Influence of number of individuals and observations per individual on a model of community structure

Sunga, Julia; Webber, Quinn M. R.; Broders, Hugh G.

doi:10.1371/journal.pone.0252471

Cited by 10 publications

(8 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several studies have suggested that social networks are quite robust to sampling subsets of the population—Silk et al (2015) estimated that studies sampling 30% or more of a population can generate reliable estimates of individuals' social network positions—so GPS devices can theoretically be deployed to characterize even large social networks. Subsequent studies highlighted that repeated sampling and avoiding misidentifications are critical for producing accurate networks (Davis et al, 2018; Sunga et al, 2021)—both are the strengths of GPS tracking data. However, GPS positions are also prone to error and sampling discontinuity (i.e.…”

Section: Social Network Structuresmentioning

confidence: 99%

A guide to sampling design for GPS‐based studies of animal societies

Klarevas‐Irby

Papageorgiou

et al. 2022

Methods Ecol Evol

View full text Add to dashboard Cite

GPS-based tracking is widely used for studying wild social animals. Much like traditional observational methods, using GPS devices requires making a number of decisions about sampling that can affect the robustness of a study's conclusions. For example, sampling fewer individuals per group across more distinct social groups may not be sufficient to infer group-or subgroup-level behaviours, while sampling more individuals per group across fewer groups limits the ability to draw conclusions about populations. Here, we provide quantitative recommendations when designing GPS-based tracking studies of animal societies. We focus on the trade-offs between three fundamental axes of sampling effort:(1) sampling coverage-the number and allocation of GPS devices among individuals in one or more social groups; (2) sampling duration-the total amount of time over which devices collect data and (3) sampling frequency-the temporal resolution at which GPS devices record data. We first test GPS tags under field conditions to quantify how these aspects of sampling design can affect both GPS accuracy (error in absolute positional estimates) and GPS precision (error in the estimate relative position of two individuals), demonstrating that GPS error can have profound effects when inferring distances between individuals. We then use data from whole-group tracked vulturine guineafowl Acryllium vulturinum to demonstrate how the trade-off between sampling frequency and sampling duration can impact inferences of social interactions and to quantify how sampling coverage can affect common measures of social behaviour in animal groups, identifying which types of measures are more or less robust to lower coverage of individuals. Finally, we use data-informed simulations to extend insights across groups of different sizes and cohesiveness. Based on our results, we are able to offer a range of recommendations on GPS sampling strategies to address research questions across social organizational scales and social systems-from group movement to social network structure and collective decision-making. Our study provides practical advice for empiricists to navigate their decision-making processes when designing GPS-based field studies of animal social behaviours, and highlights the importance of identifying the optimal deployment decisions for drawing informative and robust conclusions.

show abstract

Section: Social Network Structuresmentioning

confidence: 99%

A guide to sampling design for GPS‐based studies of animal societies

Klarevas‐Irby

Papageorgiou

et al. 2022

Methods Ecol Evol

View full text Add to dashboard Cite

show abstract

“…Several studies have suggested that social networks are quite robust to sampling subsets of the population- Silk et al (2015) estimated that studies sampling 30% or more of a population can generate reliable estimates of individuals' social network positions-so GPS devices can theoretically be deployed to characterize even large social networks. Subsequent studies highlighted that repeated sampling and avoiding misidentifications are critical for producing accurate networks (Davis, Crofoot & Farine 2018;Sunga, Webber & Broders 2021)-both are the strengths of GPS tracking data. However, GPS positions are also prone to error and sampling discontinuity (i.e.…”

Section: Social Network Structuresmentioning

confidence: 99%

A guide to sampling design for GPS-based studies of animal societies

Klarevas‐Irby

Papageorgiou

et al. 2022

Preprint

View full text Add to dashboard Cite

GPS-based tracking is a widely used automated data collection method for studying wild animals. Much like traditional observational methods, using GPS devices to study social animals requires making a number of decisions about sampling that can affect the robustness of a study's conclusions. For example, sampling fewer individuals per group across many distinct social groups may not be informative enough for inferring behavioural patterns at a finer social organizational scale, while sampling more individuals per group across fewer groups limits the ability to draw conclusions about populations. Here, we integrate previous insights from animal social network analysis with simulated and empirical data to provide quantitative recommendations when designing GPS-based tracking studies of animal societies. We outline the trade-offs faced by researchers, and how these trade-offs should vary across social organizational scales and social systems in relation to questions of interest, such as those defined within versus among groups, and spanning from cohesive, stable groups through to more open societies. We discuss three fundamental axes of sampling effort requiring consideration when deploying GPS devices to study animal societies: 1) Sampling coverage—the number and allocation of GPS devices among individuals in one or more social groups; 2) Sampling duration—the total amount of time over which devices collect data; 3) Sampling frequency—the temporal resolution at which GPS devices record data. Exploring each axis, we quantify the relationship between sampling effort and sampling error, giving recommendations on GPS sampling design to address research questions across social organizational scales and social systems—from group movement to social network structure and collective decision-making. In doing so, we also highlight practical limitations on deploying GPS tracking. Our study provides practical advice for empiricists to navigate their decision-making processes when designing GPS-based field studies of animal social behaviours, and highlights the importance of identifying the optimal deployment decisions for drawing informative and robust conclusions.

show abstract

“…1), this was only a partial focus of our results. Animal social systems vary widely, and while sampling effects optimal sampling strategies are likely to vary with social structure (Clements et al, 2022;Silk, 2018;Sunga et al, 2021), this has remained understudied. Similarly, while we varied network dynamics in our simulations, the network of every individual was drawn from the same probability distribution.…”

Section: Future Stepsmentioning

confidence: 99%

Generation and applications of simulated datasets to integrate social network and demographic analyses

Silk¹,

Giménez²

2022

Preprint

View full text Add to dashboard Cite

Social networks are tied to population dynamics; interactions are driven by population density and demographic structure, while social relationships can be key determinants of survival and reproductive success. However, difficulties integrating models used in demography and network analysis have limited research at this interface.Here we introduce the R package genNetDem for simulating integrated network-demographic datasets. By generating co-capture data with known statistical relationships it provides functionality for methodological research. We demonstrate its use with case studies testing how imputation and sampling design influence the success of adding network traits to conventional Cormack-Jolly-Seber (CJS) models.We show that incorporating social network effects in CJS models generates qualitatively accurate results, but with downward-biased parameter estimates when network position influences survival. Biases are greater when fewer interactions are sampled or fewer individuals observed in each interaction. However, model performance is robust to network covariance, network structure and sampling design. Longitudinal imputation of missing network measures is more stable than cross-sectional imputation but neither approach successfully generates accurate parameter estimates.While our results indicate the potential of incorporating social effects within demographic models, they show that imputing missing network measures alone is insufficient to accurately estimate social effects on survival. This points to the importance of incorporating network imputation approaches. genNetDem provides a flexible tool to aid these methodological advancements and help researchers testing other sampling considerations in social network studies.

show abstract

Influence of number of individuals and observations per individual on a model of community structure

Cited by 10 publications

References 41 publications

A guide to sampling design for GPS‐based studies of animal societies

A guide to sampling design for GPS‐based studies of animal societies

A guide to sampling design for GPS-based studies of animal societies

Generation and applications of simulated datasets to integrate social network and demographic analyses

Contact Info

Product

Resources

About