He teaches them. In my experience, only a few people in an undergrad EE EM class understand divergence and curl. Those that don’t probably won’t do well in certain EE disciplines.

When I was a young kid I came across a mini golf game for the computer (can't remember which platform, could have been Atari 130XE, IBM PC XT, or Apple iigs) that represented sloped surfaces as vector fields. So I knew if there were arrows pointing left, my golf ball going right would slow down and roll back to the left.

That kind of early exposure made it much easier to understand vector fields later on.

That is part of the teaching process, no?

I'm asking how he assesses their capacity, not asserting that he is not able to.

Not a statement hiding in a question :|