A QUANTITATIVE COMPARISON OF OBSERVATIONAL METHODOLOGIES FOR STUDIES OF WORKER HONEYBEES
S A KOLMES
In order to determine whether data sets collected by 'focal-animal' and 'instantaneous and scan' techniques are comparable, a single age-cohort of worker honeybees (Apis mellifera) was observed using both techniques. Scan sampling involves recording the first behaviour pattern seen during an observation period for each of a number of individual animals. Focal sampling involves recording all the behaviour patterns performed by an individual animal during an observation period. Statistical analysis showed that the two data-sets were indistinguishable. Data gathered in scan samples are independent in a statistical sense, and a wide variety of hypotheses and statistical tests are compatible with data collected by this method. Frequency data gathered in focal samples suffer from the difficulty of sequential non-independence, which makes them incompatible with many statistical tests. Focal-sample data are nonetheless required in studies addressing behavioural transition frequencies, bouting and duration of behaviour patterns.