11. Conclusions

Estimating colony loss rates reliably depends both on selecting representative samples of beekeepers and also on using suitable methods of estimation. Examining potential risk factors for losses also requires use of suitable statistical methodology. Standardisation of methodology will enable valid comparisons of loss rates to be made across time and between/within countries. Standardisation of terminology avoids confusion and facilitates the required comparisons.

In this manuscript, we have defined terminology associated with colony losses, and have presented the concepts involved in conducting a survey. The latter ranges from choosing the method of data collection to designing the questionnaire and how to select a representative sample and guidelines for choosing the sample size. Practical suggestions and examples are given.

We have examined many of the difficulties of conducting surveys and the important or most likely sources of error in surveys. Being aware of the potential for error makes it more likely that the survey organiser will be careful to avoid practices which are likely to introduce error into a survey, and therefore should achieve a more reliable result. We have also reviewed relevant methods for assessing the quality of the data and for statistical analysis, and have illustrated the more advanced techniques proposed for analysis.