New NAEP Dataset Offers Insights into Students' Mathematics Reasoning

Researchers now have a rare opportunity to access students’ reasoning and responses to mathematics questions on The Nation’s Report Card. The National Center for Education Statistics released a new restricted-use dataset (RUD) that contains over 250,000 grade 4 and 8 students’ answers to constructed-response questions from the 2017 and 2019 NAEP Mathematics assessments.
This dataset contains student answers to 10 short and extended constructed-response items. It also comprises student demographic information such as their gender, race or ethnicity, English learner status, and disability status. There is also a detailed data file for each response that contains:
- the original item students saw,
- the rubric that human raters used to hand score the items, and
- several exemplar student responses that represent each of the scoring bands.
Researchers could use the data to support a deeper understanding of students’ mathematical problem solving or to identify trends in student misconceptions—information that facilitates new approaches to teaching complex mathematical concepts. This dataset could also inform the development of assessment items.
This dataset was used for the Mathematics Automated Scoring Challenge in May 2023. Participants were charged with designing an automated approach to scoring open-ended answers to NAEP short and extended constructed-response questions. Their methods had to be fair and valid ways to modernize what has been a time-consuming and expensive process.
Read an overview of the challenge on Github and explore information about the results.
For more information and to access the dataset (NCES 2024072), visit the
NCES website. Researchers will need to apply for an
RUD license or amend their current one.