Lexical Diversity and Sophistication in State Accountability Testing: A Corpus-Based Investigation

High-stakes, state-level accountability assessments play a central role in U.S. educational accountability systems, shaping instructional priorities and influencing opportunities for many students. Despite their significance, relatively little is known about the linguistic complexity of these assessments, particularly in light of natural language processing (NLP) tools that allow systematic analysis of lexical demands. To increase our understanding of the language features and demands of state accountability assessment, the present study conducts a corpus-based analysis of lexical diversity and sophistication in a large-scale state accountability assessment across five subjects (Reading, Math, Science, Writing, and Social Studies) and Grades 3 through 8. Specifically, using two NLP tools, the study examined how lexical diversity and sophistication features vary across disciplines and grades. Results from ANOVAs revealed systematic variation across subjects and grade levels. The findings provide a comprehensive lexical landscape, revealing the uneven distribution of lexical density, diversity, and sophistication across academic disciplines.