Controllo di qualità dei dati: errori non campionari

Non-sampling errors (NSE), joining with sampling errors, contribute in determining ‘Total Survey Error’. NSE arise from the combination of elements which could potentially distort measurement procedure, even the most accurate one. The complex set of measuring operations in any sample survey, in fact, entails the involvement of several actors (e.g. interviewers and data enterers) that operate in a not reproducible and often not fully testable way, so affecting the results. NSE can be generated at each stage of the research distorting inevitably the final estimates. Some scholars proposed several classifications of NSE based on different criteria. In this study, three types of NSE are reviewed: list errors, non-response errors and measurement errors. Data are drawn from a study about the characteristics of the tourism demand in Sicily.