The aim of the national survey of radon concentration in Polish flats is to evaluate the possible adverse health effect to the inhabitants of Poland, arising from exposure to radon and its progeny. As people are exposed to radon mainly in houses, evaluation of indoor radon concentration in a number of houses should enable us to assess the average level of exposure to radon of an inhabitant of Poland, provided a suitable population-weighted sampling strategy is chosen. A different strategy would have to be applied if, e.g., the distribution of radon concentration over the area of Poland were to be studied in relation to geological structures. The sample should then contain a large number of houses of typical Polish architecture, located with the same density over the whole area of Poland. Generalization of the results is possiblle only if the set of experimental results constitutes a representative group for the population examined. The studied sample contained 5100 houses, i.e. one detector per about 2200 households in Poland. The representativeness of the sample was evaluated using the statistical c²-test, where the null-hypothesis statistical data were taken from the Polish Yearly Book of Statistics. The following parameters were compared: age of building, distribution of population of cities and villages in Poland, and the distribution of the number of rooms and of the type of sanitary installations in Polish dwellings. It was shown that the studied sample would have to be modified in order to represent a statistically random population-weighted sample.