[Statistical analysis of the results of radionuclide studies using small computers]

The paper is concerned with a method of the choice of an optimum criterion of difference with regard to the law of the distribution of samples, degree of their correlation and the limits of application of the criterion. Student's t-criterion is usually used for comparison. Examples are cited to the effect that the formal use of Student's t-criterion disregarding the limits of its applicability can sometimes result in errors in statistical conclusions. The authors give an algorithm to process the results with the help of a small-size computer.