Can you write a test that does more harm than good?
In the intro to his upcoming book The Art of Unit Testing, Roy Osherove describes a case where poorly-designed tests actually hurt his team's progress:
Worse yet, some tests became unusable because the people who wrote them had left the project and no one knew how to maintain the tests, or what they were testing. The names we gave our unit test methods were not clear enough. We had tests relying on other tests. We ended up throwing away most of the tests less than 6 months into the project.
I'm on the fence. Anything that exercises your code in a automatic, repeatable way has got to be useful. On the other hand, poorly written tests cost time and effort to understand, refactor, and fix.
Are the drawbacks to bad tests worse than having no coverage at all?
I think the answer is that in the short term, even bad tests are useful. Trying to squeeze a extra life out of them beyond that, however, pays diminishing returns.
Just like other software, your tests should be built for maintenance, but in a crunch, you can punch something in that works. It's better to have bad tests than to have untested code.