QMNET Talk: Some Applications of Latent Variable Modeling Using Mplus (Video recording available)

Image for QMNET Talk: Some Applications of Latent Variable Modeling Using Mplus (Video recording available)

More Information

mcds@unimelb.edu.au

Melbourne Centre for Data Science is proud to support the University of Melbourne Quantitative Research Methods Network (QMNET). The 2021 QMNET seminar series ran as weekly talks on a variety of topics that span disciplinary boundaries.

The 2021 series kicked off with a two-part session focused on Mplus. Starting at 10am Melbourne time on Thursday 25 and Friday 26 March, this two-part Mplus feature included talks from preeminent methodologists and Mplus program developers Bengt Muthén and Tihomir Asparouhov.

Bengt Muthén is Professor Emeritus at UCLA, former President of the Psychometric Society, creator of the Mplus program, and holds a PhD in statistics from the University of Uppsala.

Tihomir Asparouhov has made significant contributions to data analytics methods across multiple social and health science areas, including unique innovations in Mplus, and holds a PhD in mathematics from Caltech.

As many will know, Mplus is now the most commonly used latent variable modeling program in psychology and other areas, offering a variety of unique features for mixed data types under single- and multi-level models (including mixture models with observed and unobserved classes). With the upcoming release of Mplus version 8.6, a variety of new features will be implemented including automated procedures for Latent Transition Analysis and advances in Bayesian analysis.

Seminar 2

Bengt Muthén discussed some Applications of Latent Variable Modeling using Mplus

Seminar Title: Some Applications of Latent Variable Modeling Using Mplus

When: Friday 26 March, 10:00am - 11:30am AEDT (Melbourne time)

Where: This talk was delivered via Zoom Webinar


Abstract:
This talk gives an overview of some recent and ongoing latent variable research. Borrowing ideas from multilevel factor analysis, longitudinal SEM in a single-level, wide format is formulated in a new way that finds a well-fitting model 45 years after the writing of the classic Wheaton, Muthen, Alwin, and Summers article.  This segues into a generalisation of latent transition analysis using the multilevel notion of a random intercept while staying in a single-level, wide format.  Turning back to multilevel modeling, the talk considers time series analysis of intensive longitudinal data with a focus on modeling cycles.  This is illustrated by daily cycles of tiredness and positive emotions in teenagers, intervention data on electricity consumption, a randomized intervention related to positive and negative affect, and a study of smoking cessation.

Talk 1 - details can be found here

View past talks on QMNET’s YouTube channel