7.3.2. Loaded questions

Questions can often be framed in such a way that the respondent is guided towards selecting a particular response, even when that response does not reflect the true state of affairs of interest to the investigator. Box 7 gives an example.

Box 7.  Example: Case study: Experience in a Scottish survey.

An example is provided by a question used in a recent Scottish survey in which respondents were asked in what year they had first become aware of varroa infestation of their colonies. The question was intended to discover how far in the past it was when this parasite had first been detected in that area of the country, since there are still remote areas of Scotland where it has not yet been found.  However some newly established beekeepers interpreted this as meaning that they were expected to have personally observed the parasite, and so were inclined to respond that the parasite had “not yet been found” - a biased answer leading to an over-optimistic interpretation of the extent of the parts of the country which were still free of varroa.

Critical analysis of the original questions for possible loading, and careful analysis of pilot survey results, with subsequent revision of the questionnaire where necessary, are essential.