It is widely acknowledged that subgroup bias can influence hiring evaluations. However, the notion that bias still threatens equitable hiring outcomes in modern employment contexts continues to be debated, even among organizational scholars. In this study, we sought to contextualize this debate by estimating the practical impact of bias on real-world hiring outcomes (a) across a wide range of hiring scenarios and (b) in the presence of diversity-oriented staffing practices. Toward this end, we conducted a targeted meta-analysis of recent hiring experiments that manipulated both candidate gender and qualifications to couch our investigation within ongoing debates surrounding the impact of small amounts of bias in otherwise meritocratic hiring contexts. Consistent with prior research, we found evidence of small gender bias effects (d = −0.30) and large qualification effects (d = 1.61) on hiring managers’ evaluations of candidate hireability. We then used these values to inform the starting parameters of a large-scale computer simulation designed to model conventional processes by which candidates are recruited, evaluated, and selected for open positions. Collectively, our simulation findings empirically substantiate assertions that even seemingly trivial amounts of subgroup bias can produce practically significant rates of hiring discrimination and productivity loss. Furthermore, we found contextual factors can alter but cannot obviate the consequences of biased evaluations, even within apparently optimal hiring scenarios (e.g., when extremely valid assessments are used). Finally, our results demonstrate residual amounts of subgroup bias can undermine the effectiveness of otherwise successful targeted recruitment efforts. Implications for future research and practice are discussed.

人们普遍认为，小组偏见会影响招聘评估。但是，即使在组织学者中，偏见仍然威胁着现代就业环境下公平的招聘结果的观点仍在争论中。在本研究中，我们试图通过估计偏见对现实世界中的招聘结果的实际影响来对这场辩论进行情境化（a）在广泛的招聘场景中，以及（b）在面向多样性的人员配备实践中。为此，我们对最近的招聘实验进行了有针对性的荟萃分析，该实验操纵了候选人的性别和资格，使我们的调查在围绕少数偏见在其他情况下的精英化招聘环境的影响的正在进行的辩论中进行。与先前的研究一致，我们发现了性别偏见影响较小的证据（d = −0.30）和较大的限定效应（d= 1.61）。然后，我们使用这些值来告知大型计算机仿真的起始参数，该计算机仿真旨在对常规过程进行建模，通过常规过程招募，评估和选择候选人以建立空缺职位。总的来说，我们的模拟发现从经验上证实了这样的断言：即使看似微不足道的亚组偏见也可以产生很大的雇佣歧视和生产率损失。此外，我们发现情境因素可以改变，但不能消除有偏见的评估的后果，即使在表面上看似最佳的招聘场景下（例如，当使用极其有效的评估时）。最后，我们的结果表明，剩余的亚组偏见量可能会破坏原本成功进行的有针对性的招募工作的有效性。讨论了对未来研究和实践的影响。