“…Reliability, defined in this way, has been used in a number of applications aimed at understanding the extent of randomness in survey response (e.g. see Achen, 1975;Alwin, 1989aAlwin, , 1989bKrosnick, 1989a, 1989b;Alwin and Thornton, 1984;Andrews, 1984;Asher, 1974;Bieiby and Hauser, 1977;Featherman, 1977a, 1977b;Borus and Nestle, 1973;Converse and Markus, 1979;Corcoran, 1980;Erikson, 1979;Hauser, Tsai and Sewell, 1983;Jagodzinski and Kuhnei, 1987;Marquis and Marquis, 1977;Siegel and Hodge, 1968;Smith and Stephenson, 1979). The general conclusion of this body of work is that unreliability of measurement occurs with respect to a wide variety of content areas, and while the measurement of attitudes may be particularly difficult (Converse, 1964), substantial amounts of error occur in the measurement of other phenomena as well.…”