Přehled o publikaci

Advancements in statistical modeling tools have enabled researchers to delve into complex analyses more effortlessly. This presentation introduces shinyLavaan, a user-friendly Shiny R application for seamless confirmatory factor analysis (CFA) utilizing the powerful lavaan R package. This presentation aims to showcase an intuitive and versatile tool that simplifies the process of specifying and estimating CFA models using lavaan syntax. Users can define their CFA models through an interactive interface, customizing various settings such as the choice of estimator, handling of missing values, and treating indicators (as continuous or ordinal variables). In addition, the application also allows to conduct multigroup analysis. Output is presented in clear tables and insightful plots. As for tables, users can obtain parameter estimates, global fit measures, and information on local lack of fit such as residual correlations and modification indices. The application also generates diagnostic plots, including hopper plots (revealing the strongest residual correlations), structural plots (illustrating relationships between latent variable estimates), and measurement plots (to assess if the model-implied relationship between a factor and its indicator matches the empirical one). An interactive path diagram enhances the visualization of the overall model structure. The application also provides confidence intervals (or bands) for both tables and plots, and users can specify the confidence level they want to work with. In summary, shinyLavaan makes CFA modeling more accessible. Researchers and practitioners can benefit from a user-friendly interface that facilitates model specification, estimation, and comprehensive diagnostics.