DSNAEP Data Portfolio

Dataset

The public data behind this project.

Primary dataset: NAEP (National Assessment of Educational Progress)

  • Name: NAEP — "The Nation's Report Card"
  • Source / link: https://www.nationsreportcard.gov/ (data via the NAEP Data Service API)
  • Created and maintained by: the National Center for Education Statistics (NCES), part of the Institute of Education Sciences in the U.S. Department of Education
  • What it measures: academic achievement of U.S. students in grades 4, 8, and 12, reported as average scale scores (0–500) and achievement-level percentages for the nation, each state, and a set of large urban districts
  • Important variables: subject (mathematics / reading), grade (4 / 8), jurisdiction (national / state), year, and the mean score
  • Time period: biennial administrations — this project uses 2015, 2017, 2019 (pre-COVID) and 2022, 2024 (post-COVID)
  • Unit of analysis: one row is a jurisdiction × subject × grade × year × student-group summary (e.g., "Texas, grade 8 mathematics, 2024, all students, mean score")

Strengths. NAEP is the gold-standard, nationally representative measure of U.S. student achievement; its scale is comparable across states and years, which makes "before vs. after" comparisons legitimate.

Limitations. It is sample-based (different students each cycle, no individual or school-level scores), it skips years (no 2016, 2018, or 2025), and results for small subgroups are sometimes suppressed.

Secondary dataset: Stanford Education Data Archive (SEDA)

  • Source / link: https://edopportunity.org/ (Educational Opportunity Project at Stanford)
  • Created and maintained by: Stanford University's Educational Opportunity Project, led by Sean Reardon
  • What it measures: nationally comparable district- and state-level achievement and achievement gaps, harmonized onto a common "grade-level" scale
  • Time period: 2009–2019 (version 5.0) plus 2019–2024 "recovery" releases

SEDA is used only as a triangulation check — an independent dataset (built from state tests, not NAEP) to confirm that the patterns I found are not an artifact of one test.