Individualized Assessment: Detection of Learning via Score-Based Tests

• In Event: Marriott Roundtable Session 41
  In Roundtable Session: Modeling and Detecting Measurement Invariance

Abstract

Change detection is of central interest in longitudinal assessment: to study
whether and when the the mechanism underlying a set of observations
changes. In assessment settings, the upward change is often referred as
“learning”. While many methods have been proposed to detect change, they
typically require prior definition of changing point. However, this piece
of information is typically unknown in practice. As an alternative, this
paper focuses on a family of recently-proposed tests based on stochastic
processes of case-wise derivatives of the likelihood function (i.e., scores).
These score-based tests only require estimation of the null model (when
parameter stability is assumed to hold). In this proposal, we aim to detect
change and specific changing point in individual’s longitudinal assessment
framework.

Authors

• Ting Wang, American Board of Anesthesiology

• Huaping Sun, The American Board of Anesthesiology

• Yan Zhou, The American Board of Anesthesiology 

• Ann Elizabeth Harman, American Board of Anesthesiology (ABA)