I would add that you should choose a lab with lots of funding. Big labs might mean you don't spend much time with your on-paper supervisors, but you will get a

I would add that you should choose a lab with lots of funding. Big labs might mean you don't spend much time with your on-paper supervisors, but you will get access to post-docs who can mentor you better than any professor who doesn't step into a lab. Big labs with funding also have lots of machines which can help you churn out data. Churning out data is very very important! Source: I failed my PhD partly because there was no one to mentor me, and no resources to produce results. As a PhD student, you don't know what you are doing and you need those resources so you don't waste your time.