making the evidence fit the theory