Estimating intakes of chemicals in food requires residue data, giving the concentration of the chemical in food, and consumption data for foods which could contain the chemical. Residue data are often restricted to only a few foods and even then may not be representative of the foods of interest due to limitations in the sampling or analytical procedures. Obtaining reliable consumption data can also be complex. People normally eat a wide range of different foods, which may each contain a range of concentrations of the chemical of interest. Both sets of data can be combined using one of several approaches. This paper reviews the nature and relevance of the factors affecting the collection of these two data sets and the procedures available for combining them to estimate consumer exposure. Considerable expertise and judgment are necessary to obtain an estimate of exposure that is robust enough for risk management purposes.