The best generalized linear mixed-effects models (GLMMs).

A) In 'lme4' syntax: match/mismatch of choice ∼ experience + (tutor number −1 | area)
B) In 'lme4' syntax: match/mismatch of choice ∼ experience + (tutor number | area)
*P value from permutation test, see Methods for details
Estimated fixed effects for experience variable of the two best GLMMs explaining the probability of flycatchers to copy the tutors' symbol choice. The models were fitted with Laplace approximation (R function ‘glmer’), binomial error distribution and a logistic link function. Only area-specific random slopes in relation to the number of tutors at the time of choice were included in model 1, whereas both area-specific random slopes and intercepts in relation to the number of tutors at the time of choice were included in model 2.