Missing data are inevitably ubiquitous in experimental and observational epidemiological research. Nevertheless, despite a steady flow of theoretical work in this area, from the mid-1970s onwards, recent studies have shown that the way partially observed data are reported and analysed in experimental research falls far short of best practice (Wood et al. 2004; Chan and Altman 2005; Sterne et al. 2009). The aim of this Chapter is thus to present an accessible review of the issues raised by missing data, together with the advantages and disadvantages of different approaches to the analysis.