Report: Flawed Essay Scoring Algorithms Adopted by Exam Makers

The Associated Press
The Associated Press

According to a report by Motherboard, a flawed electronic essay scoring algorithm has been adopted by several states. Experts say the algorithms stress the ability to write in whole sentences, not creative or insightful thinking by students.

According to the report, several states that administer the GRE (Graduate Record Examinations) and other important exams, score essays with the aid of an algorithm.

The problem with automated essay-grading is that it scores solely on a piece of writing’s surface characteristics, such as sentence length, spelling, vocabulary, and grammar. This means that a student’s creativity or insightful thinking will be ignored during the grading process.

AI has the potential to exacerbate discrimination, experts say. Training essay-scoring engines on datasets of human-scored answers can ingrain existing bias in the algorithms. But the engines also focus heavily on metrics like sentence length, vocabulary, spelling, and subject-verb agreement—the parts of writing that English language learners and other groups are more likely to do differently. The systems are also unable to judge more nuanced aspects of writing, like creativity.

Norbert Elliot, a professor at the New Jersey Institute of Technology that previously served on the GRE committee, argues that machines simply cannot recognize the quality of a student’s writing beyond a few simple conventions.

“Automated writing evaluation is simply a means of tagging elements in a student’s work. If we overemphasize written conventions, standard written English, then you can see that the formula that drives this is only going to value certain kinds of writing,” Elliot said in a comment. “Knowledge of conventions is simply one part of a student’s ability to write … There may be a way that a student is particularly keen and insightful, and a human rater is going to value that. Not so with a machine.”

Tests conducted by researchers revealed that automated test-scoring algorithms will give high scores to essays that feature strong, coherent sentences, even if each sentence is completely unrelated from each other.

Stay tuned to Breitbart News for more updates on this story.

COMMENTS

Please let us know if you're having issues with commenting.