Management of missing data in Stata and R in SEM for longitudinal data

[6 hours] - [English]

You use SEM and your database contains missing values?

Multivariate confirmatory analyses are applied when a theoretical model already exists before data collection (for example, the direction of causal effects between variables is supposed known and well discussed).  This type of statistical analyses is of particular interest when a researcher adopts an explanatory approach.

Many databases contain missing values for certain variables of interest. Ignoring them during analysis (by only including individuals with no missing values) can, in many cases, lead to biased results. Methods exist to take account of the presence of missing values in the analyses of interest and thus correct the results obtained.

Training aims

The aim of this course is to introduce the solutions available in R and Stata for estimating structural equation models in the presence of missing data. The particular case of the Cross-Lagged Panel Model applied to longitudinal data will be discussed.

Prerequisites
Before registering for this training course, be sure you already have a level of knowledge equivalent to the following course(s):

Content

• General information on the types of missing data and their possible consequences.
• Solutions for SEMs (in particular Full Information Maximum Likelihood (FIML) and Multiple Imputation (MI)) and their use in Stata and R.

Rate

Tools used during training
R / Stata

Methods and method families discussed
Multivariate confirmatory analysis
   SEM - Structural equation model
   Path analysis