[jamesash]

An early stage lab has a lot in common with a startup. The PI is given seed money to validate an ambitious idea and has limited time and resources to do it. The work is extremely labor intensive and in the early stages the PI will be in the trenches with the early employees working in the lab and setting an example for them to follow. Fast failure and pivots are extremely common. Furthermore, at any time the research idea could get scooped by another hard-working research lab, resulting in publication in a lower-tier journal. The PI must get early results that can be used to obtain further research grants. A good PI in this situation will set expectations accordingly that 60+ hour weeks will be required and hopefully those who are not willing to do so will self-deselect. From their perspective I completely not understand not wanting to invest $50k/year of their startup funds in a graduate student who is not putting in the necessary time to succeed. The prospective graduate student's maximum power is at the time when they are choosing a research advisor. Choose wisely, and remember you are always free to choose not to go if you don't see a good situation. I advise that people think long and hard about whether or not they truly want to do graduate school. There are far better ways to spend your 20's than in a lab. I had two great experiences in "intense" labs in my field (organic chemistry) after a very poor experience with a prestigious but absentee graduate advisor in another field - but it is definitely not for everyone.

[BeetleB]

> An early stage lab has a lot in common with a startup.

The problem is you also see this behavior in well established lab, where the PI has very little to lose if he quits research altogether.

In my experience, the correlation with whether the professor was an assistant professor seeking tenure vs a well established full professor and the amount of work they make PhD students do is almost nil. It's mostly purely driven by the will of the PI, not the demands of the funding situation. A lot of professors choose to have few graduate students so they don't have to stress too much about funding.

Remember: Once you get tenure, your job is not at risk if your research output is low (although it does affect compensation). If you get a $1M grant and end up with no research papers out of it, there is no accountability - except the granting agency will deprioritize you in the future (if even that).

If you're seeking an advisor, ask his/her current students what life is like. Also find people who graduated under him/her and ask them.

[jamesash]

It's pick your poison, like choosing a startup vs BigCorp. My observations pertain to organic chemistry specifically, I'm less familiar with other fields.

With an assistant prof you will be asked to do unreasonably long hours on ideas that haven't been fully validated, but you get a lot of face-time with the PI (for better or worse) and have a chance to get on the ground floor of a potentially great research program. Your boss has a lot more invested in your success because their career depends more on you. It can be exciting.

With a more established prof you are more likely be plugged into the n-th iteration of an established research program. You're more expendable because the prof has more options to hire your replacement; you can get lost in the crowd; the prof won't be in the lab as much (travel, service responsibilities) so it's more sink or swim. Lots can go wrong in this environment. As the group gets larger, internal politics can create frictions, especially if the supervisor is away a lot.

I see A) as better than B) but that is just me.

[BeetleB]

Yeah, it will vary with the discipline/major and possibly also the rank of the university. My point was that from what I've seen, the startup vs BigCorp dichotomy wasn't as clear cut as you describe. Also, there is the third category: Small business. Few grad students, and relatively low pressure.

[jamesash]

>Also, there is the third category: Small business. Few grad students, and relatively low pressure.

These can be the saddest cases in my opinion. Lower downside, but lower upside too. You work your butt off and knock the ball out of the park with your project: who notices? A few experts, in a very narrow field, may recognize the significance of what you've done. Can be lonely.

[tptacek]

This is the understandable part of researcher and grad student abuse and the sense I get talking to people is that they do indeed enter the field expecting to duck in on weekends to tend to experiments, and expecting not to be well compensated.

The part that doesn't fit is their PIs being continuously abusive, or, worse, destructive: the PI's goals are ostensibly those of a supposed startup founder, to drive a team to the redline to get something delivered in an unreasonably short period of time, fair enough. But PIs deliberately fuck things up, disrupt experiments, chase team members out of the lab for no apparent reason, berate team members to the point where they can't be productive.

The culture is much more reminiscent of a 1980s fine dining restaurant kitchen than of a startup. The chef paid his dues, and if having a pot half-filled with hot stock thrown at his head was something he had to deal with, by god, so will his line cooks and stages.

[jamesash]

Two most extreme stories I am most familiar with (did not experience first-hand) were of a PI throwing a chair in the direction of a postdoc, and of a PI having a screaming fit directly in the face of a graduate student. Both assistant profs.

One got tenure and became more mellow in a relative sense. The other was subsequently unable to attract any more graduate students to their lab, and left for another, slightly less prestigious institution.

I am also familiar with one case of graduate students resigning en masse, leaving a PI with a hollowed-out research group of 2-3 that they had to rebuild. Prof had a subsequent attitude adjustment.

In all these cases the stories spread pretty quickly and profs realized that there were costs to abusive behavior. These stories pre-date social media BTW.