In their paper, de Vocht and Kromhout (2013) conclude that Benfords law can be used to evaluate the quality of data on occupational exposure. We agree with this conclusion in general but would like to stress that this is true only if the data set does not contain too many imputed values in the case of measurements below the limit of detection (LOD). In this letter, we will first derive theoretically why compliance with Bedford's law cannot be expected for such data sets. Secondly, we will show some examples of exposure data from the MEGA database that confirm our theoretical considerations. Thirdly, we want to explain why we think that our considerations about imputed values could be a reasonable explanation for the deviations from Benfords law found by de Vocht and Kromhout (2013) in the MEGA n-nitrosamine data. In the fourth place, we have a question regarding the selection of data sets from the ExAsRub study. Finally, we have some minor remarks regarding the publication of de Vocht and Kromhout (2013).
展开▼
