Published Online:
Published in Print: September 3, 2003, as Essays on New Teachers' Test To Be Graded by Computers

Essays on New Teachers' TestTo Be Graded by Computers

Article Tools
  • PrintPrinter-Friendly
  • EmailEmail Article
  • ReprintReprints
  • CommentsComments

An organization criticized for purportedly trying to shortcut the preparation of teachers now plans to bypass the human factor in grading the essays in its exams.

The American Board for Certification of Teacher Excellence is believed to be the first accrediting body in the nation to assess prospective teachers using artificial intelligence.

The Washington-based organization announced a partnership with Vantage Learning, in Newtown, Pa., last month. The company has devised and will score the essay portion of the ABCTE's Passport to Teaching exam, which was launched in August and is now being taken by a handful of future teachers in Pennsylvania.

"The benefit of using this type of technology ... is that it ensures consistency," said Kimberly B. Tulp, a spokeswoman for the Education Leaders Council, which formed the ABCTE together with the National Council on Teacher Quality. "Second, it is time- efficient."

Human assessors first set up guidelines to distinguish high- quality work from essays that are deficient, Ms. Tulp said. Then, the computer system is programmed to recognize those features.

"We're really digitizing the human experience," added Harry Barfoot, the vice president of marketing for Vantage Learning.

That is not necessarily a good idea, some experts say.

"I'm very skeptical that any artificial-intelligence program could do a good job evaluating essays," said Richard Alan Peters II, an associate professor of electrical engineering and the director of the robotics lab at Vanderbilt University in Nashville, Tenn.

Such a program, he said, "could probably be used to detect really egregious misuse of language or terrible grammar, ... but you can't develop meaning unless you start as an infant and interact with a number of humans for a number of years."

Better Than Human?

The new partnership fits in with the ABCTE's decision to rely heavily on computers to assess prospective educators on both aspects of the exam, Ms. Tulp said.

The organization, established in 2001 and financed with a $5 million grant from the U.S. Department of Education, aims to offer a national, portable credential for new and veteran teachers. To earn the credential, aspiring teachers must pass two exams—one in pedagogy that includes a written portion, and a multiple-choice test in the subject they wish to teach. Experienced teachers will take only the pedagogical exam.

In addition, both groups of teachers will be required to show proof of preservice instructional experience, which may include work in a preschool, a college, or the military, or online courses.

The ABCTE has been criticized for adopting a certification method that permits people to enter the classroom without formal training by taking what are seen as less-than-rigorous exams. ("New Teacher Board Parts Ways With ACT," April 23, 2003.)

The essays will be graded on six factors: overall analysis, focus and meaning, development and content, organization, language use and style, and mechanics and conventions, Ms. Tulp said.

A sample question requests test-takers to write a letter to parents and caregivers informing them of the school's attendance policy and of the importance of children's regular presence.

"We're not sitting here claiming that we can grade poetry," Mr. Barfoot said, "but when it comes to writing with components to it, [computers] provide the ability to grade much faster, with more consistency, and more accuracy than humans."

Many within the education community aren't sold on the development.

"If one is looking for a cheap, superficial way to assess composition, then one might use computer scoring," said David Bloome, the president of the National Council of Teachers of English, a professional organization based in Urbana, Ill. "But if one is seeking something more in the assessment of composition, at this point, it is truly going to require human eyes and human experience."

Computers are so predictable that test-takers may simply aim to beat the system rather than prove their worth, added Stephanie Hirsh, the deputy executive director of the National Staff Development Council, a group based in Oxford, Ohio, that emphasizes high-quality professional development for educators.

"We have," Mr. Barfoot countered, "over 120 studies we've published or used internally that talk to both the consistency of our scoring and accuracy."

Vol. 23, Issue 1, Page 11

You must be logged in to leave a comment. Login |  Register
Ground Rules for Posting
We encourage lively debate, but please be respectful of others. Profanity and personal attacks are prohibited. By commenting, you are agreeing to abide by our user agreement.
All comments are public.

Back to Top Back to Top

Most Popular Stories

Viewed

Emailed

Commented