Data from: A new method for statistical detection of directional and stabilizing mating preference

Estimation of mating preferences is a prerequisite for understanding how sexual selection through mate choice shapes both mating systems and sexual dimorphisms. Most studies of mating preferences assay mate choice using either a no choice or a binary choice design. Binary choice trials typically employ either an artificial signal or some fixed difference (e.g. color or size) between the signaling individuals. Although statistically more powerful than no choice designs, such experiments cannot be used to detect stabilizing preference. Further, the use of artificial signals is problematic because signal components tend to be varied in isolation, and hence do not reflect natural variation. Here we present a new method that uses natural variation among individuals in choice trials to determine if mating preference is absent, directional and/or stabilizing. The protocol is tested using simulation and shown to be robust to the preference function, to have the required statistical power, to be unbiased in almost all cases, and to give confidence regions that modestly overestimate the desired 95% criterion. We demonstrate the use of the method with data from mate choice trials of the sand cricket, Gryllus firmus. Software to apply this new approach is provided in Dryad.评估择偶偏好（mating preferences），是理解通过配偶选择产生的性选择（sexual selection）如何塑造交配系统（mating systems）与性二态性（sexual dimorphisms）的核心前提。当前绝大多数择偶偏好研究，均采用无选择（no choice）或二元选择（binary choice）实验设计来开展配偶选择行为的测定工作。二元选择实验通常会设置人工信号，或是在信号发出个体间预设固定差异（如体色或体型大小）。尽管相较于无选择设计，这类实验的统计效力更优，但却无法检测稳定型择偶偏好（stabilizing preference）。此外，人工信号的应用存在固有局限：其信号组分往往仅被单独调控，无法反映自然种群中的真实变异模式。本研究提出一种全新分析方法，通过利用选择实验中个体间的自然变异，可精准判断择偶偏好是否缺失、是否存在定向型择偶偏好（directional preference）或稳定型择偶偏好。本研究通过模拟实验对该方案进行了验证，结果显示该方法对偏好函数具有优异稳健性，具备预设的统计效力，在绝大多数场景下均无偏倚，且所生成的置信区域会适度高估目标95%置信准则。我们以沙地蟋蟀（Gryllus firmus）的配偶选择实验数据为实例，展示了该方法的实际应用流程。本方法配套的分析软件已上传至Dryad数据库。