Toward less misleading comparisions of uncertain risks: the example of aflatoxin and alar