In this study, the uncertainties associated with through-diffusion, the steady-state method used for determination of the pore diffusion coefficient (D p ), were analysed and evaluated. The diffusion of strontium through the sample of compacted bentonite was studied to evaluate various sources of uncertainties and compare their contribution with uncertainty of D p . Different diffusion experiment arrangements were used to test the effect of solution accessibility to the filter membranes. Three main general factors influencing the uncertainty of D p at given conditions were identified: (a) the compacted sample preparation and its physical properties; (b) the properties of the diffusion cell; and (c) the strontium concentration analyses. Some of the identified sources can be treated as fixed with limited extent (e.g. the deformation of components of the diffusion cell), but some sources must be treated as variable with wide extent depending on the experimental conditions (e.g. the physical properties of the compacted bentonite sample -moisture, density, porosity). The variable uncertainty sources have to be taken into account, especially in repeatability and reproducibility tests. It was concluded that each of the diffusion cells with compacted bentonite sample represents a unique diffusion system. It is preferable to present the results of diffusion experiments as a range of values rather than as one average value with an uncertainty.