Niektóre problemy metodyki prób reprezentujących publiczność pism tygodniowych
MetadataShow full item record
A sample of readers buying a weekly at news stands provides the sole technique to reach this kind if elite publics. Such samples are burdened, however, with a systematic bias because those persons who buy a magazine more often have a better chance to find themselves in the sample. Besides that, various other circumstances (like the specific content of a given number of the magazine, timing of the survey, etc.) may also bias the sample. In the 1961 survey on the readership of four literary weeklies („Przegląd Kulturalny”, „Nowa Kultura”, „Życie Literackie”, „Polityka”) a procedure of statistical weighting of sub-groups of readers with a differential frequency of reading was employed, its aim being to reduce the percentage of regular readers (viz. those reading more frequently) who were, over-represented in the sample. Employed were also procedures of weighting statistical distributions of other traits (e. g. sex) highly correlated with the reading frequency, as well as some methods, based on data from such surveys, of making estimates of an absolute number of all readers living in a region under investigation. A model of procedures aiming at the empirical verification of assumptions underlying this kind of statistical weighting is presented.