Results given by analytical chemistry must be of recognized quality. This quality is obtained from a correct estimation of uncertainties affecting the results. But in our domain, the strength of habit, the lack of teaching, and the particular difficulties linked to chemical metrology: absence of standards and reference materials, difficulties in traceability and sampling... have led to some isolation of chemists and misuse of statistical tools and associated vocabulary. Owing to the development of regulations and quality assurance, our results must be guaranteed and given under understandable manner. Le goal of this paper is to give references of the tools allowing a good estimate of uncertainties and of performance characteristics of analysers, especially basic books of statistics for chemists and Web sites dealing with chemical metrology. The best tools can give bad results if they are not used properly, so we will describe some misuse of statistical tolls and their specific vocabulary, in order to sensitize the reader on the fact the statistics are not a "black box" in which we can enter doubtful data and exit quality.