When: June 9th between 16:00 and 18:00 hrs.
Where: Communication Department, VU.
To register: If you are part of VU, then fill out this form: https://forms.office.com/e/uciB7nhjS5. Otherwise, send us an email: analytics-lab.fsw@vu.nl
Extra: With pizza and drinks at the end.
Requisites:
- Bring your own computer.
- Basic R knowledge is necessary.
- Install the package latent. Follow the installation instructions from GitHub before the workshop https://github.com/Marcosjnez/latent.
Description
In this workshop I will present the basic concepts of LCA and its application with the
latent
R package. LCA is a Categorical Latent Variable Model, where we identify "unobserved" groups in function of response patterns from a set of indicators. As an exploratory method, we test multiple solutions and evaluate the statistical and theoretical relevance to select the "best" solution. We will see how to run LCA with only categorical, only continuous, and combination of item types. We will then discuss how to choose the "best" number of classes, model fit evaluation, and result interpretation. Lastly, we will discuss how to do regressions on the latent classes with the 2-step approach.
Mauricio Garnier-Villarreal, PhD
Associate professor at the Sociology department comes from the field of Quantitative Psychology, with two main areas of research. First, he works on the development and test of advance data analysis methods, as well as the development of open-source software for its application. Second, he collaborates in applied fields to apply these advance data analysis methods, as part of this he has work in areas like labour market, Alzheimer's, special education, physical therapy, counselling, nursing, and more. He is a contributor in the development of several R packages, like blavaan, semTools, nonnest2, and latent.
