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

I. THE COMPREHENSIVE REWIEV OF THE DATA COLLECTION AND PROCESSING - 8. The electronic data processing of the census of 1970

2 per cent were qualified as having more than 4 per cent of errors; 28 per cent were qualified as having 1-4 per cent of errors; 70 per cent were qualified as having 0, 7 per cent of errors, on the average. Punching, control of punching, testing for completeness The recording on the primary data carrier took place centrally. 4-4 persons belonging to the census staff performed duties in two shifts in order to correct the errors made during the data processing. One of them generally cared for the correction of the indecipherable code-numbers. Others had other duties to perform. By means of sorters provided with counters they performed the testing procedure of completeness, i. e. they checked whether the census district has as many punch cards at disposal as persons or flats are belonging to it. The checking of the absurd code-number of several important criteria, as for instance of the basis criteria took also place here. These criteria served as the starting point of the automatic correction of errors. It was the last opportunity for their correction, i. e. to get in possession of adequate information needed for the proper coding, by means of reading the questionary fixing the record. Automatic correction of errors Simultaneously with the conclusion of the punching and the test of completeness, an "escorting punch card of the census district" was prepared which showed the number of entries and other characteristics of the census districts. Concurrently with the recording of the information on magnetic tapes, the "escorting punch card" has checked, by census districts, whether the number of records prepared is conform to the number of the punched cards. This fact was accordingly recorded (on the tables) and accepted. The machine has checked not only the completeness but the continuity of the seriaj. numbers ensuring the order of the hierarchy. In so far as it found serial numbers exceeding the intervals, it has corrected them according to the lacking serial number. The number of the cases occured was negligible: by technical units comprising 40 000 records but 10-15 erroneous serial numbers could be found. When planning the programme of the automatic correction of errors, we aimed at the checking of the validity of the data originating from the survey and the coding, as well as at eliminating the logical contradictions of the criteria. By means of confronting the different aspects and views of many hundreds of people belonging-to the coding staff with the sound logics of the propositions, the unambiguous interpretation of the notions, further the uniform correction of the errors, performed by the computer - a qualitative difference between the two techniques could be defined which was unmeasurable hitherto, in objective manner. The large number of the tests and corrections performed by the machine within this short term, would be impossible to be carried out by manual techniques. It may be attributed to this fact that the computer was capable of clearing up rather large number of errors and could perform the respective corrections. (As far as the amount of corrections and controls perst formed in several sections are concerned, they are to be found in the 1 Annex, while their proportions in the 2

n

d.

). 79