Toxic Snapshot Grades
Everyone reads Shawn at ThinkThankThunk, right? Somehow he manages to post something this awesome in the same week a new BABY joins his family. Let’s all celebrate the fruits of his non-existent sleep schedule.
Shawn’s graph distills from 3 years of SBG data and captures my central problem with “grading” as a time sensitive activity.
When we grade students piece by piece, unit test after quiz after homework, we create the illusion that any subset of class time serves as a scale model for the entirety.
Teachers know that’s crap model.
Students should know that’s crap, but they’ve also been raised in a factory farm that conditions them to dislike 3 question quizzes because they “get a D for missing one question.”
Parents have NO WAY of knowing that model is crap, because we present them with a system with quarter/mid-quarter/weekly progress reports and built-in parent notifications for grades lower than XX%. Why would we do that if week 12 wasn’t just a scaled up version of week 3?
If a student’s grade at any moment is supposed to serve as a predictive indicator for how they’ll perform at the end of the term, then we’re hosed. The direct, obvious interpretation is that if a student’s grade ever falls below some threshold, then they’re off course and need fixing.
Learning doesn’t work like that. When students honestly encounter a new subject, their understanding is close to 0%. Every step up from that is progress, but grade snapshots present those learning moments as failure unless they leap straight to mastery. Learning is rock climbing, but for some reason our systems present it like the 200m hurdle.
This is toxic on every level, pushing all parties to optimize around precisely the wrong metric. Students focus on points per class session, which fuels the culture of HW plagiarism and extra-credit addiction. Parents see developing understanding as “falling behind,” a situation which demands extra workload, extra assessment, and extra instruction (probably from tutors). In economic terms, families are handed system where a improving a mid-quarter C carries a significant (financial/opportunity) cost.
The scale model fallacy may be most toxic for teachers. The practice is so habitual and deeply tied to institutional structure that we fail to recognize it as a design choice. When everything we do is warped to fit this model, we can’t see the stretch marks. Assessments become thin on content mastery but cluttered with topics, because each one needs to be broad enough to float most students through with only developing understanding. Those already confused grades become more diluted on the report cards, mixed with obedience components or flat-out leavened with extra credit. The scale-model interpretation of our muddled and commingled grades constrains teachers, to the point where we’ll schedule when to assess based on how many points need to be in the heap by week 3, 6 or 10.
I stumbled through a grading conversation a few months back where teachers were bemoaning the difficulty of grading rich lab reports, essays or problem solving. ”How do you decide what gets a 92 or a 94?” Yeah, no kidding! It’s hard to make minuscule arbitrary distinctions while pretending they’re completely consistent and meaningful! Why do we do that again? Because our classes rely on a fantastically small number of massively conflated assessments to communicate the full picture of a student’s developing understanding. Shawn’s graph suggests one way out of that invisible trap. Limiting grades to skill mastery can lead to better student engagement and understanding, but it demands a shift in the meaning of grades and how we communicate with students and parents.