I was finally able to pull up that study and it looks like the quant reasoning section on the CogAT plays an even bigger role than the nonverbal section. So for the nonverbal section, if you look at the tail end of the distribution (the top 1 or 2 percent), there will be 1.2 or 1.3 boys for every girl but for the math section it's even higher with more like 2 or 3 boys for every girl. Since our district is very strict about wanting CogAT scores in the top 2 percent (even for gifted programming in the normal classrooms), I guess that explains all the boys in the program. I'm not sure what role the achievement testing plays and whether there is a gender difference.