I don't know whether this is better. On the one hand, you don't need it. On the other hand, keeping the symmetry and stating the value explicitly (not only as a comment also has some appeal.

Execution-wise it makes no difference. There is however a difference in case the if-else-chain gets out of sync with the check_in_list above. When using else other values will be treated like "equal". When using elif mode other values don't get that execution block and the function will continue. Both leads to an incorrect result, but the elif solution will most likely blow up in one way or the other, which is preferable IMHO.

The reason was to be able to get full test coverage. Now I tried another pattern that I've seen used: placing the check_in_list in a final else-clause, so more used to raise an error than to check the arguments beforehand.

I guess that in general there should be a preferred way, but it seems like there isn't.

Personally, I don't worry too much about full test coverage. It's onerous in some cases, and full test coverage does not mean your code is working correctly. Test coverage is only a weak indicator and we should be pragmatic about it.

I agree there should be a preferred way. Maybe we should briefly discuss this in a dev call. Unfortunately, there is no really good solution. You always have the duplication between the if/else cases and the values in check_in_list with the danger of being out of sync. For readability, I prefer check_in_list at the top, because that limits the scope what can happen in the code afterwards and thus reduces mental load.