Pete Koval from University of Melbourne is offering a 2-day seminar that he recently delivered for Instats titled Intensive Longitudinal Methods: Design and Analysis in Mplus, which is now available on demand – i.e., it can be completed any time in a self-paced manner.
The seminar covers the following:
Sessions 1-3 provide a general introduction to intensive longitudinal methods, including their key strengths and limitations, and an overview of key issues in intensive longitudinal study design (e.g., survey items, sampling frequency, study duration, etc). We also discuss issues of sample size planning and power, reliability of intensive longitudinal scales, and the multilevel structure of intensive longitudinal data. And, finally, there is a brief intro to Mplus (e.g., data format, key components of Mplus input/output files, etc.) in preparation for Sessions 4-6.
Sessions 4-6 comprise a detailed introduction to multilevel analysis with Mplus, including over 25 annotated example Mplus model inputs. We start from the most basic unconditional (‘null’) models through to multilevel regressions including both within- and between-person predictors, fixed and random effects, and cross-level interactions. Next, we move onto multilevel structural equation models (MSEMs) with manifest and latent variables, and indirect effects. Finally, the seminar ends with an introduction to dynamic structural equation modelling (DSEM) covering univariate autoregressive models (with random residual variances, adjusting for unequal time intervals, and single- and multiple-indicator measurement models), multivariate vector autoregressive models, and dynamic models with indirect effects.