糟糕考试的高风险

Journal of Human Resources · 2023 · [{"name": "Jack Rossiter", "affiliation": ["Center for Global Development", "University of Cape Coast"]}, {"name": "Might Kojo Abreh", "affiliation": ["University of Cape Coast"]}, {"name": "Aisha Ali", "affiliation": ["Center for Global Development"]}, {"name": "Justin Sandefur", "affiliation": ["Center for Global Development"]}]

中文摘要

每年，西非有两百万名中学生参加统一考试。通过率波动极大，引发了关于作弊与短期政策变动的猜测。为检验这些假说，我们构建了包含 2011 至 2019 年试题的混合考试，并对 4,380 名学生施测。仅考试难度一项就解释了加纳通过率波动的 80%，而在尼日利亚及其他地区，仍有其他因素发挥影响。2015 年数学未通过的考生中，有一半若参加 2019 年的考试将会通过。基于模型的估计表明，提高考试可比性将使中学毕业生的技能明瑟回报提高 6 个百分点。

Abstract

<h3>Abstract</h3> Each year two million secondary-school students across West Africa sit coordinated exams. Pass rates fluctuate enormously, fueling speculation about cheating and short-term policy changes. To investigate these hypotheses, we construct hybrid exams containing items spanning 2011-2019 and administer these to 4,380 students. Exam difficulty alone explains 80 percent of pass rate fluctuations in Ghana, while additional factors remain influential in Nigeria and elsewhere. Half of the candidates who failed mathematics in 2015 would have passed in 2019. Model based estimates imply that improving exam comparability would increase the Mincerian return to skills among secondary school graduates by 6 percentage points.