Statistical tests for individual lateralization in behavioural studies: pitfalls and guidelines

Laterality at the individual level refers to a quantitative measure of the degree of bias with which an individual uses laterally paired organs. Studies of laterality commonly involve observing each individual's behaviour multiple times and calculating the mean index value based on observations from several individuals. Due to the simplicity of this observation process, research in this area, particularly those on invertebrate taxa, is increasing. However, it is possible that statistical missteps, including those related to null hypothesis formulation and weighting schemes, may have contributed to less robust conclusions. This short note provides a brief and practical guideline for evaluating the population- and individual-level laterality using the simple laterality index. The simulation results show improper null hypotheses, a priori assumptions of the parity (equal frequencies in the usage of right- and left-side), and a lack of appropriate weighting can distort our interpretations of the evolution of individual-level lateralization, which is usually evaluated by the absolute values of the lateralization indices. Parametric bootstrapping, which allows accurate significance testing of various laterality indices without relying on unrealistic assumptions, is recommended. Python and R codes for these purposes are provided, together with statistical critical values for conventional situations. • Lateralization indices must be tested against appropriate null hypotheses. • Misassumed parity and unequal sampling require careful statistical handling. • This study offers practical guidance for testing individual-level laterality. • R/Python codes as bootstrap tools for robust inference are also provided.