AI 中文总结
研究高中 STEM 教育中传统人工批改限制形成性考试频率问题，通过准实验将学生分两组，对比人工与自动批改。发现自动批改组成绩显著提升，周转时间大幅缩短，还能加速纠正误解、提高自我效能，验证自动评分系统可助力提升学校考试成绩。
AI 中文摘要
在高中 STEM 教育中，传统人工批改形成结构性瓶颈，限制了形成性模拟考试频率。本准实验、混合方法纵向研究（N = 142）调查部署全自动手写评估评分平台以消除此瓶颈的效果。将准备 STEM A 级课程的学生分为对照组（N = 70，人工批改）和干预组（N = 72，自动批改），为期 16 周。结果显示干预组在最终模拟 A 级成绩上有显著提升（p <.01），平均比对照组高 16.8%。评估周转时间从 11.2 天降至< 0.1 天，练习量增加四倍。自动系统的详细旁注功能，尤其是其错误结转（ECF）逻辑，加速了程序误解的纠正并显著提高了学生自我效能。实际上，这些发现验证了自动评分系统可作为高速形成性循环加速器，为国际学校提供高度可扩展、低成本干预，提升机构考试成绩且无需增加员工人数。
英文摘要
Traditional human marking in upper-secondary STEM education creates a structural bottleneck that restricts the frequency of formative mock examinations. This quasi-experimental, mixed-methods longitudinal study (N = 142) investigates the efficacy of deploying a fully automated, handwritten assessment marking platform to remove this bottleneck. Students preparing for STEM A-levels (Mathematics, Further Mathematics, Biology, Chemistry, Physics) were divided into control (N = 70, human marking) and intervention (N = 72, automated marking) groups over a 16-week term. Results indicate that the intervention group achieved a statistically significant improvement in final mock A-level marks (p < .01), scoring on average 16.8% higher than the control group. Furthermore, assessment turnaround time dropped from 11.2 days to < 0.1 days, enabling a fourfold increase in practice volume. The automated system's capacity for granular marginalia, specifically its Error Carried Forward (ECF) logic, accelerated the correction of procedural misconceptions and significantly boosted student self-efficacy. Practically, these findings validate that automated marking systems serve as high-speed formative loop accelerators, providing international schools with a highly scalable, low-cost intervention to elevate institutional exam outcomes without necessitating increases in staff headcount.
Comments10 pages, 2 figures
