> if a study can't be replicated then the study is wrong isn't it?
Not necessarily. Maybe the replications were flawed. Characteristics of a population can change over time. Blah, blah, blah. It's incredibly hard to say something is flat wrong or absolutely correct. A good scientist uses what others might call "weasel" words (I hate that term), like, "The data is (in)consistent with the hypothesis."
The more evidence that mounts for or against a hypothesis, well, it's up to you to decide how to act.
either the sample was too small, they got unlucky or the result is stated too generally.