Estimation and applications of a specialized mixture IRT model and its extensions in Mplus
Presenter: Minjeong Jeon (UCLA)
Wednesday, February 3, 2021, 10:00 am-12:00 pm (PST)
Zoom host: email@example.com; 310-848-8991
Description: Finite mixture models have been widely utilized in IRT analysis for studying sub-populations (or latent classes) of respondents and/or items. While most applications of mixture models in IRT seem to serve on exploratory purposes, e.g., for identifying the number and nature of latent classes are the main goals of data analysis, confirmatory uses of mixture models have also been found to be useful. In this workshop, I will focus on a special type of confirmatory mixture IRT model, called the Saltus model (Wilson, 1989). The original Saltus model was developed within the context of investigating children’s differential developmental stages. In this workshop, I will argue and show that this useful model can be utilized in much broader applications. In addition, I will explain that the Saltus model can be modified and parameterized as a constrained mixture model, and show how it can be extended in a variety of ways and applied in a wide range of situations. Estimation of the Saltus model and its extensions will be demonstrated with Mplus (and possibly with an R package).
Berkeley Assessment System Software/BASS
Presenter: David Torres Irribarra (Pontificia Universidad Católica de Chile)
Monday, February 8, 2021, 9:00 am-12:00 pm (PST)
Zoom host: firstname.lastname@example.org; 415-533-4508
Description: The BEAR Assessment System Software (BASS) encompasses a theoretical basis for the domain modeling logic, and offers tools for development, delivery, scoring, reporting and use of learning evidence. Specifically, BASS employs the UC Berkeley BEAR Assessment System (BAS) to coordinate the instrumentation with the domain modeling, which is a four-part approach to modeling that involves a combined domain and student model (Wilson, 2005). The structure of the software system is designed to highlight the four parts of the BAS, and to allow educational practitioners and developers to implement the processes. In this workshop, we provide hands-on activities with the software BASS 2.0 while we cover its theoretical bases of construct mapping of the domain and student models, and how these serve as a basis for the “full cycle” of instrument development, deployment and reporting. Participants will learn how the software enables practitioners and developers to implement the full range of assessment processes.