NCTE Main Study

该数据集涵盖了美国四个学区大约五十所学校和三百个四至五年级数学课堂连续三年的数据收集和观察，其中包括专家对个别视频捕捉的课堂课程的评分。此外，该数据集还包含了使用两种观察工具——CLASS框架和数学教学质量（MQI）的专家评分。规模上，大约有40万个独特的人类评分标签。该数据集的任务是通过对人类和模型注释的评价，来评估课堂教学质量。

This dataset contains three consecutive years of data collection and observational data from approximately 50 schools and 300 Grades 4 to 5 mathematics classrooms across four U.S. school districts. It includes expert ratings of classroom lessons captured in individual video recordings. Additionally, the dataset features expert ratings derived using two observation tools: the CLASS framework and the Mathematics Quality of Instruction (MQI). There are roughly 400,000 unique human annotation labels in total. The core task of this dataset is to evaluate classroom teaching quality by assessing both human and model annotations.

搜集汇总

背景与挑战

背景概述

NCTE Main Study数据集是一个大规模教育研究数据集，覆盖美国四个学区约50所学校、300个四至五年级数学课堂的三年观察数据，包含约40万个独特人类评分标签。它使用CLASS框架和数学教学质量（MQI）两种观察工具进行专家评分，旨在通过人类和模型注释评估课堂教学质量，具有长期、多工具、聚焦数学教育的特点。