Why did the Survey of Employees use a fixed sample size of 25 employees in each workplace?
In designing the WERS98 Survey of Employees, much was learnt from the 1995 Australian Workplace Industrial Relations Survey (AWIRS95). The Australian survey had incorporated a survey of employees in which a fixed sampling fraction was used within workplaces of a similar size. This sampling fraction ranged from 1 in 4 within workplaces having 20-99 employees, to 1 in 33 within workplaces having 500 or more employees. The use of a fixed sampling fraction within workplaces of a similar size meant that the weighting of each employee s data was relatively straightforward, since all employees within a particular stratum would be given the same weight. This, in turn, helped to keep the average design factor low at around 1.3. However, the sampling fractions that were used inevitably meant that, after taking account of non-response, the number of employees returning questionnaires in each workplace was, in many cases, too low to be able to compile variables that gave reasonable estimates of th
