As institutions increasingly use predictive algorithms to allocate scarce resources, scholars have warned that these algorithms may legitimize inequality. Although research has examined how elite discourses position algorithms as fair, we know less about how the public perceives them compared to traditional allocation methods. We implement a vignette-based survey experiment to measure perceptions of algorithmic allocation relative to common alternatives: administrative rules, lotteries, petitions from potential beneficiaries, and professional judgment. Focusing on the case of schools allocating scarce tutoring resources, our nationally representative survey of U.S. parents finds that parents view algorithms as fairer than traditional alternatives, especially lotteries. However, significant divides emerge along socioeconomic and political lines—lower socioeconomic status (SES) and conservative parents favor the personal knowledge held by counselors and parents, whereas higher SES and liberal parents prefer the impersonal logic of algorithms. We also find that, after reading about algorithmic bias, parental opposition to algorithms is strongest among those who are most directly disadvantaged. Overall, our findings map cleavages in attitudes that may influence the adoption and political sustainability of algorithmic allocation methods.

中文翻译：

---

预测算法和公平感知：家长对 K-12 教育中算法资源分配的态度


随着机构越来越多地使用预测算法来分配稀缺资源，学者们警告说，这些算法可能会使不平等合法化。尽管研究考察了精英话语如何将算法定位为公平的算法，但与传统分配方法相比，我们对公众如何看待它们知之甚少。我们实施了一个基于小插图的调查实验，以衡量相对于常见替代方案（行政规则、抽签、潜在受益人的请愿书和专业判断）对算法分配的看法。针对学校分配稀缺辅导资源的情况，我们对美国家长的全国代表性调查发现，家长认为算法比传统替代方案更公平，尤其是彩票。然而，在社会经济和政治方面出现了显著的分歧——较低的社会经济地位 （SES） 和保守的父母更喜欢辅导员和父母所持有的个人知识，而较高的 SES 和自由派父母更喜欢算法的非个人逻辑。我们还发现，在阅读了算法偏见之后，父母对算法的反对在那些最直接处于不利地位的人中最强烈。总体而言，我们的研究结果绘制了可能影响算法分配方法的采用和政治可持续性的态度分裂。