. 2020 Mar 23;74(4):46. doi: 10.1007/s00265-020-2829-y

Table 1.

Summary of the likelihood of dyadic association. Multimodel inference of the GLMM and LMM. Expected weights for each predictor variable are indicated in italics, and a sum of Akaike weights (based on AICc) per predictor that are larger than expected given the model set are indicated in bold. Estimates are the average estimates produced by the repeated random selection approach

		Significant association		Pairwise affinity value
Term	Expected akaike weight	Estimate	Summed akaike weights	Estimate	Summed akaike weights
Kinship	0.38	2.4	1	0.39	1
Sex f_m⁽¹⁾ * Rank Difference	0.04	0.76	0.95	0.05	0.85
Sex m_m⁽¹⁾ * Rank Difference		1.12		0.08
Sex f_m⁽¹⁾ * Age adult_subadult⁽²⁾	0.01	− 1.41	0.98	−0.12	1
Sex m_m⁽¹⁾ * Age adult_subadult⁽²⁾		− 0.11		0.09
Sex f_m⁽¹⁾ * Age subadult_subadult⁽²⁾		− 1.50		−0.20
Sex m_m⁽¹⁾ * Age subadult_subadult⁽²⁾		0.48		0.09
Rank Difference	0.40	− 1.18	1	−0.10	1
Sex f_m⁽¹⁾	0.30	− 0.12	1	0.05	1
Sex m_m⁽¹⁾		0.61		−0.07
Age adult_subadult⁽²⁾	0.28	0.61	1	0.10	1
Age subadult_subadult⁽²⁾		2.35		0.51
Newborn Infants yes_no⁽³⁾	0.27	− 0.47	1	0.01	1
Newborn Infants yes_yes⁽³⁾		1.34		0.11

(1) = reference level is female_female

(2) = reference level is adult_adult

(3) = reference level is no_no