Table 4.

Effect of ‘part of the day’ (morning or afternoon), ‘round of tests’ (first, second or third), ‘sex’ (female or male), ‘social context’ (individual, unfamiliar, familiar), ‘hour’ (first or second hour of the test) and interaction between ‘social context’ and ‘hour’ on number of following bouts recorded in one hour

Fixed effect	Comparison	Estimate	2% CI	98% CI	P value
Part of the day	Morning vs afternoon	0.043	−0.029	0.116	0.221
Sex	Female vs male	0.007	−0.085	0.098	0.877
Hour	Second vs first	−0.063	−0.089	− 0.037	<.0001
Round	First vs second	0.060	−0.033	0.154	0.259
	First vs third	0.121	0.024	0.219	0.007
	Second vs third	0.061	−0.040	0.162	0.307
Social context	Unfamiliar vs familiar	0.059	0.001	0.121	0.045
Social context × hour	First hour: unfamiliar vs familiar	0.074	0.003	0.144	0.032
	Second hour: unfamiliar vs familiar	0.018	−0.045	0.081	0.554
Hour × treatment	Familiar: second hour vs first hour	−0.035	−0.064	− 0.006	0.014
	Unfamiliar: second hour vs first hour	−0.091	−0.134	− 0.048	<.0001
Random effect		Variance	± SE
Individual identity		0.008	± 0.087
Test		0.007	± 0.085

Coefficients and 96% confidence intervals are presented; statistically significant comparisons (zero is not included in the interval) are in bold. P values obtained with Tukey method adjusted for multiple comparisons. Results are in the log (not in the response) scale. ‘Individual identity’ and ‘test’ are fitted as random effects; variances associated with them are shown