Physiological measurements of both stress and sex hormones are often used to estimate the consequences of natural or human-induced change in ecological studies of various animals. Different methods of hormone measurement exist, potentially explaining variation in results across studies; methods should be cross-validated to ensure that they correlate. We directly compared faecal and plasma hormone measurements for the first time in a wild free-living species, the Adelie penguin (Pygoscelis adeliae). Blood and faecal samples were simultaneously collected from individual penguins for comparison and assayed for testosterone and corticosterone (or their metabolites). Sex differences and variability within each measure, and correlation of values across measures were compared. For both hormones, plasma samples showed greater variation than faecal samples. Males had higher mean corticosterone concentrations than females, but the difference was only statistically significant in faecal samples. Plasma testosterone, but not faecal testosterone, was significantly higher in males than females. Correlation between sample types was poor overall, and weaker in females than in males, perhaps because measures from plasma represent hormones that are both free and bound to globulins, whereas measures from faeces represent only the free portion. Faecal samples also represent a cumulative measure of hormones over time, as opposed to a plasma 'snapshot' concentration. Our data indicate that faecal sampling appears more suitable for assessing baseline hormone concentrations, whilst plasma sampling may best define immediate responses to environmental events. Consequently, future studies should ensure that they select the most appropriate matrix and method of hormone measurement to answer their research questions.