[Klaus]

Distortions will not vary between samples in any substantial way.
Of course, there'll be slight variations - lenses aren't digital after all - but there'll be no qualitative difference.

Can you quantify the "non-substantial" way? To me the diff between  -1.28%, -1.35% makes 0.07%. I would not call that "substantial". That was my reason to ask for sample variation. Be it that, or variation in measuring method, I still don't believe that distortion is absolutely excluded from any variation. The differences might be very small and neglible, but there's absolutely no reason why this property should not be affected. There are differences in resolution big enough to appear on charts...

An obvious example is zoom lenses - there are slight variations in the zoom stops at the end of the range.
I can't really quantify what this means - 0.1-0.2% if we are talking about an ultra-wide lens.
I reckon the spacers used for calibrating a lens can also make a slight difference albeit surely less than 0.1%.

Please also note that the distortion analysis is the least accurate with Imatest - but that has nothing to do with the lenses, of course.

---

Back to the Sony FE 2/28 ....

I owned this lens and sold it, resolution in the outer image field is not good at F2 or F2.8 and there's an amount of LoCAs or even purple fringing at F2 as well.

But it's one of the cheapest offerings for native Sony FE (by Sony) so it might be interesting to see some test results given the affordable price.

That'll be pretty much the verdict.