It seems one of my Devoted Readers is in the business of catching cheaters. Don Sorensen emailed me to let me know that he is a founder in a new company, Caveon Test Security. A bi-weekly newsletter called "Cheating in the News" is available to those in the testing and education industries. Archives may be found here. The June 4th archive alone contains seven articles, a few of which I might have missed, with my searches because they focus on certification exams. Dan mentioned that he'd like to point some readers my way, so I thought I'd do the same. Click here to read everything in my archive that's in the "Cheaters" category (although you'll need to enter the keyword "cheat" in my search engine to get the stuff that was written before May of 03).
I thought this story from Dan's June 4th archives was particularly compelling. It's about a student, Alexis Martin, who blew off her first SAT and then aced the second one, only to have ETS accuse her of cheating.
Read the whole thing. It seems to be, at first, a typical tearjerking tale about an innocent student vs. evil ETS, but then a real psychometrician got involved. He didn't take sides, but instead analyzed the data, and concluded (tactfully) that there is solid evidence to suggest that the "innocent" student may not be. An equally-tactful opposition to ETS's policies, though, is provided by a (named) member of the test prep community; nice to see that the typical "critics say" method wasn't resorted to here.
What's also interesting is that the student admits to one boneheaded action that, while it wasn't cheating, didn't help her case. She figured out during the test which SAT section was the variable one and blew it off, not realizing that if she were ever accused (and she was) of copying off a seatmate, it was her performance on this variable section (the only section differing from her seatmate's) that would suggest the true measure of her skill. This article is definitely impressive in the amount of solid information (lots) it presents vs. unsupported anti-testing cliches (none).
The tale ends with another tearjerker of a comment, of course - got to push the human interest side to keep the readers - but the article lays the framework well enough that a sympathetic yet logical reader can understand why the cheating detection system is not perfect, and why it is sometimes necessary to "catch" the innocent and "jeopardize a child for improving her score" in order to make sure that the real cheaters don't slip through the net.
Posted by kswygert at June 14, 2004 02:14 PM