A Markov Estimation Strategy for Longitudinal Learning Diagnosis: Providing Timely Diagnostic Feedback

Abstract

Timely diagnostic feedback is helpful for students and teachers, enabling them to adjust their learning and teaching plans according to a current diagnosis. Motivated by the practical concern that the simultaneity estimation strategy currently adopted by longitudinal learning diagnosis models does not provide timely diagnostic feedback, this study proposes a new Markov estimation strategy, which follows the Markov property. A simulation study was conducted to explore and compare the performance of four estimation strategies: the simultaneity, the Markov, the anchor-item, and the separated estimation strategies. The results show that their performance was highly consistent, and they presented in the following relative order: simultaneity > Markov > anchor-item ≥ separated. Overall, although accuracy in parameter estimation is sacrificed slightly with the proposed strategy, it can provide timely diagnostic feedback to practitioners, which is in line with the concept of “assessment for learning” and the needs of formative assessment.