Il ritorno di Bertoldo. Il campionamento multi-stage a raggio d’azione

Abstract

The article proposes a new multi-stage sampling procedure in face to face surveys at national and regional level. This proposal foresees to assign to each interviewer a range of action where to realise their interviews, thus widening their concerning area, which isn’t anymore identifiable with a sampling point (often corresponding to the interviewer’s residence municipality), but with a sampling area where n municipalities are located and may be reached by the interviewer. The idea is to pass from an interviewer network, developed by market research agencies and constituted by single sampling points, to a network formed by sampling areas. This change of perspective allows the contractor to have more control when choosing the area and the municipalities where to submit the questionnaire. This method is therefore able to contrast eventual incorrect behaviours in data collecting procedures such as, for example, the development of «professional interviewees» networks (friends and acquaintances who steadily collaborate with them), created by some interviewers. The main objective of this proposal is to give to the contractor/researcher, who has a relationship with the market research agency, an active role in the sampling definition and realisation, offering statistical solutions (evaluation of sampling networks and estimation of the optimal size of the range of action) in order to reach a higher data quality. The aim is to adopt a disenchanted and critical point of view towards a certain sampling oratory, where terms like « representative sample» and «random sample», which are often used in publications of different survey agencies, have been transformed (many times for economic reasons) in lexical habits which uncritically compose the repertoire of the field research.