Is it acceptable to mark off or comment on someone's presentation based on their non-standard English accent?

There are two ways to interpret the issue that is being raised. One is that the people simply have accents and not completely correct grammar in spoken English... but not to the point that it seriously impedes intelligibility. The other is that their English is so poor that it does greatly impede intelligibility.

The former is of little consequence, of course. The latter is a serious problem, in fact. It's not about bias for/against languages/nationalities, but about the context in which one operates. If one cannot communicate effectively, that's a minus. Even if/when people are "understanding", it's still not a plus, but a minus. Overcoming such an impediment should have a pretty high priority... as opposed to simply believing that such an issue solves itself in time. In particular, this will not happen for people whose English is sufficiently compromised that they spend nearly all waking hours with their fellow speakers, and thereby never practice the local language (as Dir. of Grad. Studies in Math. at my university on two different occasions, and participating in such issues for 35+ years, I've seen many cases).

In particular, outside of super-special cases, it serves no grad student well to have English be an obstacle. It will significantly obstruct their hiring on teaching-communication grounds, if nothing else. And, unavoidably, it reduces the impact of their portrayal of their research work.

Some decades ago, my professor summarized it roughly like this:

• The professional vocabulary must be right.
• The grammar should be right, but errors are acceptable if they don't affect meaning.
• The pronunciation is optional.

The goal was to train students to write acceptable papers, and to understand (or be understood by) other students who are benevolent and largely have the same accent.

I haven't been in academia for some years, so standards might have shifted at least in some places. Being able to present your results is part of scientific process.

I wish to elaborate on paul garrett's point, which I agree with:

It's not about bias for/against languages/nationalities, but about the context in which one operates. If one cannot communicate effectively, that's a minus..

Specifically, I do not agree with a popular approach that assigns some fixed fraction of the marks to "presentation" and the remainder to "content". In my opinion, marks should instead be assigned for "effort" and "content". Majority of the marks should be for "content", and should accurately reflect the amount of content that the student successfully conveys to the target audience. This applies to all types of student work.

For presentations in particular, this does mean that if a student is unable to convey any content whatsoever to the target audience, it is fair to give zero marks for "content". Poor English grammar or pronunciation is not an issue until it begins to affect the ability to correctly understand the content that the presenter intends to convey.

This also means that if the presenter successfully conveys the content via slides and hand gestures, then an inability to convey well that same content verbally is less of an issue. More often than you might think, presenters who speak perfect intelligible English fail to convey their presentation content well.

I do not think that anyone would dare to accuse you of unfairly giving lower marks if practically the whole class cannot understand what the presenter is saying. But you must make it clear beforehand that the grading will be based on how well the content is presented.

To preemptively avoid any accusation of racist bias, you can offer to give some feedback on students' presentations prior to the actual presentation date, but students have to request in advance if they want this feedback. During these feedback sessions you can specifically point out how they can improve their delivery of the presentation. But this is extra work, and not necessary for fair grading.