Chris Rush, head of something called New Classrooms, an outfit dedicated to pushing personalized learning in schools, said this at the NY Edtech Week global innovation festival when asked what “tech advances are missing in school”:
“Teachers spend a significant amount of time scoring papers rather than spending time with students. Automating not only multiple-choice test scoring but the grading of essays and project work would give teachers more time to focus on the student interaction that they’re uniquely capable of.”
If I had a time machine, I’m not sure what I’d use it for, but it’s pretty much a coin toss between killing baby Hitler and snuffing out the idea that we should automate the grading of writing in order to free up teachers to do the “important stuff.”
Because when it comes to teaching writing, grading is the important stuff.
Asking writing teachers to work with students without directly assessing their writing is like asking a coach to try to develop their players by telling them the score of the game, but not letting them observe any of the action. It is simply nonsensical.
As most know, but Chris Rush apparently doesn’t understand, grading is not just “scoring,” but involves evaluation, assessment, and what I think of as “diagnosis,” a chance to glean insights into the specific struggles of specific students. Spending time reading my students’ work is spending time with my students. It’s also a way to make the face-to-face time one spends with their students more productive and beneficial. To have something to discuss, teachers must be familiar with the ideas students are wrestling with, and how they’re going about the struggle.
Rush’s suggestion that outsourcing grading to an algorithm will benefit teachers and students by freeing teachers to do what they’re “uniquely capable of” somehow misses the fact that one of the unique things teachers can do which algorithms cannot is read.
It is disheartening to see this kind of thing showing up again and again in discussions of educational technology and educational innovation. I would like us to dispatch that myth that the quest for algorithmic grading of student writing is driven by such noble goals as “freeing up teacher time.” It simply isn’t true. Algorithms can kind of/sort of put a semi-accurate “score” on writing – they can also be fooled by nonsense – so those invested in education technology must create a rationale for its use.
We should not be taking statements like Rush’s seriously. If I am being charitable, they’re based in deep ignorance about teaching and learning. If I am being uncharitable…well, let’s not go there.
As I’ve said many times in many different ways, we know how to teach writing. We know that learning to write is difficult and complicated, and improvement takes sustained effort which is ideally supported by a close collaboration between students and their instructors.
Teachers do need the time and space to do what they’re uniquely capable of, but outsourcing grading to algorithms is not a route to achieving this goal. It’s movement in the opposite direction.
There is one benefit to the continued obsession in some corners of edtech with algorithmic grading. As soon as someone proposes such a thing, it’s a clear signal they don’t know what they’re talking about and should be dismissed from the conversation.