A number of recent studies have made use of multivariate statistical methods to seek to identify locally elevated element concentrations in marine sediments. These anomalies are usually attributed to anthropogenic contamination. Multivariate statistics can be a powerful tool in discerning patterns in large collections of data. Studies of sediment geochemistry do not however always make use of the most appropriate statistical methodology. In consequence, the effects of sediment mineralogy are often confounded with effects of contamination. To avoid this, data need to be analysed in ways which take account of the underlying sedimentology. Suggestions are made as to how this might be done.