Investigating structure and temporal scale in social organizations using identified individuals

Studies of individually identified animals can produce substantialdata sets containing information on the structure and temporalscale of social organizations. However, methods of analyzingsuch data are not well established. Important features of asocial organization are revealed by plotting the rate of persistenceof the associations between pairs of individuals over a rangeof time lags (lagged association rate). The consistency of long-termrelationships can be characterized using the rate of associationof pairs of individuals between their first and last observedassociations (intermediate association rate). A hierarchicalseries of models featuring exponentially decaying lagged associationrates may be fitted to these data. This technique retrievedthe essential parameters of five simulated social organizationsand, when used on real data, portrayed the essential featuresof the patterns of temporal change in relationships betweenanimals. The method should be especially useful for analyzingfissionfusion societies containing 10–10, 000 individuallyidentifiable animals.