1970 HUNGARIAN CENSUS OF POPULATION Information on the data collection and processing (1977)

I. THE COMPREHENSIVE REWIEV OF THE DATA COLLECTION AND PROCESSING - 6. Sampling techniques applied in the census

For the sake of correction, a reserve has been provided in the course of the sample design of 1 per cent representativity: namely, by having had the data copied referring to each 1 000^ flat or each 500

tJ

linhabitant of the institutional household. Consequently, they have formed a random sample. The selection of the flats included or omitted also took place by using a table of random numbers. The method used is fully justified, if the distribution of the flats according to the number of inhabitants in the sample during the correction procedure is identical to the distribution of the total number of flats. In this case, the distribution of the flats in the sample comes closer to the distribution of the population after the correction than before. In order to perform this operation, the number of flats having different number of inhabitants to be included to or removed from the sample had to be fixed during the procedure of the correction. The mathematical principles of the correction are presented in the 2°°' volume of the "Cennd sus of 1970" (pp. 226-227.). A more detailed expounding of the method is to be found in the 2 volume of the Census of 1960 (pp. 152-154.) where the operation is also numerically illustrated. The correction procedure was different from the techniques applied in 1960 in the respect that it was made by means of distinction according to communities and towns - and not according to counties as before. In addition to the towns having county rank, the more populous towns (over 60 000 inhabitants) and the districts of Budapest also formed separate units. In this manner, the operation had to be carried out in 71 territorial units. The difference between the hundredth part of the preliminary results and the sample of 1 per cent representativity amounted to 54 flats (private) and 220 persons ("present population"). These figures are smaller than the differences which had shown themselves in 1960, (- 195 dwellings and - 540 persons), but considering that corrections were made in three times as many territorial units, a greater part of the dwellings had to be included to or removed from the sample. (Namely, among the 71 territorial units, in 37 cases a positive, in 34 cases a negative deviation presented itself in the number of the persons, which became equalized, however, on nation-wide level. As far as the number of flats is concerned, a deviation presented itself in 35 territorial units; for the most part, in negative direction as expected. ) The checking up of the population being temporarily present or absent as reflected by the sample In order to check up the registration of persons having double residence, on the purpose of completing the registration procedure of the persons being temporarily present or absent, the method of the "control sheets", as described in a separate chapter, has been used. The design of the 1 per cent sample took place prior to the activity of control which called for its completion by the supplementary i. e. corrected sheets. As far as private dwellings are concerned, this operation took place by means of the accordance of the flats included to the 1 per cent sample with the full-scale registration. The sample was completed by the missing questionaries of the persons being absent, and the corrections occured in several categories of the population have been carried out accordingly (as for instance if a person has been registered as present or absent in both places.). The missing persons of the institutional households have been completed by random selection, according to the difference presenting itself between the full-scale survey and the sample results. 50