We experimentally analyze leading by example in a public goods game with two permanent and two temporary group members. Our results show that leadership when permanent and temporary members interact leads to lower contributions than interaction without leadership.