One‐sided comparisons for treatments with a control