Naar aseptische en realistische methoden in de toegepaste statistiek

Summary Applied statistics is not pure mathematics but the application of suitable methods of mathematical statistics to practical problems. In choosing the method one should be extremely suspicious in regard to septic methods and give strong preference to realistic methods. Septic methods are those which by their very nature introduce properties alien to the magnitudes concerned e.g. addition of non‐additive variables, differentiation of stochastic variables, stochastic treatment of indeterminate variables, etc. Realistic methods are those based on a preliminary analysis of the practical problem which should bring to light the essential mutually independent variables controlling the outcome of the problem. Undue simplification at the cost of the accuracy required in actual practice should be avoided; especially weak correlations between variables shown by the deductive analysis as being essentially mutually independent should be discarded. On the other hand the mathematical treatment should be as simple and direct as possible. The professional mathematician can become a real danger if he wants to guide the application of mathematical statistics to certain fields of applied science in which he is not sufficiently specialised and experienced. Contrariwise, if he acquires this knowledge in business, the highest positions are open to him.