gazit@cs.duke.edu (Hillel Gazit) (04/13/91)
In article <1991Apr11.043308.16202@beaver.cs.washington.edu> fester@wolf.cs.washington.edu (Lea Fester) writes: >I have heard computer science professors note with amazement that the >verbal tends to predict academic success for PhD candidates better >than the math does. I suspect that it is more than a Politically Correct Legend than truth. I know a guy who was in the *bottom* 13% of the Verbal test and in the top 3% of the CS advanced test. If the Verbal measurement was a good predictor then that alone had to rule out that guy. It did not rule him out because some universities prefer not cross out their best advance test candidate. >I have no argument with using a test that is an accurate predictor, >assuming it really is accurate. In *my* opinion the advance test result is a better predictor, but it is Politically Incorrect to show that.