Models for Discrete Choice Experiments

The package includes the following models:

  • Multinomial (conditional) Logit (MNL)
  • Mixed (random parameters) Logit (MXL)
  • Generalized Multinomial Logit Model (GMXL)
  • Latent Class (LC)
  • Latent Class Mixed Logit (LCMXL)
  • Multiple Indicators Multiple Causes (MIMIC)
  • Hybrid Multinomial Logit (HMNL)
  • Hybrid Mixed Logit (HMXL)
  • Hybrid Latent Class (HLC)

The models are estimated using maximum likelihood method and work with the following specifications:

  • preference or WTP space
  • multiple distribution types (for random parameters)
  • non-linear transformations of explanatory variables
  • covariates of means, scale, and scale variance (where applicable)
  • impose equality restrictions or constraints
  • flexible data types (panel structure, non-constant number of choice tasks or alternatives per respondent, missing data)
  • various estimation and numerical optimization algorithms and options
  • parallel computing
  • and more ...


The latest Matlab version is available from GitHub