The aim of the present paper is to present a comprehensive review of the issues involved in exposure assessment for occupational epidemiology studies and to provide an example. Exposure assessment for occupational epidemiology studies is becoming more quantitatively refined. This paper discusses important issues that need to be taken into account for exposure assessment, with particular reference to occupational asthma. It discusses issues such as survey design, data collection, the effect of measurement error and data interpretation. It presents recently developed methodology to evaluate exposure variability and its effect on the attenuation of risk estimates. It also presents methodology to control for such variability. It uses examples from a recent cohort study of flour millers and bakers. This example shows various characteristics of exposure and demonstrates that various measures of exposure, such as peak and full-shift exposure measurements, are regularly correlated, which has consequences for the analyses of exposure-response relationships. This paper stresses the importance of the recognition and evaluation of exposure variability and its effect on risk estimates and shows that with different exposure grouping schemes, different health risk estimates can be obtained. Quantitative exposure assessment is generally difficult, time-consuming and expensive and many issues need to be taken into account, but it can be rewarding and has become an absolute necessity for many occupational epidemiology studies. Evaluation of components of exposure variance is absolutely necessary. Exposure variability could lead to serious attenuation of risk estimates.