Diane Ravitch’s Blog includes two items of considerable interest on the topic of the computer grading of essays.
In the first post, titled simply “Why Computers Should Not Grade Student Essays” [http://dianeravitch.net/2014/09/03/why-computers-should-not-grade-student-essays-2/], Ravitch chronicles the efforts to create software that can generate essays that the grading software will evaluate as excellent. Although the computer-generated essays are structurally sound and grammatically correct, they are actually largely gibberish, and the purpose in creating them is to emphasize all of the elements of effective writing that cannot be reduced to patterns and rules. A computer program cannot distinguish nuances in tone nor can it recognize the subtle ways in which support for an argument can be turned so that rather pedestrian details become charged with meaning.
In the second post, titled “Gates Funds Programs to Grade Student Essays” [http://dianeravitch.net/2014/09/03/gates-funding-machines-to-score-writing/], Ravitch provides overviews of four examples of how funding from the Gates Foundation is supporting the development of this sort of grading software—not by non-profit institutes or institutions but by for-profit enterprises. This unusual allocation of funding from a Foundation serves to emphasize the degree to which the automation of education, like the “reform” of public education, is being driven by corporate interests, rather than purely altruistic intentions.
My previous posts on the computer-grading of essays have included the following items from the NCTE:
“Machine Scoring Fails the Test”: https://academeblog.org/2013/05/19/machine-scoring-fails-the-test/
“Annotated Bibliography of Machine-Grading of Essays, Part 1”: https://academeblog.org/2013/05/19/annotated-bibliography-on-machine-grading-of-essays-part-1/
“Annotated Bibliography of Machine-Grading of Essays, Part 2”: https://academeblog.org/2013/05/19/annotated-bibliography-of-machine-grading-of-essays-part-2/