T H E M O D E L
GANZ RASCH covers two models for categorical data analysis, the Rasch-Model (RM) and the Latent-Class-Model (LCA).
The Rasch Model (RM; Rasch, 1960, 1966) is a probability model for a dichotomous random variable involving the two real-valued latent parameters theta_v (v = 1 ... n) and beta_i (i = 1 ... k). Due to its mathematical properties, the RM allows for a rigid assessment whether model assumptions hold for a given data set.
In contrast, the LCA assumes the latent variable to be categorical, hence allows to classify respondents into the so-called latent classes. It can be viewed as a kind of cluster analysis for categorical data involving a latent factor.
Online resources (offering numerous links)
|ra pinxit — Mail to Webmaster|