In a recent New York Times, there was an article by Randall Stross, “The Algorithm Didn’t Like My Essay.” I found it interesting because I had posted back in May on a similar article about “robograding.” Stross writes about how consistent computer-scored essays are to human-scored essays in standardized tests such as the ACT and SAT. He then postulates that using computerized essay grading could free up human resources to interact with students on more meaningful issues at more meaningful levels.
While I don’t disagree with him on a philosophical level–what teacher doesn’t want ways to work smarter and not burn out on repetitive tasks–the fact remains that at present there is no substitute for human interaction when it comes to written expression. For the computerized scoring functions, the grammar and style checker that comes standard with Word does about as well as the standalone programs that supposedly excel at finding errors or at rating levels of sophistication. But none of these programs actually “reads” what a student has written. They can’t detect irony, they can’t identify falsehood, and they can’t recognize empty ideas.
The dilemma is that there are certain assessments that cannot be done quickly or automatically–at least not if they are to mean anything. Perhaps we should eliminate written essays from any kind of standardized tests. Stroll points out that even human scorers of such tests typically spend three minutes or less on each essay. Having been one of these scorers, I would agree with that estimate. In such settings, you train yourself to look for very specific things in a piece of writing; most of them relate more to modes of expression rather than modes of thought. It reminds me of Billy Crystal’s line from years ago, “it is better to look good than to feel good, dahling.”
In the next week, the students in my summer online composition course will turn in their rough drafts of their first essay, and they will then spend two days reading each others’ writing and commenting on it (as will I). Then they will go back to the drawing board (I hope) and spend time wrestling not just with comma splices but with the ideas that need further development. I am just nosy enough to want to know what they really thought about the readings I chose for the course, and I am naive enough to believe our future leaders in business, technological innovation, parenting, and education will be those who were more concerned with the content of their ideas than with the “color” of those ideas’ skin (apologies to Martin Luther King, Jr.).