ISO 15796:2005 specifies generic methods for detecting and correcting bias (systematic errors) of analytical procedures for the analysis of gases, using reference gas mixtures or reference analytical procedures, as well as for estimating the correction uncertainty. The main sources of (and parameters affecting) bias of analytical procedures are instrumental drift (time) and matrix interferences (matrix composition). Moreover, bias normally varies with analyte concentration. ISO 15796:2005 therefore establishes protocols for detecting and correcting drift for an analytical system of limited stability, and for investigating and handling bias of a stable analytical system for a specified range of sample composition. These protocols are intended to be used in method development and method validation studies, either separately or sequentially. ISO 15796:2005 specifies procedures for two options, applicable to systematic effects, as follows: 1) tracing the observed pattern of deviations and correcting for their effect, 2) averaging over their effects and increasing the uncertainty, where normally the first option entails lower uncertainty at the expense of higher effort. For the convenience of the user, the methods specified in ISO 15796:2005 are described for procedures of composition analysis, i.e. procedures for measuring the concentration of a specified analyte in a gas mixture. However, they are equally applicable to measurements of physico-chemical properties of a gas or gas mixture relevant to gas analysis, and translation into this subject field is straightforward.