Is it also possible to run this test with different settings for sharpness? Or isn't the program 'fooled' by this because the actual resolution doesn't change?

You can run the test at any level of sharpening, but it normalizes the result to a standard 2 pixel sharpening, so the short answer is no, the program is not fooled.

That's not to say your own eyes can't be fooled. Since sharpness emphasizes the contrast on edges, higher contrast and sharpening lends more punch and the perception of resolution.

You'd have to use Hornby or Marklin when testing the PAL version... :-)