Chen Daniel L, Ertac Seda, Evgeniou Theodoros, Miao Xin, Nadaf Ali, Yilmaz Emrah
Toulouse School of Economics, Institute for Advanced Studies, 21 allée de Brienne, 31015 Toulouse cedex 6, Toulouse, France.
Koc University Department of Economics, Sariyer, Istanbul, 34450, Turkey.
NPJ Sci Learn. 2024 Sep 13;9(1):57. doi: 10.1038/s41539-024-00265-3.
Grit, a non-cognitive skill that indicates perseverance and passion for long-term goals, has been shown to predict academic achievement. This paper provides evidence that grit also predicts student outcomes during the challenging period of the Covid-19 pandemic. We use a unique dataset from a digital learning platform in the United Arab Emirates to construct a behavioral measure of grit. We find that controlling for baseline achievement, students who were grittier according to this measure before the pandemic, register lower declines in math and science scores during the coronavirus period. Using machine learning, behavioral data obtained from the platform prior to the pandemic can explain 77% of the variance in academic resilience. A survey measure of grit coming from the same students, on the other hand, does not have significant predictive power over performance changes. Our findings have implications for interventions on non-cognitive skills, as well as how data from digital learning platforms can be used to predict student behavior and outcomes, which we expect will be increasingly relevant as AI-based learning technologies become more common.
毅力，一种表明对长期目标坚持不懈和充满热情的非认知技能，已被证明可以预测学业成绩。本文提供的证据表明，在新冠疫情这一充满挑战的时期，毅力也能预测学生的成绩。我们使用来自阿拉伯联合酋长国一个数字学习平台的独特数据集构建了一种衡量毅力的行为指标。我们发现，在控制了基线成绩之后，根据这一指标，在疫情之前更有毅力的学生在新冠疫情期间数学和科学成绩的下降幅度更小。利用机器学习，疫情之前从该平台获得的行为数据可以解释学业恢复力差异的77%。另一方面，来自这些学生的一项关于毅力的调查指标对成绩变化没有显著的预测能力。我们的研究结果对非认知技能的干预措施具有启示意义，同时也关乎如何利用数字学习平台的数据来预测学生的行为和成绩，随着基于人工智能的学习技术越来越普遍，我们预计这些结果将越来越具有相关性。
