Table 1.

Mate preferences – minimal adequate model

	Estimate	Num d.f.	Denom d.f.	Test statistic	P value
Minimal adequate model
Intercept	−1.63	1	66.46	−47.52	< 0.0001
HL_focal	0.14	1	204.90	2.47	0.01
HL_stimulus	0.03	1	160.20	0.05	0.96
HL_focal × HL_stimulus	6.76	1	62.93	2.76	0.008
Relatedness	0.30	1	63.05	1.65	0.10
Relatedness²	−2.33	1	274.20	−2.39	0.02
Relatedness × HL_stimulus	−3.10	1	277.10	−2.06	0.04
Dropped terms
HL_offspring	−0.13	1	66.30	−0.24	0.81
Sex (female)	0.02	1	61.65	0.23	0.82
HL_stimulus × sex (female)	0.95	1	164.80	0.93	0.35
HL_focal × HL_stimulus × sex (female)	5.67	1	61.06	1.05	0.30
HL_offspring × sex (female)	0.90	1	43.50	0.94	0.35
Relatedness × HL_focal	−1.42	1	63.37	−0.66	0.51
Relatedness × sex (female)	−0.29	1	62.97	−0.68	0.50
Relatedness² × sex (female)	1.55	1	282.00	0.71	0.48

Table consists of all factors tested in the binomial mixed model with proportion of time spent with each of the stimulus birds as the dependent variable (N _focals = 116, N _tests = 359). Given is the estimate, the degrees of freedom (d.f.), the test statistic (F‐value) and the significance (P‐value). Significant terms (P<0.01) are indicated in bold, and marginally significant terms (P= 0.05 ‐ 0.01) are indicated in italics (and described in Appendix S4). A random effect for stimulus bird identity (mean ± SE; 0.22 ± 0.04) r and random slopes for focal bird identity with respect to the stimulus bird's heterozygosity (2.41 ± 0.70), relatedness (1.46 ± 0.47) and offspring heterozygosity (3.23 ± 1.98), and a random effect for test number (to allow for negative correlations among association times within one six‐choice test; (−13.89 ± 0.52) and an extra scale parameter on the original scale, were included in the model. Using backwards elimination of factors, the P‐values, d.f. and test statistics given come from the last model in which the factor or interaction was included. Degrees of freedom for F‐ and t‐tests were calculated using the degree of freedom approximation proposed by Kenward and Roger (1997).