Enkele statistische aspecten van sociaal onderzoek

Samenvatting

Part I of an article about the statistical aspects of sociological research; it deals with the random sample on the basis of the existing literature. The random sample is evaluated by the criteria of faithful representation ('cleanliness') and efficiency. It is explained why 'typical' data may be inconvenient for a random sample since they may be too small. 'Representativity' in itself is not enough and the quota method may be 'unclean.' To obtain a 'clean' sample there must be agreement on what factors are under investigation and what degree of approximation is sought. 'Nonresponse' must be eliminated as a vitiating factor. Efficiency may be obtained by 'stratified' or '2-stage' sampling. I. Langnas.