When measuring end-to-end, any part of that may influence the final outcome. If, and I am not saying this is true, if there is a tendency to have a few more errors (especially grammatical with it's its) in a non-native tongue, success outcomes MAY point to a bias in interviewing, or in how people interact internally.
Or not. It is just something I wondered, and something it doesn't seem like you've looked at specifically, and understandably as well, because who wants to open themselves to accusations of bias with their own data?
That said, I totally agree with your comment here that the preparation time available should render resumes free of such errors. They should be peer reviewed, as well. Problems caused by a lack of either indicate a character flaw to me.
decent -> descent.
Muphry's law strikes again. https://en.wikipedia.org/wiki/Muphry%27s_law
The USA is more permissive of double negatives for example.
And this ignores the fact that some candidates will have dyslexia.
If there are people of non-English decent, I wonder if they would be more likely to have errors, especially grammatical, that are understandable. If that is the case, this may simply be a correlation to poor interview performance, and a bias towards the singular group of native English speakers, and that they may have let great programmers go.
Now, I'm not saying it is true, I am just saying that is a possibility that would alarm me when errors on a resume is such a huge indicator, and wonder whether that is not something they should be correcting for.