Skip to main content

Data from: Aggressive interactions influence cognitive performance in Western Australian magpies

Cite this dataset

Speechley, Elizabeth Marie et al. (2024). Data from: Aggressive interactions influence cognitive performance in Western Australian magpies [Dataset]. Dryad.


Extensive research has investigated the relationship between the social environment and cognition, suggesting that social complexity may drive cognitive evolution and development. However, evidence for this relationship remains equivocal. Group size is often used as a measure of social complexity, but this may not capture intraspecific variation in social interactions. Social network analysis can provide insight into the cognitively demanding challenges associated with group-living at the individual-level. Here, we use social networks to investigate whether the cognitive performance of wild Western Australian magpies (Gymnorhina tibicen dorsalis) is related to group size and individual social connectedness. We quantified social connectedness using four interaction types: proximity, affiliative, agonistic, and vocal. Consistent with previous research on this species, individuals in larger groups performed better on an associative learning task. However, social network position was also related to cognitive performance. Individuals receiving aggressive interactions performed better, while those involved in aggressive interactions with more group members performed worse. Overall, this suggests that cognitive performance is related to specific types of social interaction. The findings from this study highlight the value of considering fine-grained metrics of sociality that capture the challenges associated with social life when testing the relationship between the social environment and cognition.

README: Aggressive interactions influence cognitive performance in Western Australian magpies

Dataset contains the R script, data files (.csv) for all analyses and the supplementary material.

Description of the data and file structure

Dataset includes:

  • SN - Data for individuals with social network metrics
  • GS Stable - Data for individuals with stable group sizes
  • RPT COG - Cognition repeatability data
  • RPT AG ALL - Repeatability for agonistic data
  • RPT V ALL - Repeatability for vocal data

Description of the data and file structure

The SNP & Cognition Code is the main R code file use for analysis.

Each data file (.csv) is labelled according to the analysis (SN for social network GLMM, GS stable for group size analysis and RPT for repeatability analysis). SN is the main analysis file containing scores and social network metrics for GLMMs. GS stable is the file for analysis of the subset of individuals with stable group sizes. RPT files are used for repeatability analysis. For example, RPT COG is the repeatability data for cognition scores, RPT AG ALL is the repeatability data for agnostic metrics and RPT V ALL is the repeatability data for vocal metrics.

The additional word doc is the supplementary material containing supplementary figures, tables and analysis.

Abbreviations in data files include:

  • id=individual identity
  • grp=group identity
  • score=associative learning score
  • colour=colour combination used
  • shade=shade of rewarded well (D=dark, L=light)
  • year=year tested (1 = 2020 and 2 = 2021)
  • season=season tested (B=breeding, NB=non-breeding)
  • site=testing site (U=UWA, G=Guildford)
  • mass=body mass (grams)
  • fe=foraging efficiency (amount of biomass caught in grams per minute foraging)
  • sex=sex of individual tested
  • sr=sex ratio in group (females to males)
  • grp.size=group size at time of testing
  • gscurr=group size at time of testing (2020-2021)
  • gs2016=average group size during 2016
  • change=changing in group size from current group size (gscurr) to 2016 (gs2016)
  • neo=neophobia (time taken (seconds) to approach test apparatus from 1m away)
  • order=order tested in group
  • temp=average ambient temperature at time of testing (°C)
  • agd*=agonistic degree
  • agid*=agonistic in-degree
  • agod*=agonistic out-degree
  • ags*=agonistic strength
  • agis*=agonistic in-strength
  • agos*=agonistic out-strength
  • agcov*=agonistic coefficient of variation
  • vd*=vocal degree
  • vid*=vocal in-degree
  • vod*=vocal out-degree
  • vs*=vocal strength
  • vis*=vocal in-strength
  • vos*=vocal out-strength
  • vcov*=vocal coefficient of variation

*node strength for each individual was divided by the highest node strength within that group, thus scaling node strength between 0 and 1 (where individuals with a score of 1 showed the most social connectedness in that group).


Custom R code used for analysis can be accessed here and is labelled "SNP & Cognition Code".


Australian Research Council, Award: DP200100566

Ecological Society of Australia

Australian Government

Australian Research Council, Award: DE220100096