The insidious nature of grades

by Ripal Nathuji | August 28, 2018

August marks the start of a new academic year for most students in Texas. It’s a transition period where they say goodbye to the relaxing days of summer and return to structured schedules of morning commutes and school bells. What is striking is how many of these underlying structural components have gone unchanged across multiple generations. Indeed, there are many aspects of our modern educational system that, as a society, we’re conditioned to rarely, if ever, bother to question. Perhaps one of the most pernicious is also a common source of stress and discouragement for many students — the pervasive application of tests and quantitative “all or nothing” scores as a means to evaluate outcomes. In this article, I’d like to revisit the the use of grades in our schools, and why they may actually lead to unintended negative consequences. As this is a topic of fierce debate in both education (teachers, schools, etc.) and corporate (companies that hire graduates, etc.) communities, I’ll caveat that the discussion herein is largely based upon my personal perspective which leans towards the idealist versus pragmatist region of the overall design space. No matter your personal predilections on this topic, however, I encourage you to treat this as a thought exercise in exploring one of the most fundamental and implicit conclusions we communicate to students in our school system today — that grades are what matter most.

Modeling the learner’s knowledge

As a starting point, it’s useful to contextualize this discussion by abstracting the role of feedback in learning. In the most general sense, feedback mechanisms in education are meant to help the learner (as well as other stakeholders) gain a continued understanding of what is “known” versus unknown. This understanding can then be made actionable in a variety of ways. The following diagram outlines a basic framework that can be used to organize a student’s understanding of their own knowledge corpus across four quadrants:

A key point when considering this model is that the content of each of the four regions is continuously changing. The goal of assessments external to the learner is to help the individual gain clarity with regards to their own personal state of knowledge. The specific implementation of an assessment might vary, of course. For example, when we hear and talk about experiential learning, we typically imply the use of hands-on applications to both provide an avenue for assessment (e.g. through some form of reflection) as well as the development of new knowledge through explorative experiences. Quadrant II might seem counterintuitive at first glance. After all, how can we not know we know something? In practice, this might be latent knowledge that goes unacknowledged simply due to lack of reflection or experience applying the knowledge. Similarly, quadrant IV might include items that a learner may simply not know about, or potentially things that they falsely believe they do know until they attempt to apply the knowledge. For the remainder of this post, I’ll assume we all agree that the purpose of education is to learn, and that learning can be framed as the expansion of what we know we know (e.g. quadrant I) via the various transitions illustrated using green arrows in the diagram.

The observer effect in education

So where do assessments tie into this discussion? All forms of assessment, whether they be derived from self-reflection, peer feedback, or inputs from a teacher / mentor, are critical towards establishing the boundaries of our self-assessed knowledge base. An accurate self-assessment is a prerequisite for understanding and prioritizing where we can increase what we know. Therefore, assessments in general are a critical component of education.

In physics there is something called the observer effect, which is the theory that simply observing a situation or phenomenon necessarily affects that same phenomenon. A basic example is when measuring the tire pressure on a car, you let out a small amount of air thereby causing a change to the pressure during the action of measurement. Often times, these observer effects are tiny and negligible, at least until you get into quantum phenomena. However, drawing a parallel to education, I believe that how we measure students via assessments has a dramatic, and potentially detrimental, impact on the learner.

Modeling the learner’s mindset

Before diving into the impact of the observer effect in education, let me first introduce a basic model for how students might perceive their own abilities and intelligence. Here I’ll borrow from the psychologist Carol Dweck who posed two categories of self-views that learners might adopt. The first is called the entity view which treats one’s intelligence as static and endowed at birth (this is also referred to as the fixed mindset). A symptom of this mindset is a need to continuously “prove” what you know to others and to be seen as smart while avoiding a demonstrable lack of knowledge, since any signs of the latter are an unaddressable flaw. On the other end of the spectrum is the incremental view which treats intelligence as fluid and dynamic. A symptom of this mindset is a desire to find “growth opportunities” where one can try new things knowing they might well result in failures. This mindset aligns with the “it’s about the journey not the destination” mantra that we often encourage students to espouse. In a previous post, I’ve already advocated for the benefits of instilling and nurturing the incremental view of self, particularly in STEM education. Here I’ll take for granted that as a shared goal we want to avoid instilling the fixed, entity view of self with students, and instead we desire to encourage growth mindsets via the incremental view of self.

Mapping back to our knowledge model, the fixed mindset in effect places a “wall” around quadrants I and II. Since intelligence is innate in this view, there’s really no need to understand (or divulge) what you don’t know. The lone transition arrow is valuable in so much as one might be asked to prove knowledge that is unknown to be known. The “winning strategy” so to speak is to establish expertise, and then minimize exposure to any perceived weaknesses in one’s intelligence. This leads quickly to diminishing positive returns in real learning and, in practice, intellectual boredom.

In contrast, the incremental view is all about a free flow of movement across the quadrants to support the growth of known-knowns. Here, a clear understanding of quadrants III and IV (unknowns) is the most important thing, since they represent the upside potential for a learner. Luckily, in today’s world of explosive knowledge creation, there’s no dearth of things to learn, so a growth-minded learner will never have to endure boredom. The challenge, in fact, is in determining a direction to pursue based upon an accurate understanding of one’s current knowledge base and personal objectives.

Quantitative feedback: Reinforcing the entity view of self

Let’s first start with the use of grades as a quantitative feedback mechanism in education. What I specifically mean here is the widespread use of metrics and rankings which revolve around seemingly “objective” tools such as numeric grades and GPAs. While these allow for the ability to sort and rank at various granularities (individuals, schools, etc.), operations that are seen as useful from a systems or administrative vantage point (e.g. to determine the “best” students or schools), they’re relatively meaningless in terms of serving as truly actionable feedback for the learner. Indeed, if we investigate how exams can be applied in practice to help learners identify what they know and don’t know, it inexorably leads us to a qualitative assessment of what knowledge was missing or misapplied in the context of the test. In this case the grade itself doesn’t matter — it’s about the analytics and insights that can be derived from the data. From an information theory perspective, then, the worst outcome is to get 100% on a test! While this helps the student reinforce quadrant I (what I know I know) and perhaps gain some insights into knowledge that can move from quadrant II (what I don’t know I know) to quadrant I, there’s no impact on quadrants III and IV which are the most critical for growth. In fact, the perverse incentive of grades when used as a tool to rank and sort is that this is exactly the type of behavior that gets nurtured. After all, there’s no rank benefit to gain a better understanding of what you don’t know — the best strategy in this game is to continuously prove what you know while minimizing downside risk by avoiding opportunities that might force you to demonstrate a lack of knowledge or that require purposefully exploring unknown areas (e.g. quadrants III and IV). Returning to our learner’s mindset model, this line of reasoning and behavior aligns perfectly with the fixed entity view. We often see this in practice when students pursue “easy A” teachers or classes to pad GPAs in lieu of opportunities to grow and challenge themselves which might carry downside risk to their grades. We’ve precisely encouraged the exact behaviour we (presumably) don’t want to see. Whoops!

Qualitative feedback: Reinforcing the incremental view of self

If not already apparent, let’s now contrast the disadvantages of a culture based on grades and rankings by exploring the use of qualitative feedback in education. In fact, to further note that I’m not advocating the position that tests and structured assessments are necessarily bad, let’s take the same exam scenario where I mentioned receiving a 100% is a poor outcome. Per our knowledge model, the role of feedback of any form is to help the learner gain a clearer understanding of themselves so that they can make an informed decision on the next course of action for self-improvement. Getting a perfect score simply means the assessment (in this case a test) simply wasn’t robust or challenging enough for the learner. Therefore, it’s imperative that the selected assessment includes knowledge a student might (or should) know along with items that might be new or unlikely to be known. From a purely quantitative standpoint this leads to a suboptimal score (e.g. a 70%), but we must separate the tool from the feedback mechanism. If we take that same exam score and extract the qualitative data regarding what topics the learner demonstrated versus what may be areas of growth and exploration, we can toss the grade aside altogether (after all, it’s completely meaningless in this use case). In practice, in an increasingly interdisciplinary world, designing and delivering exams that probe all potential permutations of exploration and application isn’t even feasible, so these types of assessments are largely a dead end in general. This is where qualitative feedback from teachers, facilitators, and mentors can be incredibly valuable. These stakeholders can probe a student’s knowledge in a more personalized manner compared to a one-size-fits-all exam. Since the end goal is to provide a qualitative feedback mechanism, the lack of objective, normalized scoring is removed, alleviating the instructor from comparing apples to oranges across students. From a behavioral standpoint, the use of qualitative feedback nurtures a continuing emphasis on the identification of unknowns so the learner can effectively and actively incorporate that information into their own, personalized learning plans. With the correct support, this allows the learner to pursue aggressive expansion of the boundaries of quadrant I which is exactly the incremental view of self we want to reinforce!

Why any of this matters

If you’ve read up to this point, and you generally agree with the argument I’ve laid out, you may be left wondering what practical implications any of this might have. After all, I began by acknowledging that readers should treat this as an idealistic thought exercise. A common conclusion when I have these types of discussions with people amounts to some derivative of: Like it or not, in the “real world” grades are what we have and therefore they define the boundaries within which we (as a community of learners and educators) must act. But, though we often forget it, we don’t live in a static world. Similarly to how there was a time when we all knew that transportation required horses, today we know that our educational (and, really, our professional) pipelines require grades. Every system seems impenetrable until it eventually collapses, and I’m 100% certain the same applies here (there are early examples in the market, but I’ll defer that discussion for a future post). Knowing and acting on the understanding that change is brewing leads to a different decision making process versus blindly following the status quo. So for now, I urge students of all ages to compartmentalize the actions required to “play the game”, and focus as much energy as possible towards identifying and integrating feedback mechanisms that encourage a growth mentality into their personal development and learning strategy. By doing so, they’ll find that removing grades as the ultimate goal doesn’t necessarily mean sacrificing them altogether. Instead, they’ll simply be the positive side effect of the kind of learning practices that will benefit them greatly in their personal and professional futures.