| dc.description.abstract | To discover errors that diminish the cognitive value of data obtained in questionnaire-based surveys, two procedures may be applied: (1) external analysis and (2) internal analysis. The first one consists in the confrontation of the answers given by respondents in a questionnaire with information on the same topics obtained from sources independent from the researcher and considered to be objective. The second procedure uses the existing logical relationships between answers to questions which occur at the very moment of the construction of a questionnaire. As a result of these relationships, co-existence of certain answers in a filled-out questionnaire leads to the contradiction. In the internal analysis, by an error is understood the situation of co-existence of contradictory sentences (answers).
In a questionnaire interview three elements may be distinguished: (1) interviewer, (2) respondent, (3) the set of questions in the questionnaire all of them entangled in the psycho-social situation of interview. Methodological innovations introduced and a methodological evaluation should relate to each of them. The internal analysis consists in the methodological evaluation of a questionnaire treated as a carrier of a communication process.
The principles of internal analysis are implicitly observed in the current research practice in the form of a check on information obtained in the course of interviewing. The checking is done by the interviewer as well as, at the later stage, by the interviewer’s supervisor. This is reflected in the use of filtering questions, control questions and instructions for interviewing. Nevertheless, the analysis of logical relationships between answers to questions in a questionnaire constitutes a higher level of methodological consciousness for it takes into account more complicated situations, such as, for example, enthymematic relationships; in addition, it makes use of logical calculus.
Using the classical sentential calculus, the author introduces the concepts of thesis, tautology and counter-tautology which serve to define two basic logical relationships: the conclusion and the contradiction. Then he undertakes an attempt at the typology of contradiction relationships between answers to questions in a questionnaire. We have to do with a direct contradiction in such a situation only where questions and answers are identical; otherwise it is necessary to accept certain additional assumptions. These assumptions may be based on definitional terminological conventions as well as on empirical regularities. Among the latter, of especial importance for socio-medical research are interrelationships discovered in medical science.
Quantitative results of analysis of questions in the socio-medical questionnaire are presented: 455 possible pairs of contradictory sentences have been identified, discussed and exemplified within the framework of the above-mentioned typology. Finally, prospects for treatment of the results of internal analysis together with the results of other methodological studies are discussed. | |
