I think it depends heavily on what you're mocking and what the code you're testing looks like. I think Martin Fowler has a nice dichotomy where he splits tests

I think it depends heavily on what you're mocking and what the code you're testing looks like. I think Martin Fowler has a nice dichotomy where he splits tests up into solitary (generally using mocks for deps) and sociable (generally using real code for dependencies): https://martinfowler.com/bliki/UnitTest.html https://martinfowler.com/bliki/UnitTest.html