Links between international large-scale assessment (ILSA) methodologies, international organization (IO) ideologies, and education policies are not well understood. Framed by statistical constructivism, this article describes two interrelated phenomena. First, OECD/PISA and UNESCO/TERCE documents show how IOs’ doctrines about the value of education, based on either Human Capital Theory or Human Rights, shape the design of the ILSAs they support. Second, quantitative analyses for four Latin American countries show that differently designed ILSAs disagree on the effectiveness of a specific policy, namely, grade retention: PISA’s achievement gap between repeaters and nonrepeaters doubles TERCE’s. This matters and warrants further research: divergent empirical results could potentially incentivize different education policies, reinforce IOs’ initial policy biases and provide perverse incentives for countries to modulate retention rates or join an ILSA on spurious motivations. In summary, ILSA designs, shaped by IOs’ educational doctrines, yield different data, potentially inspiring divergent global policy directives and national decisions.

中文翻译：

---

政策证据设计：国际大规模评估和成绩重复

国际大规模评估 (ILSA) 方法、国际组织 (IO) 意识形态和教育政策之间的联系尚不清楚。本文以统计建构主义为框架，描述了两个相互关联的现象。首先，OECD/PISA 和 UNESCO/TERCE 文件显示了 IOs 关于教育价值的学说，基于人力资本理论或人权，如何影响他们支持的 ILSA 的设计。其次，对四个拉丁美洲国家的定量分析表明，设计不同的 ILSA 对特定政策的有效性存在分歧，即保留成绩：PISA 重读者和非重读者之间的成绩差距是 TERCE 的两倍。这很重要并值得进一步研究：不同的实证结果可能会激励不同的教育政策，强化 IO 最初的政策偏见，并为各国调整保留率或加入虚假动机的 ILSA 提供不正当的激励措施。总之，由 IO 的教育学说塑造的 ILSA 设计产生不同的数据，可能会激发不同的全球政策指令和国家决策。