Measuring “Noncognitive” Skills at Scale: Building Longitudinal Student Behavior Composites Using Administrative Data

July 2025

“Noncognitive” skills, especially student behavior, are critical predictors of academic and life outcomes. However, measuring student behavior at scale remains challenging, particularly for longitudinal research. This study uses a demographically diverse sample of students followed from kindergarten to eighth grade in the Boston Public Schools (N=12,232) to examine trade-offs between two psychometric methods—principal components analysis (PCA) and factor analysis—and the use of raw measures for constructing longitudinal composites of student behavior from administrative data. We compare the structure of each approach and their predictive validity in forecasting eighth grade standardized test scores. PCA composites offered slightly stronger predictive power, but factor analysis provided stronger theoretical grounding with statistically comparable predictive validity. Eighth grade unexcused absence rate, a driver of both composites, was also among the strongest predictors overall. Findings offer practical guidance for researchers and policymakers seeking scalable, interpretable measures of student behavior to inform interventions and policies.

Keywords

noncognitive skills, student behavior, PCA, factor analysis, psychometrics

Education level

K-12 Education

Topics

Methods