Important features of the multivariate approach are discussed, and an extension to this approach is proposed whereby the total uncertainty in site investigation methods due to spatial averaging is assessed prior to its adoption. Results from a site investigation of spatially averaged values of undrained shear strength (S-u) and the corresponding coefficient of variation (COVSu) in Veda sulphide clay were used as a practical illustration of the extended multivariate approach and provide a basis for discussion. The inherent variability and scales of fluctuation for different methods are presented. The study shows the usefulness of the extended multivariate approach for the evaluation of representative values of S-u and COVSu based on results from different methods. It is also a way of implicitly reducing the transformation errors that arise when a property is derived from measurement results. Nevertheless, considerable care must be taken as a much lower COV for one method will have a significant impact on the results.