This project seeks to improve our methodological understanding of how to do generic representative organizational surveys. The few general population private sector organizational surveys which have been done and report response rates tend to have response (or cooperation) rates significantly lower than the seventy to seventy-five percent which is normal for general population surveys of individuals. It is likely that some of the high level of non-response in organizational surveys represents the application of survey techniques developed for populations of individuals to organizations. The application of survey methods to organizations must come to grips with the differences between organizations and individuals. The project has two objectives. The first is to evaluate the representativeness of a final sample of establishments relative to the population in terms of the sources of response bias. These sources include omissions from the sampling frame and non-response to the administered survey or to particular items on the survey. The second goal is to develop a theory of organizational survey non-response. Once a sampling frame has been constructed, high response rates are essential in order to approximate the characteristics of a random sample and to insure population representativeness. It is anticipated that organizational characteristics associated with survey and item non-response will be identifiable. Since, almost nothing is now known about non- response in general organizational samples the probabilities of making a significant contribution to survey methodology and organizational research appear to be high.
