To advance ethical leadership using signaling theory, the current work presents a mixture of inductive and deductive studies. Using a constant comparative analysis method, Study 1 involved coding CEO letters to shareholders (n = 10,919 sentences). Eight verbal ethical leader signals (ELSs) emerged and were associated with emotions (e.g., righteous anger, pride). In a set of preregistered experiments, ELSs were found to lead to evaluations of ethical leadership (Study 2: n = 264; Cohen’s d = 0.26). Study 3 illustrated that ELSs led to a reduction in financial theft (n = 434; Cohen’s d = 0.20). Study 4 showed that ELSs led to an improvement in performance (n = 434; Cohen’s d = 0.18) but had little effect on extra role behavior (Cohen’s d = 0.06). Finally, in Study 5 a machine learning algorithm, DeepEthics, was created to automatically score text (ROC =. 84; r = 0.85 between human and algorithm scores), such as emails and meeting transcripts, for ELSs in future research. Recommendations for theory and practice are discussed.

为了使用信号理论推进道德领导，目前的工作提出了归纳和演绎研究的混合体。使用恒定的比较分析方法，研究 1 涉及编码 CEO 致股东的信（n = 10,919 句）。八种口头道德领导信号 (ELS) 出现并与情绪相关（例如，正义的愤怒、骄傲）。在一组预先注册的实验中，发现 ELS 会导致对道德领导力的评估（研究 2：n = 264；Cohen 的d  = 0.26）。研究 3 表明 ELS 导致金融盗窃减少（n = 434；Cohen 的d  = 0.20）。研究 4 表明，ELS 导致绩效改善（n = 434；Cohen's d  = 0.18）但对额外角色行为几乎没有影响（Cohen's d = 0.06）。最后，在研究 5 中，创建了一种机器学习算法 DeepEthics，用于自动对文本（ROC =.84；人类和算法分数之间的r  = 0.85）进行评分，例如电子邮件和会议记录，用于未来研究中的 ELS。讨论了理论和实践建议。