A new way of assessment of the state of a granular mixture is presented in this paper. It is based on the method of statistical reduction and it enables a more precise assessment to be made of the state of a grain mixture as compared with other methods, especially when the mixture represents a large population. The new method is illustrated by a computational example.