7.3.1. Ambiguity of interpretation

If respondents can interpret a question in various ways, the returns made will not be easy to interpret and the analysis can become difficult or impossible to conduct. Box 6 gives an example. The way to minimise ambiguity is, first of all, to ensure that the early drafts of a questionnaire are always criticised by an independent evaluator before they are used, and once all obvious ambiguities have been removed, to pilot the questionnaire (see section 7.7.) in order to try to detect any remaining problems with the questions.

Box 6.  Example: Colony management in Canada.

An example is provided by recent COLOSS surveys in which beekeepers were asked about increases and decreases during a certain timeframe. All Canadian respondents who reported increases or decreases during the defined wintering period were contacted to verify whether such changes truly reflected the dynamics of the wintering population. Invariably, these changes reflected spring-time activities (typically splitting colonies), where these activities could occur in warmer areas of the country prior to the defined end date of the wintering period. Moreover, these changes were not reflected in total colony counts at the end of the wintering period. The question was clear about the timeframe, but a substantial number of beekeepers ignored this information (van der Zee et al., 2012).