The methodology is flawed but the kernel of the idea is good - which universities provide the best platform for would-be founders?

Things like proximity to VCs, labs with a history of spin-offs, successful alumni angel networks, etc would all play a role.