Data R Project for the National Center for Education Statistics
For decades, the National Center for Education Statistics (NCES) has provided researchers with invaluable datasets reflecting the achievement of American students and the experiences of students, teachers, and school administrators in the American education system.
Due to the scope and complexity of NCES datasets, researchers often must use several different software tools to access, clean, and analyze the data. To help NCES and researchers when accessing the datasets, AIR developed the Data R Project, a set of specialized open-source R tools that streamline the data analysis process.
Developing such open-source software for statistical analysis is critical to advancing research because it fosters transparency, reproducibility, and collaboration. Open-source software democratizes access to sophisticated analytical capabilities by both offering users effective software without charge and allowing them to access and reuse the code for their own purposes.
Our Work
To help NCES and researchers when accessing the datasets, AIR developed the Data R Project, a set of specialized open-source R tools that streamline the data analysis process.
During our efforts to introduce AI into our open-source solutions, AIR determined that even more important than creating AI agents to work with the code was learning how to prepare user documentation that AI assistants can scan and understand to return accurate results. In doing so, we can help not only users who ask us questions but also users who ask chatbots questions.
We also created supplementary R packages—Dire, WeMix, wCorr, and NAEPirtparams—which each offers specialized functionality, such as conducting multilevel modeling and determining weighted and unweighted correlations
AIR intentionally developed these tools as an open-source suite to provide sophisticated analysis and visualization software at no cost to the user, thus providing access to NCES’s large-scale assessment and survey data tools to a broader and more diverse audience. We also developed documentation and provide in-person professional development at conferences to help users navigate the tools.
The tools in our R package reflect advances in computing and meet the shifting trend in higher education to use open-source software packages to make data more accessible.
wCorr: Weighted and unweighted correlations
NAEPirtparams: Provides NAEP Item Response Theory parameters
Impact of Our Work
AIR’s development of EdSurvey and its ancillary R tools supported NCES by bringing innovation to the analysis of complex data structures and broadened NCES’s user base by expanding awareness of its data sets to new audiences.
In 2018, Peggy Carr, who was the NCES associate commissioner at the time, noted that EdSurvey “represents a huge technological innovation in the analysis of survey data collected using a complex sample design” and that the “availability of EdSurvey substantially reduces the burden on the individual researcher” to analyze data collected with a complex sample design.
Since 2014, AIR has rolled out six R packages, with more than 100,000 downloads from R CRAN.
AIR intentionally wrote the Data R Project software to support collaboration with other researchers. In his book Practical Multilevel Modeling Using R, Francis L. Huang extends the WeMix package to work with plausible values outside of the context of EdSurvey and several psychometrics packages using our weighted correlations software wCorr.
Examples like this showcase how open-source software effectively leverages innovation by facilitating global communities of experts who work together to improve tools, fix issues, and expand capabilities faster than traditional closed-source approaches.
