Unfortunately its very hard to make a rational sort of choice at the junction point where you choose, because by then you've already decided you want to do a PhD and so any offer seems like a good one.
But you really shouldn't do this: you need to turn a critical eye to the facilities if you're looking at physical sciences, get a feel for the environment (basically, if people are joking something is always broken, that's a red flag because it won't be funny 18 months in), and try and figure out where your supervisor is at regarding publications - anyone who's established but say, lacks a Nature paper, or has one, needs to be treated suspiciously because their incentive to help you get papers published is low compared to wanting you to pursue high-risk/high-reward type activities. You also really need to pay attention if the project you're being asked to do matches up with what the lab/group you'll be in can do.
A PhD is taking effectively a massive paycut for nebulous future earnings potential, not to mention is asking a lot of your enthusiasm and creativity for a long duration.
I think this is the best way to sum up my original message. I wrote this blog post out of frustration.
When people criticize academia I listen because I think everybody's got something to say, even if part of their message is self-justification. In non-CS sciences with fewer pathways to industry, it takes guts to speak out about the truth of the situation because the culture selects so strongly against it.