Dear MARMAM'ers, We invite you to read our new publication on bottlenose dolphin social networks. We studied a variety of weighted metrics (centrality, transitivity, affinity) and argue that many so-called "network" metrics are highly redundant to simple individual-attributes, and are not informative of dolphin social structure. We also compare how social networks' weight information (strength of connections) and binary information (who is connected to who) are differently important for inferences about community-partitioning and network structure.
Rankin, RW, J Mann, L Singh, EM Patterson, E Krzyszczyk, and L Bejder. 2016. The role of weighted and topological network information to understand animal social networks: a null model approach. Animal Behaviour 113:215–228. DOI:10.1016/j.anbehav.2015.12.015 ABSTRACT: Network null models are important to drawing conclusions about individual- and population-(or graph) level metrics. While the null models of binary networks are well studied, recent literature on weighted networks suggests that: (1) many so-called ‘weighted metrics’ do not actually depend on weights, and (2) many metrics that supposedly measure higher-order social structure actually are highly correlated with individual-level attributes. This is important for behavioural ecology studies where weighted network analyses predominate, but there is no consensus on how null models should be specified. Using real social networks, we developed three null models that address two technical challenges in the networks of social animals: (1) how to specify null models that are suitable for ‘proportion-weighted networks’ based on indices such as the half-weight index; and (2) how to condition on the degree- and strength-sequence and both. We compared 11 metrics with each other and against null-model expectations for 10 social networks of bottlenose dolphin, Tursiops aduncus, from Shark Bay, Australia. Observed metric values were similar to null-model expectations for some weighted metrics, such as centrality measures, disparity and connectivity, whereas other metrics such as affinity and clustering were informative about dolphin social structure. Because weighted metrics can differ in their sensitivity to the degree-sequence or strength-sequence, conditioning on both is a more reliable and conservative null model than the more common strength-preserving null-model for weighted networks. Other social structure analyses, such as community partitioning by weighted Modularity optimization, were much less sensitive to the underlying null-model. Lastly, in contrast to results in other scientific disciplines, we found that many weighted metrics do not depend trivially on topology; rather, the weight distribution contains important information about dolphin social structure. Download the free PDF here: http://authors.elsevier.com/a/1SWvl_4tkzCKe Visit on Mendeley: http://mnd.ly/1QwiaR3 Visit on ResearchGate: http://www.researchgate.net/publication/294088685_The_role_of_weighted_and_topological_network_information_to_understand_animal_social_networks_a_null_model_approach Import the following bibtex into your reference manager: @article{rankin_networks_2016, title = {The role of weighted and topological network information to understand animal social networks: a null model approach}, author = {Rankin, Robert W. and Mann, Janet and Singh, Lisa and Patterson, Eric M. and Krzyszczyk, Ewa and Bejder, Lars}, journal = {Animal Behaviour}, year = {2016}, pages = {215--228}, volume = {113}, doi = {10.1016/j.anbehav.2015.12.015}, url = {http://www.sciencedirect.com/science/article/pii/S000334721500456X}, keywords = {bias, bottlenose dolphin, community structure, maximum entropy, network topology, social network} } -- "You could give Aristotle a tutorial. And you could thrill him to the core of his being ... Such is the privilege of living after Newton, Darwin, Einstein, Planck, Watson, Crick and their colleagues." -- Richard Dawkins
_______________________________________________ MARMAM mailing list [email protected] https://lists.uvic.ca/mailman/listinfo/marmam
