I guess was looking at it more like of those that dropped out, how many ended up not succeeding, with the correlation implying that dropping out wasn't the right choice.

Agreed that there are quite a few confounding variables though. Any stats majors out there that know how to adjust for socioeconomic status + other biases?