Signal to noise ratio: Lessons Learned from NAEP Process Data
Abstract
The introduction of digitally based assessments has led to large amounts of data collected on students’ behavior using digital platforms. Therefore, it provides opportunities to address existing questions in new ways or generate new questions; thus, producing knowledge bases that have no previous counterparts. Although the data exist, they are not easily accessible. Most of the data are large and messy; to improve the art of assessment, the data need to be understood, transformed, and used wisely. This talk provides examples to mine actionable insights from NAEP Process Data on different aspects (data itself, the analytics of the data, and value creation from the data). The goal is to generate information from Process Data that are of interest to various stakeholders (e.g., policy makers, test/item/platform developers, and educators).