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

I. THE COMPREHENSIVE REWIEV OF THE DATA COLLECTION AND PROCESSING - 11. The control of the reliability of the census

•the returns of the post-enumeration control of a criterion under survey. The items taken up identically will be situated naturally in the transversal rubrics of the table, while the cases in which the criteria determined by the census and the post-enumeration control differ will be situated above or below the diagonal. Thus the table shows also the dispersion of the responses. Machine processing made possible the determination of the indexes of (partly mathematical) evaluation without lengthy manual calculation. The international literature of the examinations describes the mathematical methodology in depth, however, here only the presentation of the most important indexes necessary for the evaluation of the results is possible. The following rates and indexes were use d in the analysis: - The percentage distribution of the criterion according tothe census and the post-enumeration control; the difference between the two percentage rates. - The number of items identically grouped according to the results of the two censuses (they are situated on the diagonal of the table mentioned before) and their percentage proportion. - Gross errors: the total number as well as the percentage proportion of the items erroneously to be found in, or missing from, the rubrics of the criterion concerned. It must be remarked that the foregoing presupposes that the deviations between the two surveys are owing to the census while the post-enumeration control is without errors. In similar examinations abroad the two operations are frequently considered as surveys in similar circumstances and half of the deviations are considered as owing to the census itself. It seems justified to attribute the greater portion - even if not the totality - of the errors to the census because of the circumstances described in the information on the post-enumeration survey in Hungary (statistician censustakers, inspection of documents in part, occupation data obtained from employers etc. ) - Net error: the difference between the items erroneously to be found in, and missing from, the rubrics of the criterion under survey. Namely not the gross error but the net difference is to be taken into consideration in the census returns. The figures of both the gross and the net error may be compared to the data of the criterion in question on the one hand and to the total number of the items (persons) comprised by the sample on the other. The latter plays a more important part in mathematical examinations, however, comparison with the number of the criterion under survey is more descriptive, for this reason this is used in most instances. It follows from the definitions above that as a matter of fact that the "gross error" includes the errors found twice, once as items erroneously missing from the rubrics and secondly as items erroneously to be found in other rubrics. The net error, however, only expresses the net balance of errors as well as + deviations. The gross and net errors can be formulated too. If a = is the number of identically entered items, b = is the number of items erroneously entered in the rubrics, c = is the number of items erroneously left out from the rubrics, n = is the total number of items (individuals) in the sample, 96