TAs aren't overworked, even without union representation. You'll never be asked to do more than an average of 20hrs/week, ever. You can always just refuse to sign the end-of-semester certification.

RAs are overworked, but unions can't do anything to help that. As a grad student, you need to publish papers, which means you need the cooperation of your advisor. If you bring a grievance or work to rule, there are a million innovative ways your advisor can sabotage you as retaliation.

They can let you pursue ideas that they know are not novel (and the papers will be rejected). Or suggest you try something they know won't pan out. Or run out of budget before ordering that $THING you need for your experiment. When you still haven't defended after ~8-10 years, the university's policies will automatically kick you out. Problem solved!

As for a living wage: I wasn't represented by a union in grad school, but all the commenters here who were have pointed out that their unions didn't fight for higher wages.

> If you bring a grievance or work to rule, there are a million innovative ways your advisor can sabotage you as retaliation.

That sounds like exactly the kind of thing a union could address.

> I wasn't represented by a union in grad school, but all the commenters here who were have pointed out that their unions didn't fight for higher wages.

It's possible that they were fighting for other things considered by most members to be more important. Ultimately though, if a group has a union and they feel that the union isn't representing their interests there are remedies for that. Members can vote to change leadership, or in extreme circumstances dissolve the union and form a new one designed to prevent the problems you had with the old one.

> > They can let you pursue ideas that they know are not novel (and the papers will be rejected). Or suggest you try something they know won't pan out.

> That sounds like exactly the kind of thing a union could address.

Short of having a group of union members becoming your defacto supervisor and doing the guidance your supervisor should be doing, I don't see how they could help in this situation.

Union puts rules in their contract saying 0 retaliation when grievances are filed or disputes arise, explicitly calling out commonly seen retaliations while still leaving room to account for other sneakiness or else X and if students who have complaints feel retaliated against they can now appeal to the union who (if convinced) proceed to X where X can mean anything from disciplinary action against the supervisor, lawsuits, or strike.

Supervisors who want to avoid X will then be forced to reconsider retaliating against students or at the very least find ways to do it where students don't feel like they're being retaliated against which I suppose is the best you can expect since assholes exist and some percentage of them will end up as supervisors.

Yes, I understand the idea of such clauses. My specific argument was that for academia it's very easy for them to retaliate in ways that are very non-obvious as being retaliations, especially to outside observers:

> > They can let you pursue ideas that they know are not novel (and the papers will be rejected). Or suggest you try something they know won't pan out.

in my experience, the overwork of the RAs wasn't being pushed by a ruthless taskmaster advisor. it was that students understood the hypercompetitive academic job market and knew that they had to build out their CV if they wanted to be lucky enough to get a postdoc once they defended.

Grad student unions can stop TAs from being overworked for their classes. But most of the overwork actually stems from trying to start a research career.

there’s nothing a union can do to fix research-based overwork because youre a) doing it to yourself, not because a boss is making you and b) you are competing with researchers from other institutions that aren’t in your union to build your CV.

"There's nothing a union can do to fix structural inefficiencies" seems like an obvious and irrelevant take. Unions are for securing benefits within the existing system, not for changing the system per se.