The reliability of psychological instruments in community samples: A cautionary note