Table 4.

Results of the ‘what-makes-top-associates’ (Model 2a) and ‘what-makes-significant-associates’ (Model 2b) models fitted to test for species differences in the characteristics of top and significant associates for individuals of each sex. Significant p-values are indicated in italics. p-Values are only given for terms not included in an interaction. Results are based on PAV values derived from the ‘individual randomization’. Results for other randomizations are provided in electronic supplementary material, tables S3 and S6.

Model 2a what-makes-top-associateModel 2b what-makes-significant-associates
sample sizetotal number of periods (quarter or year) = 47 845, number of dyads = 7055, number of individuals = 302, number of communities = 7.
null versus full modelχ2d.f.p-valueχ2d.f.p-value
33.816<0.00143.336<0.001
estimate ± s.e.χ2p-valueestimate ± s.e.χ2p-value
intercept−3.82 ± 0.42−2.34 ± 0.39
Sex 1 (male)−0.55 ± 0.28−0.04 ± 0.15
Sex 2 (male)−0.22 ± 0.32−0.06 ± 0.15
species (chimpanzee)0.05 ± 0.490.54 ± 0.46
kin (Yes)3.67 ± 0.2721.28<0.0013.83 ± 0.58
top/significant associates in previous quarter (Yes)1.62 ± 0.1220.99<0.0010.74 ± 0.0915.85<0.001
kin × species−1.37 ± 0.683.010.083
Sex 1 × Sex 2−0.23 ± 0.48−0.17 ± 0.27
Sex 1 × species−0.42 ± 0.31−0.52 ± 0.17
Sex 2 × species−0.68 ± 0.37−0.50 ± 0.17
Sex 1 × Sex 2 × species2.72 ± 0.5227.83<0.0011.84 ± 0.2938.31<0.001