I read something similar, only in this case GMAC was going to add retinal tracking equipment to track your eye movements. They want to determine who is using process of elimination as a test-taking technique and penalize them for not working out each problem using every possible answer. They are also testing seat sensors for the chairs to determine if someone is pulling something out of their a$$ during the AWA. I think I read this in Jackass Quarterly.
Seriously, how the heck would they determine whether someone guesses or not? I don't agree with your assertion that the computer would "clearly" know that there would be guessing involved in a situation where one answers 5 questions in a minute. Is it likely? Yes, but certainly not clear. How would it determine this?
GMAC tracks how long each individual takes to answer the question correctly or incorrectly. But even with that information there will be wide discrepancies in the time used, from very low to very high. They could use this possibly to see if someone falls under the expected time range. But this assumes they can determine whether the numbers used to determine this range is free of guessers' times, or geniuses' times, or those who thought they knew the answer but didn't and just stumbled upon the right one. HOW WOULD THEY DETERMINE THIS?
This idea also throws in a degree of subjectivity into a standardized test. How to determine how much of a penalty to assign a supposed guesser who gets an item correct? I would contend that this issue, even more than the actual mechanics of identifying guessers, would be the single strongest reason to reject any such proposal.
