There’s a lot of banter going on right now about a new study that suggests automated essay graders can be as effective as humans.
My colleague Erik Robelen wrote about the study from the William and Flora Hewlett Foundation last Thursday, prompting a rapid string of comments from readers who, it’s fair to say, are quite invested—positively or negatively—in the concept. The Hewlett Foundation also helps fund Education Week‘s news coverage.
Meanwhile, ed-tech opinion blogger Justin Reich has published the first of three posts on the topic of automated essay grading in response to the study, explaining the basic theory and concepts behind how such a system analyzes and rates written work. He plans follow-up pieces on how automated grading would reshape assessment and reshape teaching.
And over at Hack Education, Audrey Waters gives a long (but very interesting) look not only into the possible ramifications of automated essay grading as an isolated practice, but how it fits within the broader movement across education to more mechanized methods.
Anything that proposes humans could be replaced by machines is bound to draw attention, and transform from a technology issue to a labor issue. But it’s less clear what the real impact would be.
Would the technology merely enable more written assessments as part of standardized testing, or lighten the teacher grading load (and possibly the size of the teaching source) by mechanizing the majority of writing feedback in our public schools?
Would it merely encourage students to write more directly, and with a greater focus on organization, for the sake of a computer evaluator that prioritizes essay structure, or would some students abuse the system and be able to write work that is rigidly organized and grammatically flawless, but makes no actual sense?
Perhaps someone should write a five-paragraph essay explaining it. Or, you know, a news story.