| dc.description.abstract | In the course of the survey on medical care utilization in Poland, abundant materials have been collected concerning the process of data gathering. The materials embrace: observation reports on interviews; personal data on interviewers; their written utterances about the questionnaire and interviews; results of a personality test given to interviewers; extensive fill-in questionnaires attached to each interview, concerning respondents’ behaviour; interviewers’ answers to certain questions of the questionnaire proper as well as their answers concerning their expectations with regard to the distribution of respondents’ answers to some questions, etc.
On the initiative of the Polish research group it was decided to elaborate these materials according to the preliminary project presented in the article.
The main emphasis of the project is on the problem of reliability of the data obtained in scheduled interviews as well as on the sources of errors. The existence of errors will be ascertained directly, i.e. through external checking of some informations as well as by means of an internal checking. The external checking relates only to the informations about medical visits which will be checked against reliable records kept in medical care dispensaries. Internal checking consists in juxtaposing informations contained in answers to various questions.
Another method of checking if an information is or is not burdened with an error consists in finding out whether the distribution of answers is related to the existence of various factors involved in the process of data collection. When there is a statistically significant non-spurious correlation between the existence of a certain factor and the kind of answers obtained, it may be assumed that informations contained in such answers are burdened with an error.
The latter type of analysis points to sources of errors inherent in various factors of a research process. Among these factors the following ones will be given special attention: (1) traits and attitudes of respondents as seen by interviewers, (2) traits and opinions of interviewers, (3) various elements of interview situation, (4) difficulty, delicateness and vagueness of questions. In addition, the influence of selection of interviewers and efficiency of the control of their work, will be also studied.
Various methods of error detection, with regard to some informations, will be applied simultaneously. Due to the abundance of data, it will be also possible to perform a combined analysis of the influence of various factors upon the research process and its deformations. The purpose of these methodological analyses will be a global evaluation of data reliability and the elaboration of ways to avoid errors in socio-medical surveys. | |
