| Summary: | The text being offered is methodological micro-tribute to the theory of market reserarch, sociological surveys and opinion polling. It analyses results of already evaluated empirical examinations about the nature of interviewing and measurement error streaming from this source. The results are discussed and completed by authors empirical experience gained in full service market research agency on both domestic and international projects. We are able to measure sampling error or, as some researchers say, errors of non-observation, but we have serious doubts about measurement and prevention of those arising in the datacollection phase. Those errors negatively affect the quality of survey results. Whether it is respondents reactions, interviewer tactics or non-respose phenomena, we are not able to foresay its volume and direction in advance. Partially, bacause those effects are interreacting. Results of sociological survey are uneasy to be generalised because of impossibility to compare data collected in diferent cultural, historical and political surroundings. The text aims to define survey quality rather in a way what it is not. The conclusions of the analyses show that it is safer to rely on prevention of the errors rather than on subsequent data processing. Also, validity and reliability schemes must be...
|
|---|
