Discrete Choice Models in LIMDEP
LIMDEP provides a wide range of estimators for discrete choice modeling including many specifications of binary choice models, multinomial choice, ordered choice and the widest selection of models for count data in any package.
Model frameworks
- Binary choice models
- Bivariate probit models
- Multinomial (logit) choice models
- Ordered choice models
- Bivariate ordered choice models
- Multivariate binary choice models
- Simultaneous equations and two step estimators
- Count data models
Tools
- Specification analysis
- Heteroscedasticity
- Robust inference tools
- Lagrange multiplier, likelihood and Wald tests
- Model simulator for binary choice models
- Matching and propensity scores analysis
- Average partial effects
- Partial effects for interactions
- Model simulation and prediction
Statistics
- Numerous fit measures
- Test statistics for specifications
- Marginal effects for all models
- Interaction terms in model specification
- Fixed effects for all models
- Conditional fixed effects
- Random effects for all models, quadrature and simulation estimators
- Random parameters models
- Latent class specifications
Specifications
- Count data models: censoring, truncation, underreporting, zero inflation, negative binomial (NB1, NB2, NP-P), generalized Poisson, Polya-Aeppli, gamma
- Binary choice: probit, logit, Weibull, others; heteroscedasticity, sample selection models
- Multinomial logit: random effects, latent classes, random utility models, model simulations
- Bivariate probit models: recursive simultaneous equations, marginal effects, heteroscedasticity
- Sample selection models
- Ordered choice: bivariate, sample selection, hierarchical, generalized ordered probit, marginal effects, panel data models