Accuracy may make a difference, but does a difference make for accuracy? A response to Pritchard and Rosenblatt