Postal surveys of physicians gave superior response rates over telephone interviews in a randomized trial.