scholarly journals Dominance network structure across reproductive contexts in the cooperatively breeding cichlid fish Neolamprologus pulcher

2015 ◽  
Vol 61 (1) ◽  
pp. 45-54 ◽  
Author(s):  
Cody J. Dey ◽  
Q.Y. Joanne Tan ◽  
Constance M. O’Connor ◽  
Adam R. Reddon ◽  
J. Ryan Caldwell

Abstract While a large number of studies have described animal social networks, we have a poor understanding of how these networks vary with ecological and social conditions. For example, reproductive periods are an important life-history stage that may involve changes in dominance relationships among individuals, yet no study to date has compared social networks of dominance interactions (i.e. dominance networks) across reproductive contexts. We first analyzed a long-term dataset on captive social groups of the cooperatively breeding cichlid Neolamprologus pulcher, and found that eviction events were significantly more common around reproduction than expected by chance. Next, we compared the structure of dominance networks during early parental care and non-reproductive periods, using one of the first applications of exponential random graph models in behavioral biology. Contrary to our predictions, we found that dominance networks showed few changes between early parental care and non-reproductive periods. We found no evidence that dominance interactions became more skewed towards larger individuals, became more frequent between similar-sized individuals, or became more biased towards a particular sex during parental care. However, we did find that there were relatively more dominance interactions between opposite-sex dyads in the early parental care period, which may be a by-product of increased sexual interactions during this time. This is the first study in behavioral ecology to compare social networks using exponential random graph modeling, and demonstrates a powerful analytical framework for future studies in the field.

2016 ◽  
Vol 7 (1-2) ◽  
pp. 29-54
Author(s):  
Yeaji Kim ◽  
Leonardo Antenangeli ◽  
Justin Kirkland

AbstractExponential Random Graph Models (ERGMs) are becoming increasingly popular tools for estimating the properties of social networks across the social sciences. While the asymptotic properties of ERGMs are well understood, much less is known about how ERGMs perform in the face of violations of the assumptions that drive those asymptotic properties. Given that empirical social networks rarely meet the strenuous assumptions of the ERGM perfectly, practical researchers are often in the position of knowing their coefficients are imperfect, but not knowing precisely how wrong those coefficients may be. In this research, we examine one violation of the asymptotic assumptions of ERGMs – perfectly measured social networks. Using several Monte Carlo simulations, we demonstrate that even randomly distributed measurement errors in networks under study can cause considerable attenuation in coefficients from ERGMs, and do real harm to subsequent hypothesis tests.


2007 ◽  
Vol 29 (2) ◽  
pp. 173-191 ◽  
Author(s):  
Garry Robins ◽  
Pip Pattison ◽  
Yuval Kalish ◽  
Dean Lusher

Author(s):  
Qiuchang Cao ◽  
Li Liao ◽  
Keith Leverett Warren

Purpose To analyze networks of social interactions between the residents of a therapeutic community (TC) for women and the way, in which such interactions predict the discussion of issues that arise in treatment. Design/methodology/approach In total, 50 residents of a corrections-based TC for women were surveyed on the peers with whom they socialized informally, shared meals, shared letters from home and discussed issues that arose in treatment over a 12 h period. The data were analyzed using exponential random graph models (ERGM). Findings Reciprocity occurred in all networks while transitivity (a tendency of two residents who are connected to both connect to a third peer) occurred in all networks measuring informal social interactions. When controlling for reciprocity and transitivity, residents avoided spending social time or sharing meals with the same peers. There was no evidence of homophily by race, age or years of education. Homophily by entrance time and case manager occurred in social time. Case manager homophily occurred in the discussion of treatment issues but disappeared when controlling for social time and sharing letters from home. Research limitations/implications Social networks in this TC arise from factors endogenous to the TC itself. It should be possible to determine the characteristics of optimal social networks in TCs. External validity is limited. Practical implications It should be possible to intervene to optimize the social networks of TC residents. Originality/value This is the first ERGM analysis of both informal and formal interactions in a TC.


2007 ◽  
Vol 29 (2) ◽  
pp. 192-215 ◽  
Author(s):  
Garry Robins ◽  
Tom Snijders ◽  
Peng Wang ◽  
Mark Handcock ◽  
Philippa Pattison

Sign in / Sign up

Export Citation Format

Share Document