Thanks for the complement, you flatter me
I think I see this from a different angle. Firstly, let's tackle that generalisation point. That's not really a fair challenge. Your point was that "Indian women" will find PTE doesn't work for them. That's a generalisation
you made that I ran with. If you're going to challenge my rebuttal, based on the idea that I am generalising, you're actually challenging your own point, because it was you that initially made the grouping. So, yeah, you're right, different accents etc. but your point was on Indian women in general.
Onto the main discussion. Your wording points to where we differ:
and
My perception is that you conflate the average profile of the test taker and the target profile for speaking "excellence". The purpose of the exam is to test capability within a native English speaking environment. In reality, what would be ideal would be to train the exam
only on Australian accents for DIBP use. That's probably prohibitively expensive and would likely get complaints for the British (they did invent the language, it would be a bit rich to claim they can't speak it.... though looking at cricket, there's a point to be made
). The Americans would likely complain also, and as they tend to run the things, I guess we're stuck with their accents being part of the training data.
I digress. Back to the point. By limiting the training data, the system is
designed to discriminate accents that are distinct from native speakers. In fact, my argument would be that it is likely the IELTS assessors are more prone to incorrectly rate strong accents because their familiarity with these accents puts them in a position where a) they have a better ear and b) they are unaware of this and assume their ability to detect meaning from the strongly accented speech is typical. It probably isn't, and that's a tough bias to compensate.
Your thoughts?
