If different laboratories using thermionic mass spectrometry determine in routine operation a 235U concentration of ≈0.7% in the same sample material, their measurement results deviate by at least 0.8% (relative) in ≈50% of the cases. The application of gas mass spectrometry reduces this deviation to 0.1 or 0.05%, the more favorable value being obtained by measurements in which the same reference material is used by all laboratories. These results were obtained by application of an empirical evaluation procedure that, contrary to the usual statistical tools, is not restricted to homogeneous data material. The data were taken from two interlaboratory evaluation programs performed recently on the isotopic abundance determination of 235U in uranium hexafluoride by mass spectrometry.