Major New Features in Latent GOLD 5.0

• Much faster estimation time
• Improved profiling of classes (see Step3 module)
• Obtaining equation for scoring new cases (see Step3 module)
• New GUI for Latent (Hidden) Markov / Latent Transition Modeling
• New log-linear scale model
• and more!

Faster Estimation Time

Latent GOLD 5.0 supports multiple processors, making estimation faster than ever! 
 

Latent GOLD Basic

• Step3 Module. (See Tutorials) After developing a segmentation model (Step 1), and classifying cases (Step 2), you can now use the latent class segments in followup analyses with the new Step3 module.
  • Get exact equation for scoring new cases
  • Properly adjust for misclassification error
  • Predict classes from exogenous variables
  • Predict exogenous variables from classes
• Select Cases. The 'Selection Variable' option makes it possible to select a specific subset of cases/records from the data file for analysis.
• New statistics:
  • SABIC(LL,N), SABIC(L2,N), and total BVR
  • Classification table for proportional assignment
• Bootstrap p values:
  • Not only of L2, but also for X2, CR2, DI, total BVR, and BVRs
  • In 2LLdif bootstrap, also L2 for H1 model
  • Critical values are reported in addition to p values

Advanced Option

• Latent (Hidden) Markov GUI. (See Tutorials) How do clusters change over time? Do persons change from one latent class to another over time? Do some persons change (movers) and others not change classes (stayers)? You can now use the new latent Markov GUI to analyze your longitudinal data to address these and many other important research questions. This Markov module is light years ahead of other programs!
  • Unique implementation, not available elsewhere
  • Fastest of all latent Markov programs
  • Quickly analyze even hundreds of time points (other programs can only handle a few time points)!
  • Unique interactive graphics
  • Mixture latent Markov containing multiple indicators
  • Mover-stayer structures
• New statistics. Based on higher-level sample size for multilevel LC models: BIC(LL,K), CAIC(LL,K), and SABIC(LL,K)

LG-Syntax Module

• Additional Scale Types. Mixture regression models for dependent variables with other distributions (gamma, beta, von Mises).
• Scaling Models. New log-linear scale model for categorical dependent variables.

Features

• Full windows implementation - point and click
• Interactive graphics provide new insights into data and powerful model diagnostic capabilities
• Flexible model structures can handle variables of different metrics
• Automatic generation of sets of random starting values
• Fast, efficient maximum likelihood and posterior mode estimation based on EM and Newton Raphson algorithms
• Use of Bayes constants to eliminate boundary solutions
• Bivariate residual diagnostic for local dependencies