Modeling, even if incorrect, is important for at least 2 reasons. The most obvious one is that we can refine the theory as new results come out of the experiments. The second is that a successful model, even if wrong for the experiment it was devised, will eventually shed light on a similar problem elsewhere.