Bias in Usability Testing

What does “bias” mean in usability testing? Is it bad? good? In between? What are sources of bias in usability testing? Which one(s) should we worry about most? How do we know our methods are any good? Especially when we all do things differently