Female preference for large waving claws in the dotillid crab Ilyoplax pusilla

Designing mate choice experiments

The important role that mate choice plays in the lives of animals is matched by the large and active research field dedicated to studying it. Researchers work on a wide range of species and behaviours, and so the experimental approaches used to measure animal mate choice are highly variable. Importantly, these differences are often not purely cosmetic; they can strongly influence the measurement of choice, for example by varying the behaviour of animals during tests, the aspects of choice actually measured, and statistical power. Consideration of these effects are important when comparing results among studies using different types of test, or when using laboratory results to predict animal behaviour in  natural populations. However, these effects have been underappreciated by the mate choice literature to date. I focus on five key experimental considerations that may influence choice: (i) should mating be allowed to occur, or should a proxy behavioural measure of preference be used instead? (ii) Should subjects be given a choice of options? (iii) Should each subject be tested more than once, either with the same or different stimuli? (iv) When given a choice, how many options should the subject choose between? (v) What form should the experimental stimuli take? I discuss the practical advantages and disadvantages of common experimental approaches, and how they may influence the measurement of mate choice in systematic ways. Different approaches often influence the ability of animals to perceive and compare stimuli presented during tests, or the perceived costs and benefits of being choosy. Given that variation in the design of mate choice experiments is likely unavoidable, I emphasise that there is no single 'correct' approach to measuring choice across species, although ecological relevance is crucial if the aim is to understand how choice acts in natural populations. I also highlight the need for quantitative estimates of the sizes of potentially important effects, without which we cannot make informed design decisions.