Assessment of bias in a survey of residential magnetic fields in Melbourne, Australia.

An important consideration in surveys of residential magnetic fields is whether the sample of homes measured is representative of the population. This study assessed the potential bias in a survey of magnetic fields conducted in 296 homes randomly selected from the telephone directory in Melbourne. The choice of the telephone directory as the sampling frame introduced some bias when compared with census information. There was also evident self-selection bias with residents twice as likely to participate if they lived in the vicinity of visible electrical infrastructure such as transmission lines. However, when corrected for bias the corrected magnetic fields were only slightly lower than the measurements with a percentage difference in the range of 2.8-6.0 %. These results provide confidence that the measurements in the survey were good estimates of the distribution of residential magnetic fields in Melbourne.