Machine learning (ML) methods for causal inference have gained popularity due to their flexibility to predict the outcome model and the propensity score. In this article, we provide a within-group approach for ML-based causal inference methods in order to robustly estimate average treatment effects in multilevel studies when there is cluster-level…

Education leaders must identify valid metrics to predict student long-term success. We exploit a unique dataset containing data on cognitive skills, self-regulation, behavior, course performance, and test scores for 8th-grade students. We link these data to data on students' high school outcomes, college enrollment, persistence, and on-time degree…

Test-taking experience is a consequence of the interaction between students and assessment properties. We define a new notion, rapid-pacing behavior, to reflect two types of test-taking experience -- disengagement and speededness. To identify rapid-pacing behavior, we extend existing methods to develop response-time thresholds for individual items…