According to a Businessweek article on the next version of the GMAT which is scheduled to be official on June 5, the test's developers say they spend "about six months to develop each question, with as many as a dozen people taking part in the process."
This makes a lot of sense to me. In my work in assessing students in their preparedness for standardized exams, I've seen that a question can test dozens of metrics at once. While no single questions can be conclusive, a full test with many questions can use underlying metrics to highlight meaningful patterns about the test-taker.
No comments:
Post a Comment