What would be two or three major differences if an e-mail survey was used instead of a mail survey? Would this improve the results? What problems would it create?