This came up frequently when I was in grad school. I doubt it is the right solution, especially for PhD students doing experimental work. For disciplines where

This came up frequently when I was in grad school. I doubt it is the right solution, especially for PhD students doing experimental work. For disciplines where PhD students are mostly independent and funded via TAships (humanities, sometimes math or social science), a lot of their aggravation is due to university or department policies: how many classes does each person need to teach, how many students per section, how are the classes assigned, how is summer funding allocated, and so on. Collective bargaining could be an effective way to sort this out. In the sciences, however, the model is very different: students are essentially apprenticed to one professor. If there are problems in that relationship, it's not clear to me how a union could successfully intervene since many of the disputes are both a) highly specialized and b) incredibly subjective. My PhD advisor, for example, sat on manuscripts forever. Could a union have forced him to provide timely, constructive comments on them? Probably not. Very few labs set explicit working-hour "quotas", which a union could regulate. Instead, there are productivity expectations: "this should just take an afternoon" and it's unclear how a union could ensure those are reasonable, especially for research stuff which, by definition, has never been done before! Finally, a lot of academia runs on reputation: recommendation letters are essential to get jobs, grants, and promotions. Even if there was an effective grievance procedure, it could get you through to a PhD--and then promptly doom your career. The biggest change I would like see is a transition from students "belonging" to one PI towards a model where all of the faculty are invested in all of the students' success. Management training would be good too. As a grad student and postdoc, you're judged on your ability to train the rats, run the reactions, or image the stars. This is how we pick faculty, and it's a surprisingly small chunk of the actual job, which is mostly people/grant management, with a lot of writing thrown in.