Analyses Using Achievement Levels Based on Plausible Values
AIR - NAEP Working Paper #2017-02, NCES Data R Project Series #01
Achievement levels are important yardsticks in the National Assessment of Educational Progress (NAEP) program. Three levels—Basic, Proficient, and Advanced—are defined for each subject and each grade, with cut scores for each level determined through a standard-setting process. The results using NAEP achievement levels are standard indicators of student performance reported in the Nation’s Report Card.
Methodologically, an analysis using NAEP achievement levels can be conducted with different approaches. In this document, we provide a detailed description of the plausible value method and demonstrate the method by replicating the analyses using SAS and comparing the replicated analysis results with those produced with NAEP Data Explorer. It is intended to provide a complete document for NAEP data users to easily adapt the methodologies with other statistical software packages such as SAS, Stata, R, and so on.
The results from the replication exercise with SAS fully match those from NDE—the official NCES online NAEP data analysis tool. This indicates that the methods discussed in this document can be used as a guide to conduct analyses using achievement levels with other statistical software packages, though there are limitations.