The Range and Directions of Bias in a Press Questionaire.

This article presents a study on the factors which cause bias in a press questionnaire in Poland. It is a common belief that questionnaires of that type do not furnish data representative of all the readers because the samples of readers who took part in the survey consists of persons who have answered by virtue of autoselection from among all people reading the newspaper with the questionnaire. In attempt at the empirical examination of the range of bias of the material obtained from a press questionnaire was taken up in spring 1964 when the Glos Szcecinski was investigated. Another interesting phenomenon is almost identical percentage of pensioners and retired persons in the questionnaire and representative sample. The bias is smallest relation to the structure of age and the period of residence in Szczecin. The results obtained in this study have allowed for a proper formulation of problems and a few suggestions as to the further directions of analysis which might bring a fuller answer to the question what corrections and what limiting assumptions and methods of calculations must be applied so that a group of readers completing the press questionnaire could be taken as a substitute for the representative sample of readers of a daily paper.