ABSTRACT
Latent GOLD® is developed by Jeroen Vermunt and Jay Magidson, with Windows programming assistance from Alexander Ahlstrom. It was designed originally with an SPSS-like point-and-click GUI interface, implementing the most important types of latent class and mixture models. The initial release in 2000 (version 2.0) contained three modules: Cluster, Discrete Factor (DFactor), and Regression, representing latent class (LC) analysis with a single nominal latent variable, LC with multiple ordinal latent variables, and mixture regression analysis of two-level datasets. Each of these modules could accommodate response variables of different scale types (e.g., continuous, categorical). The DFactor module is in fact an IRT routine implementing multidimensional discretized IRT models with fixed nodes for the discretized traits.
