Gas movement through soils is important for ecosystems and engineering in many ways such as for microbial and plant respiration, passive methane oxidation in landfill covers and oxidation of mine residues. Diffusion is one of the most important gas movement processes and the determination of the diffusion coefficient is a crucial step in any study. Five laboratory methods used for measuring the relative gas diffusion coefficient (D(s)/D(o)) were compared using a loamy sand, a porous media commonly found in agricultural fields and in several engineered structures, such as in landfill final covers. In the absence of macropores, all methods gave rather similar values of D(s)/D(o). Methods allowing the study of microscale variability indicated that the presence of macropores highly influenced gas movement, thus the value of D(s)/D(o), which, near a macropore may be one order of magnitude higher than in regions without macropores. Repacked columns do not allow the study of heterogeneity in D(s)/D(o). Natural spatial variability in D(s)/D(o) due to water distribution and preferential pathways can only be studied in large systems, but these systems are difficult to handle. Advantages and disadvantages of each method are discussed.