Читать книгу Handbook of Web Surveys - Jelke Bethlehem - Страница 62

1 1 Fanney Thorsdottir and Silvia Biffignandi presented and discussed their flowchart at a WG3 task force (TF10) meeting organized by the Webdatanet COST Action (IS1004) while chairing and co‐chairing WG3 and leading TF10 on General Framework for Error Categorization in Internet Surveys. A final draft of the flowchart was presented at the Webdatanet conference (Salamanca, May 26–28, 2015). An extension for the mode selection has been presented at the Total Survey Error Conference (Baltimore, September 2015).

2 2 The flowchart is useful not only for specialized survey research organizations but also for smaller organizations or individual researchers. Any surveyor, even for simple and rather small surveys, no matter if he is in favor for more or less sophisticated methodological solutions, should refer to it. He should go across each step and decision to run a good survey. Large statistical organizations (like National Statistical Institutes) follow a more complex survey process because of the different context, laws, regulations, and objectives governing the survey process. Therefore, a specific flowchart exists in specialized documents for official statistics.

3 3 It should be advised that it is not possible to process non‐probability‐based survey data using statistical inference methods to estimate population. Comments about probability versus non‐probability‐based samples and the exclusive advantages of probability‐based samples are in Chapter 14.

4 4 As stated above, if some units of the target population are not on Internet and/or an adequate online sampling list is available, alternative sampling frames could be available (for instance, ordinary mail addresses), and the sample could be drawn from this sampling list. In this case, sampled units that are not on the Internet could receive the invitation letter either by e‐mail or by ordinary mail.