Standard survey methods for estimating colony losses and explanatory risk factors in Apis mellifera

Authors: Romée van der Zee, Alison Gray, Céline Holzmann, Lennard Pisa, Robert Brodschneider, Róbert Chlebo, Mary F Coffey, Aykut Kence, Preben Kristiansen, Franco Mutinelli, Bach Kim Nguyen, Adjlane Noureddine, Magnus Peterson, Victoria Soroker, Grażyna Topolska, Flemming Vejsnæs, Selwyn Wilkins.

Table of contents

1. Introduction
2. Objectives and case definitions

   2.1. Objective of epidemiological studies on honey bee colony losses
   2.2. Application of definitions associated with honey bee colony loss
   2.3. Case definitions
3. Data collection methods
   3.1. Choosing the method of data collection
   3.2. Available data collection methods with advantages and disadvantages

      3.2.1. Surveyor administered questionnaires Face-to-face interviews Telephone interviews
      3.2.2. Self-Administered questionnaires
… Postal or email survey
… Internet survey
… Questionnaire published in beekeeping journals Questionnaires disseminated during meetings
   3.3. Data validity and accuracy
4. Quality issues in surveys
   4.1. Errors

…   4.1.1. Coverage error
…   4.1.2. Sampling error
…   4.1.3. Measurement errors
…   4.1.4. Non-Response errors
…   4.1.5. Errors caused by selection bias
…   4.1.6. Processing errors
   4.2. Effort and costs in data accumulation
   4.3. Issues of anonymity and ethical approval

5. Coverage
   5.1. Effective coverage
   5.2. Potential bias of the survey sample
   5.3. Identifying the target population
6. Sampling
   6.1. Random and non-random sample selection methods

…   6.1.1. A census
…   6.1.2. Random sampling
…   6.1.3. Recommended approach for random sampling
      6.1.4. Non-random methods
   6.2. Need for and use of a sampling frame in random sampling
   6.3. Availability of a sampling frame
   6.4. Sources of sampling frames appropriate for different target populations
7. Questionnaire design
   7.1. Completeness of the questionnaire
   7.2. Appropriate designs for different sampling methods
   7.3. Common problems to avoid in questionnaire design

…   7.3.1. Ambiguity of interpretation
…   7.3.2. Loaded questions
…   7.3.3. Questions on sensitive issues
   7.4. Questionnaire design for minimisation of measurement error and ease of analysis
   7.5. Need to limit data sought, for a high response rate and accurate measurement
   7.6. Problems of multi-lingual/multi-cultural questionnaires
   7.7. Testing survey questions: importance of pilot studies
   7.8. Example of a standardized questionnaire on colony losses
8. Response rates

   8.1. Use of incentives and reminders to improve response rates
9. Choice of sample size
10. Analysis of survey data
   10.1 Assessing data quality

      10.1.1. Dealing with missing data
   10.2. The use of weighting in statistical analysis
   10.3. Elementary analysis

      10.3.1. Descriptive analysis
      10.3.2. Loss calculations and Confidence Intervals
      10.3.3. Loss rate per factor including stratification on the operation size
   10.4. Advanced analysis; identification of risk factors by logistic regression
      10.4.1. Logistic regression
 … 10.4.2. Dispersion in statistical models
 … 10.4.3. Multilevel analysis
      10.4.4. Software for logistic regression models
      10.4.5. Example of advanced analysis
11. Conclusions
12. Acknowledgements
13. References