The repeatabilities (R) of behavioural measures in juvenile pike (n = 34) within each experimental situation: (a) control (no risk), (b) competition (low risk), (c) predation (high risk).

Generalised linear mixed-effects and linear mixed-effects models (rptR package in R [33]) with fish identity fitted as random effect and the behavioural measure as dependent factor were used for calculating repeatabilities, standard errors, 95% confidence intervals (CIs) and P-values. Latency to prey attack and swimming activity were log-transformed to achieve normality.