Abstract
Recommendation for number of valid digits in presented results of research in the field of a health and social sciences is given. In the case of quantitative data it depends on the resolution of the method used and the number of valid digits should be the same as the number of valid digits of input data.
If an ordinal scale of a measurement is used, only integers should be used. Special attention deserves results, which are ratios.
In this case the number of valid digits implies the size of a sample. Measures of location and variability should correspond to the type and distribution of data.
In the case that it is possible to assume their normality, the proper characteristics are arithmetic mean and standard deviation, in all other situations median, mode and different ranges are better.