Két cigányvizsgálat

(Kritikai elemzés)

Absztrakt

The paper compares the representative survey of Gypsies conducted by the Institute for Sociology of HAS in 1993-94 and the Gypsy sub-sample of the survey of nationalities conducted vy the Central Statistical Offie in 1993. It shows that the two samples differ from each other to a significant extent in respect of important social indicators. Looking for the causes of these differences the author studies three issues more closely. He considers the consequences of the differences in the sampling methods and shows that though the sampling methods of the two surveys were founded on totally different principles, yet the significant differences of the samples cannot be explained by those variations. Further on, he studies those differences which derive from the different handling of failed questionings. The paper shows that the fact that the data collection of the Statistical Office - due to the nature of the task - was unable to substitute the aborted questionings of Gypsies by additional Gypsy addresses, has significantly contributed to the distortions of the Gypsy sub-sample of the Central Statistical Office. It is also understood that the method of weighting employed by the Central Statistical Office in the case of the Gypsy sub-sample not only does not improve, but it further deteriorates representativity. Finally the last part of the paper tries to check the correctness of the two samples, more exactly the reality of the number of primary school pupils and of the data on school education of the adult population on an external source of data, namely on the time sequences of the school statistics of the Ministry of Culture and Education. This test has also unambiguously supported the greater reliability of the data of the Institute for Sociology.