Students who take the Graduate Management Admissions Test for business school admissions soon will have their essays read and graded by a computer. Under the new system, computers will be fed hundreds of essays that already have been graded by human readers and scored as either good, average or poor. The computer then will grade a submitted essay by scanning the vocabulary and syntactic structure to find key words and phrases that constitute a “good” argument. The GMAT essays currently are graded by two human readers. Essays will soon be graded by the computer and one human reader; if the two give an essay widely disparate scores, a second human will grade it.
This new arrangement sounds eerily like something out of 2001: A Space Odyssey. By focusing solely on words and phrases that comprise a “good” argument, the computer may fail to take into account an essay’s creativity or originality. No matter how many sophisticated microchips and programs a computer boasts, it can never replicate the analytical skills a person uses when he or she grades an essay. It is not key phrases that count in an essay of this type – it is the clarity and originality of the writer’s argument.
Computers already have taken over much of daily life – everything from getting money out of the bank to taking the Metro around town. Students can apply to school via computer and long-distance learning exists only because computers do. The human element slowly is being removed from many tasks, replaced by a box of wires, chips and electronic gizmos. The GMAT computer grading system may be a neat technological improvement, but it seems more trouble than it’s worth.