Can self-reported encoding strategy and recognition skill be diagnostic of performance in eyewitness identifications?