CS general knowledge tests of people going through the same program are not at all a good proxy for a GMA test - there is a lot of variance in how people actual

CS general knowledge tests of people going through the same program are not at all a good proxy for a GMA test - there is a lot of variance in how people actually go through the same program, even with the same grades. You can't abstract that down to a test and assume the difference is the result in a difference in intelligence.