In recent years, researchers have begun to use corticosteroid metabolites in feathers (fCORT) as a metric of stress physiology in birds. However, there remain substantial questions about how to measure fCORT most accurately. Notably, small samples contain artificially high amounts of fCORT per millimetre of feather (the small sample artefact). Furthermore, it appears that fCORT is correlated with circulating plasma corticosterone only when levels are artificially elevated by the use of corticosterone implants. Here, we used several approaches to address current methodological issues with the measurement of fCORT. First, we verified that the small sample artefact exists across species and feather types. Second, we attempted to correct for this effect by increasing the amount of methanol relative to the amount of feather during extraction. We consistently detected more fCORT per millimetre or per milligram of feather in small samples than in large samples even when we adjusted methanol:feather concentrations. We also used high-performance liquid chromatography to identify hormone metabolites present in feathers and measured the reactivity of these metabolites against the most commonly used antibody for measuring fCORT. We verified that our antibody is mainly identifying corticosterone (CORT) in feathers, but other metabolites have significant cross-reactivity. Lastly, we measured faecal glucocorticoid metabolites in house sparrows and correlated these measurements with corticosteroid metabolites deposited in concurrently grown feathers; we found no correlation between faecal glucocorticoid metabolites and fCORT. We suggest that researchers should be cautious in their interpretation of fCORT in wild birds and should seek alternative validation methods to examine species-specific relationships between environmental challenges and fCORT.