Statistical methods in concrete research

Summary The application of statistical methods has helped greatly in planning and interpreting concrete research at the Road Research Laboratory and, as no statistical text-book deals specifically with the problems of concrete technology, this paper aims to bring such methods to the attention of research workers elsewhere. The assessment of variation in test results is discussed and some useful statistical processes are introduced. These include a test for the difference between two sets of results, the method of analysing the causes of variation in test data and the fitting of equations to experimentally determined relationships between two variables. An appendix gives a method for fitting a Gaussian distribution to a series of test results. A subsequent paper will deal with the design of experiments in concrete research.