared all of this with me and we discovered that there was a name for what we had been trying to do for a very long time.

What comes below is an attempt to summarize what is contained in the Equitable Grading book and share this with you.  You will, hopefully, see some of these ideas in how we have structured the grading in your last couple of classes.  I will admit, I personally am having a hard time adjusting to this "whole hog."  But I also have to admit, that I am slowly discovering that many of the things I thought I originally disagreed with I am starting to come around to.  In other words, you should look carefully at what is described below and give serious thought about where/when I don't follow what is here because I may slowly becoming convinced that what I AM doing is wrong and what is here really is better.  MAYBE.

Having said all of that I am not asking you to buy into this or "drink the Kool Aid" as the common phrase goes. We just want you to be aware of all of this and give it serious consideration in your own teaching even if you don't know if you buy it yet.

What is Equitable Grading

My current understanding is that equitable grading is a set of practices that focus on accurately assessing student capability/understanding rather than some other elements. The discussion below attempts to present sufficient information about the practices to allow you to consider adopting/adapting them. If you have questions, please contact us.

Joe Feldman includes several principles (3) and a number of grading practices (18) in his grading for equity playbook. The following table presents them and selected elements are further discussed below the table (which is taken from p.72 of the book).

Much of the significance of equitable grading arises from the notions that grades should accurately reflect student understanding that the common percentage scale (100 points) does not do that. First, the range of failing scores is the same as the other four ranges combined. That is obvious, but until recently I had never really thought about it and its implications. There are 10 degrees of A or C but 50 degrees of F. Really? Does that seem right? Appropriate? What is the argument for it? Second, to be accurate, we must discriminate between scores. What is the difference between 85 and 86 (or even 87 or 88)? Or, what is the difference between 40 and 50? It is hard to find any true difference, particularly when we think of student understanding

For a long time I have struggled with the 100 point scale and its use. Should we start with 100 points and deduct 1 (or 2 or 3) points for each missing good thing or each wrong thing present? Is one more important than the other? Should we start with 0 points and add 1 (or 2 or 3) points for each good thing present? Realizing there was a problem was easy, finding a solution is much harder. The ideas here can help us explore and, perhaps, create a solution.

Grading Practices Discussion

The discussion below provides our/my current understanding some of Feldman's practices, their rationale, and perhaps suggestions for their use. Let me/us know if you have questions. Note that some practices are grouped together and that the discussion does not follow the order in the table above.

• Minimum grading;  0-4 scale

  (I like to think of minimum grading as not doing very much grading :-)  But, that's not what it is.)

  Minimum grading and using a 0-4 point scale are related. A key idea with minimum grading and using a 0-4 point scale is that in common grading systems a failing grade (an F) is about half of the percentage grading scale with the other grades each only being one-fifth the size of the F range. That means that while combining a C and an A would would result in a B, combining an F and a C (or even an A) might well result in the grade still being an F. A low F combined with four middle of the road Cs would barely produce a D. Does that seem to accurately reflect what the student has learned? That seems really out of whack when you think about it from a standpoint of accurately reflecting overall student capability.

  An example of "minimum grading" is to record an F as a score in the range 50-59. That way, all the grading levels are of the same size and the effect of a failing grade is no harder to offset than a D or C or B. For example, an F combined with an A might well result in a C (which seems sorta reasonable as opposed to still being an F).

  Using a 0-4 point range when recording scores automatically uses minimum grading and it correlates nicely with GPA scores. It also correlates well with standards-grading—Exceeds Standards, Meets Standards, Approaching Standards, Standards Not Yet Met, Insufficient Evidence (or whatever names you might choose).

• Avoiding Zeros

  The key idea with respect to zeros relates to the meaning of the zero. A zero could mean the student hasn't done the homework, learning activity, exam, etc. It could also mean that the student did something but did not demonstrate any understanding/learning. If the zero indicates lack of completion, it does not accurately communicate the student's understanding. We do not yet know the student's level of understanding.

  So, ... use zeros only when you have both assessed the student capability and there is none. For our assessment/grading to be accurate, we can not use a zero when we have no evidence of the student understanding. This means, of course, that our grading system will need to differentiate between the two cases.

• Grades based entirely on summative assessments, not formative assessments (such as homework);  Grades based on student work, not the timing of the work

  This practice may be the one that is hardest to deal with. Homework (learning activities) typically make up half or more of an overall grade. We want to give students credit for doing the learning activity (homework). We don't want to base the grade solely on a couple of exams and the final. And, how can we give feedback for learning if we don't grade the learning activities? On the other hand ... Including mistakes students make while they are learning is not fair (or right). Grading learning activity penalizes slower learners and students with temporary interruptions from real life, e.g., illness.

  It turns out there are a number of benefits to not including scores on learning activities (homework) when we grade, particularly with respect to programming.

  • Copying is not a problem

    If students copy, they learn less or nothing. And, if we don't grade homework there is no incentive to copy. Seeking help from classmates or the Web only affects their learning, not their grade.

  • We have less to grade (Duh!)

    (That's reason enough for me!) With time gained from not grading homework, teachers can focus on something that actually helps students learn, e.g., finding better ways to provide feedback on student learning/performance.

  • Students assume the responsibility for learning

    Assuming the teacher explains that learning occurs from doing the learning activity/homework and seeking help when difficulties are encountered, students will eventually recognize that doing the homework is their job and their responsibility.

  • Punishment and life factors don't directly impact the grade

    When we grade learning activity/homework students are punished for not doing it (or doing it late), even if they already can do what we want. Those who don't already know are doubly punished—once for low grades on the homework and once when they take the exam. Students for whom life interferes (e.g., illness, a death in the family, noisy neighbors keeping them awake, etc.) are also penalized, particularly by late penalties, when they could have learned just fine given a little more time.

  Traditional "homework" is meant to be an opportunity to learn and/or practice what is being learned. Because one learns more slowly, has trouble with particular elements of content, had difficulty with the type of learning activity or its directions, etc. should not negatively affect one's grade or assessment of capability. Grading learning activities should be discontinued.

  But how will students get feedback about their understanding and/or capability. That's a key idea. As teachers, we will need to figure that out. For programming, public code-walkthroughs and in-person grading can help. These ideas are discussed below. Likely, you can think of some other way(s) to provide feedback, perhaps looking over the shoulders of students as they write their programs and asking them questions. (That's more fun than grading.)

• Weighting more recent performance;  Retakes and redos

  In all of education we are concerned with developing particular student capability. Some students will come to us with nearly all that capability due to their past schooling/experience while others will arrive with essentially none of the desired capability. It should not matter where there start, only where they end. And it should not matter if they learn swiftly or slowly—only that they learned. The implication here, is that an average score over the term of the learning does not accurately reflect the ending capability.

  So, ... it makes sense to do what we can to aid student learning and to use assessment strategies that accurately measure desired capability while encouraging students (or a least not discouraging them). Using, primarily, the best performance (which is usually the most recent) and allowing multiple chances at demonstrating capability will do both.

  A few years ago, I thought I had come up with a brilliant idea--competency demonstrations. These were quizzes that I could use to assess student capability on particular topics. I could grade them quickly and students were allowed to take them repeatedly until they demonstrated competency. I talked about them with colleagues from the UK and was told, "Yeh, we do that all the time. They're called driving tests." When we take a drivers' test, we can repeat it until we pass and the final attempt is not averaged with the earlier ones. What counts is that we got to the desired capability. Seems reasonable to do the same with school learning. And, programming seems particularly amenable to such an assessment system.

• Grades based on an individual's achievement, not the group's

  This has always been hard for me—I've never actually done it. But it makes perfect sense. When we assign group work, our goals typically are to enhance student learning through a group process and/or to have students develop the "soft skills" related to group work. However, we want to assess students based on individual performance. So what do we do.

  (Not sure since I've not done it, but ...) What we need to do is to develop an assessment that allows us to examine individual capability with respect to our goal(s) for the group work. And as we assign the group work make sure students understand that the goal is individual learning, not the production of some group product. With programming, that can be relatively straightforward. (With Teaching and Learning Programming, it is less straightforward.)

• Grades based on standards scales, not (total or average) points;  Renaming grades;  Rubrics

  The primary goal here (I think) is to replace the traditional scheme of awarding points and averaging and weighting them to determine a percentage value that is translated into a grade. Whatever we replace it with needs to allow the teacher to exercise professional judgment of a student's capability. That judgment, of course, must be based on the data from the assessments we've selected/constructed.

  Rubrics can supply a mechanism for accomplishing this, and constitute the renaming of grades (e.g., A: Exceeded standard; B: Met standard; C: Gaps in capability RE standard; D: "Some" capability RE standard; F: Little/no capability RE standard). We will need, of course to identify the various expected capabilities for which we will define standards.

• Grades based on required content, not extra credit;  Alternative (non-grade) consequences for cheating;  Excluding participation and effort

  If we are to truly base grades on assessed student capability, then all the extraneous elements we may have used in the past must be omitted from our assessment practice. Thus, we no longer explicitly (or implicitly) include extra credit, participation, or effort. And, we don't penalize students for behavior not indicative of their capability, i.e., cheating or copying work from others.

  Reward is a more powerful reinforcer than punishment and we want to reward student success. Success is developing the desired capability. All other (behavior-based) considerations are immaterial insofar as student capability is concerned. If there are particular elements of participation (soft skills or self-regulation) capabilities that we wish to inculcate they would become a goal of instruction with its own assessment.