I see that a TA position requires 20h every week which almost equals 2.5 days workload. Wouldn't that be too heavy for a Ph.D. student? If I don't do the TA, I could save the time to do high-quality research but I will lose 18K AUD stipend, which would make my life difficult at Sydney. Is it worth doing a TA?
A 20 hour per week teaching assistantship is fairly common in the US. In some departments TAs actually work less than the full 20 hours per week, while in other departments the actual work load really is 20 hours per week (of classroom teaching, preparation for teaching, grading, office hours, etc.) This varies somewhat by discipline and across universities. In mathematics, TAs traditionally have heavy workloads. From what I know, the patterns are fairly similar in Australia, but you would have to determine how things actually work in the department where you will be studying.
Yes, having to work an actual load of 20 hours per week as a TA will slow down your research compared with a Research Assistantship (20 hours per week working on your thesis and closely related projects) or a fellowship (no work requirement.) However, it does pay you, and the experience in teaching can be valuable.
Students working as TAs often take fewer courses per semester than students on fellowships. For example, a TA might take 9 credit hours of courses while a student on fellowship might take 15 credit hours.
Caveat lector: I am writing from a pure mathematics perspective. It could very well be different for someone who has to spend three hours or more every day in a laboratory dissecting something.
I don't know what constitutes as "TA'ing" for you. For me it meant teaching frontal classes and revising material from the lecture or solving exercises.
I used to teach between one to three hours (the one hour came with grading homework; the three hours came with two master students to do that). But in addition there were many office hours, extended office hours, answering emails, and so on. So one could perhaps said that I worked more than 20 hours a week.
In my last year I was also assigned to teach a graduate level course. What I did, that year, was as follows:
Try to get all your duties into two days, and focus almost only on research for the other three days.
If the above is not possible, try to keep at least two hours for research every day. Or one, if you can "almost squeeze" your duties into two days.
If you need to prepare exercises, set a weekly time to discuss with your professor what you want to have in those exercises.
If you answer a lot of emails, set aside an hour a day to answer emails.
In any case, barring grading exams, the weekend is sacred time for you to kick back, watch a movie, or hang out with friends and family. Not doing so is a surefire way of getting burned out and not doing well in both teaching and research. Sure, the occasional email or research thought can find their way into the weekend, but don't force it. Except with grading. Grading sucks.
Keep in mind that things like "20 hours per week" are often convenience thresholds - at my university that's a 0.5 FTE position, which is the most we can pay a graduate student, and it means I don't have to have them do time sheets and the like.
That being said, yes, being a TA is a lot of work. It's time consuming in a way that's non-optional, which can feel like a burden. But there's a couple caveats to this:
- It gets you experience teaching, and helps you refine your teaching philosophy. This will come in handy in many job applications.
- Not being able to pay for things is also often quite time consuming
- It might actually help inform your work - my thoughts on subjects I was working on were rather nicely refined while I was a TA, and I learned how to explain them fairly well - or at least I'd like to think so.