.png){kind=icon}
.png){kind=icon}
Return
Flexible and modular latent transition analysis-A tutorial using R
.png){kind=icon}
OA
AI
DOI：10.1371/journal.pone.0317617.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Latent transition analysis (LTA) is a useful statistical modelling approach for describe transitions between latent classes over time. LTA may be characterized in terms of prevalence at each time point and through transition probabilities over time. Investigating predictors of these transitions is often of key interest. Currently, LTA can mostly be carried out using commercial and specialized software and only to some limited extent by means of open source statistical software. This tutorial demonstrates a flexible and modular approach for LTA, providing a powerful alternative using R through a combination latent class analysis and multiple logistic regression models. This approach has several advantages from a modelling perspective, as demonstrated through revisiting a previously conducted LTA, published in PLoS ONE recently. In short, results were very similar to the original analysis using commercial software although some additional novel results were also obtained. The proposed alternative approach offers more options in terms of choice of effect measures, model assumptions such as hierarchical structures and covariate adjustment, and differential handling of missing data. R code snippets are provided in the tutorial. A detailed accompanying script is also provided for full reproducibility.
Keywords:
CHAINED EQUATIONS
PACKAGE
IMPUTATION