If 30 out of a 100 students were cheating and annoyed at the anti-cheating measures, his rating would significantly drop. That alone would be a sufficient explanation and is reason to infer causation from correlation, as we humans do every day of our lives to make sense of what is happening. We constantly successfully defy Hume.

You are not accounting for the fact that his ratings dropped, which implies previous classes did not have the kind of experiences you described and rated him low because of it.