This event has completed.

8/13/2014 - 8/13/2014

"Modeling item response profiles using factor models, latent class models, and latent variable hybrids"
Wednesday, August 13, 2014 from 4PM-5PM EST
Presented by NERA Director Dena Pastor, James Madison University

There are a wide variety of latent variable models, both old and new, appropriate for modeling the relationships among observed variables. More traditional models include those using only latent continuous variables (e.g., factor models, item response theory models) or latent categorical variables (e.g., latent class analysis). More recently introduced models, known as latent variable hybrid models (Muthén, 2008), incorporate both latent continuous and latent categorical variables (e.g., factor mixture models or mixture factor models). To enhance the understanding of the hybrid models, Muthén (2008) provided a framework which highlights how the hybrid models are both similar and different to one another and to their traditional counterparts. This presentation will use Muthén’s framework to illustrate the model-implied item response profiles (IRPs) that correspond to these latent variable models used with dichotomous item response data. The IRPs associated with the more traditional latent class and factor models are presented along with the IRPs of the newer hybrid models. A graphical representation of item response profiles is first described, followed by a discussion of the ways in which IRPs can vary across examinees and the latent variable models best utilized to capture certain forms of variation. To illustrate the use of these models in practice, an exploratory model selection approach is used to determine which of the various model best represents college students’ responses to items assessing their knowledge of ethics and generalizability in psychosocial research.

Additionally, please use the following link to log in to the webinar on August 13th. We would recommend that you log in 10-15 minutes early to make sure the appropriate software is downloaded on your computer for you to participate in the webinar (Click here for additional information about this: http://www.microsoft.com/en-us/download/details.aspx?id=23396). Additionally, we will not be using the webinar audio system but so attendees should use their computer to view the presentation and call in by phone for audio:

Join Online Meeting
https://meet.lync.com/etsorg1-ets/sholtzman/28PQJJPB
AND
Join by Phone
Toll-free number: 8884625931
Participant code: 6097345037