Even in ideal conditions, it depends on the project, the language, the methodology, the architecture, on the stance on mocking, on the engineering philosophy us

Even in ideal conditions, it depends on the project, the language, the methodology, the architecture, on the stance on mocking, on the engineering philosophy used. It depends even on what a person considers a "unit test". For all I know, by "unit test" you might also mean what people call "integration test".