I doubt the compiler is responsible for assessing the test. I also doubt that gender-bias on the part of the assessor is the reason for the disparity though (assuming the disparity in this small sample is representative anyway).