A note on the triple difference in economic models