2.4. Appropriate randomization of study organisms for laboratory cage experiments involving adult workers
When designing studies it is crucial that experimenters avoid bias when choosing study subjects. Workers, for example, can exhibit significant genetic variation for expression of mechanical, physiological, immunological, and behavioural responses used in disease resistance (Evans and Spivak, 2010). This diversity can occur among workers in the same colony or among honey bees from different geographic regions. Additionally, timing and method of collection, as described in section 4, can also have a significant influence on results. Because of this, careful consideration is needed when choosing colonies from which to collect experimental honey bees. To promote a repeatable investigation that is representative of a honey bee population in question, workers should be collected from as many, and as diverse a set of, colonies as possible. It is generally recommended to randomly mix workers from all source colonies among all cages during a study to minimize potential colony-level effects on experimental results. Refer to the BEEBOOK statistics paper by Pirk et al. (2013) for details on determining number of colonies from which to source individuals and for how to properly randomize individuals and cages for experiments, as well as section 4 for choosing and obtaining workers for experiments.