Below, we describe the major new features in the subsequent versions of LatentGOLD. We start with version 2.0 because version 1.0 from 1999 was a beta version that was not distributed.

2000: Version 2.0

  • Cluster module
  • DFactor module
  • Regression module

2003: Version 3.0

  • Choice module
  • Options for saving models and output
  • Enhanced algorithms
  • Extended missing value options
  • Bootstrap L2 goodness-of-fit statistic
  • Inactive covariates option
  • R2 measures for indicators and dependent variables
  • Parameter restrictions in Regression module

2005: Version 4.0

  • Complex sampling (or Survey) options
  • Multilevel latent class analysis
  • Continuous latent variables (CFactors)
  • Knownclass option
  • Bootstrap -2LL difference statistic (BLRT)
  • Various order restrictions on parameters
  • Zero-inflated models
  • Linkage with SI-CHAID program

2008: Version 4.5

  • Syntax module
  • Latent Markov models
  • (Mixture) factor analysis and IRT
  • Flexible parameter restrictions
  • Imputation options
  • Monte Carlo simulation options

2013: Version 5.0

  • Major speed improvement by parallel computing
  • Step3 module
  • Markov GUI module
  • Obtaining scoring equations
  • Selection and holdout options
  • Flexible Step3 modeling via Syntax
  • Continuous-time latent Markov models
  • Log-linear scale models
  • Score tests
  • User-defined Wald tests
  • Power computation for Wald tests
  • Expanded bivariate residuals

2016: Version 5.1

  • 64-bit version
  • Paired comparison Wald tests
  • Power computation for likelihood-ratio tests
  • Standard error correction for Step3 models
  • Expanded (multiple) imputation options

2021: Version 6.0

  • LC tree models
  • Three-step LCA with differential item functioning (DIF)
  • Bakk-Kuha stepwise LCA
  • Mixture and multi-group exploratory factor analysis
  • B-spline effects for independent variables
  • Vuong-Lo-Mendell-Rubin tests
  • Marginal effects output
  • User-defined expressions

2025: Version 6.1

  • New licensing system
  • Three-step analysis with continuous latent variables
  • (Mixture) structural equation modeling (SEM)
  • Parametric survival models
  • Markov option in Step3 GUI
  • Expanded batch mode options