NLOGIT is an extension of LIMDEP that provides programs for estimation, model simulation and analysis of multinomial choice data, such as brand choice, transportation mode, and all manner of survey and market data in which consumers choose among a set of competing alternatives. NLOGIT has become the premier package for estimation and simulation of multinomial discrete choice models.

NLOGIT: Statistical Analysis Software

LIMDEP and NLOGIT are integrated statistical analysis software programs. They contain a large array of tools for data analysis, data management and model building from simple linear regression to maximum likelihood estimation of nonlinear systems of equations, with many extensions and variations.

NLOGIT 6 provides programs for estimation, simulation and analysis of multinomial choice data, such as brand choice, transportation mode, and all manner of survey and market data in which consumers choose among a set of competing alternatives. Since its introduction nearly 20 years ago, NLOGIT has become the premier statistical package for estimation and simulation of multinomial logit models including willingness to pay and best/worst modeling. NLOGIT supports mixing stated and revealed choice data sets.

Analysis Tools for Multinomial Choice Modeling

NLOGIT statistical software provides a wide array of tools for analysis of multinomial logit models, including nested logit, generalized mixed multinomial logit, heteroscedastic extreme value, multinomial probit, mixed logit and more. It is a simulation package that allows you to analyze alternative scenarios in the context of any estimated discrete choice model with any data set, whether used in estimation or as hold out data for examining model cross validity.

Features of NLOGIT 6 include:

With over 200 built-in estimators, you can analyze:
  • Four level nested logit models
  • Random parameters mixed logit
  • Latent class
  • Multinomial probit
  • Panel data - fixed effects MNL
  • Stated choice experiments
  • Willingness to pay
  • Heteroscedastic extreme value
  • Best/worst modeling
  • Random regret
  • Attribute nonattendance
  • Estimation and simulation
and much more In addition to the estimation programs, NLOGIT provides:

  • Data management, including input from all standard sources (such as Excel), all manner of transformations and sample controls
  • Built-in estimation programs plus a programming language, matrix algebra package and scientific calculator that allow you to write your own estimators, test statistics and simulation and analysis programs
  • Random number, vector and matrix capabilities for bootstrapping, Gibbs sampling and Monte Carlo simulation
  • A wide range of graphical and numeric descriptive statistics capabilities
  • Optimization tools that allow you to construct your own likelihood, GMM, or maximum simulated likelihood estimators
  • Analysis tools including graphics, numerical analysis and post estimation tools for specification and hypothesis testing
  • An extensive PDF documentation set, with over 3,000 pages, containing full reference guides for the programs, background econometrics, and sample applications

All the new features described for LIMDEP 10 are in NLOGIT 5. In addition, there are many new features in Version 5. We have added several enhancements to give you greater flexibility in analyzing different types of data. Many of the features of NLOGIT, existing and new, are designed to let you go beyond just computing coefficients, to analyzing and using your model. We have added many new models including the generalized mixed logit model, latent class mixed logit, randomly scaled MNL, random regret and a new multinomial choice modeling engine that allows arbitrary nonlinear utility functions. NLOGIT 5 pioneers the new developments for estimation in ‘WTP space.’ The mixed logit (random parameters logit) model is currently the most general and flexible model available for analyzing individual choice. Altogether, we have added dozens of features in NLOGIT 5, some clearly visible ones such as the new models and some ‘behind the scenes’ that will smooth the operation and help to stabilize the estimation programs. The following will summarize the important new developments.

New Multinomial Choice Models

We have added two major model classes to the package.

Generalized Mixed Logit Model

The generalized mixed logit (GMXL) model accommodates both random parameters in the utility functions and randomly variable scaling of the entire preference structure. The GMXL model is at the frontier of mixed multinomial choice modeling and NLOGIT provides many different variations on the model.

Nonlinear Utilities in the Mixed Logit Model

NLOGIT 5’s random parameters (mixed) logit model is extended to allow nonlinear utility functions. This capability vastly generalizes the model – utility functions may be any nonlinear or linear function that can be specified using the program syntax for nonlinear models.

Innovations in Multinomial Choices

New model frameworks include several innovations:

  • Latent class model with random parameters in each class
  • Scaled multinomial logit
  • Random regret MNL – this model explores an alternative to utility maximization as the model basis
  • Attribute nonattendance – this model accommodates the latent possibility that some respondents do not attend to all attributes in making their choices
  • Estimation in ‘willingness to pay space.’ This approach to model estimation works around the problem of using ratios of estimates to estimate willingness to pay. The ratios can behave erratically when they are close to zero. Estimation in WTP space, via a nonlinear transformation of the model, circumvents the problem by making the WTP coefficient the structural parameters in the model

Model Extensions

Utility Scaling

Heterogeneity in preference structures may take the form of general scaling of the entire utility framework. NLOGIT 5 provides general scaling in the multinomial logit model, the latent class model and the generalized mixed model. All of these can be layered into models based on stated choice data.

Mixed Logit Models

The mixed logit model represents the frontier in multinomial choice modeling. We have added many new features to NLOGIT’s already major implementation of this model. A partial list includes:

  • The latent class model may now have random parameters in each class
  • The generalized mixed logit model allows random parameters and random scaling of the entire preference structure
  • Heterogeneity in random parameters and generalized mixed logit models may appear in the variances as well as the means

Partial Effects and Elasticities

Elasticities have been reformatted so that tables may be exported to spreadsheet programs such as Excel. The results in the figure we exported directly to Excel. Elasticities may also be formatted as matrices to analyze within NLOGIT or export to other programs.

The NLOGIT 5 Reference Guide

We have completely reworked the NLOGIT manual. The new electronic format is portable and easily searchable. There are two major extensions. First, we have included in this manual documentation of the foundational discrete choice models described in detail in the LIMDEP Econometric Modeling Guide, including binary choice and ordered choice models. These are presented here to develop a complete picture of the use of NLOGIT to analyze data on discrete choices. Second, we have included extensive explanatory text and dozens of new examples, with applications for every technique and model presented.

NLOGIT is written for use on Windows driven microcomputers. NLOGIT is generally compatible with Windows 95 and later Windows operating systems. For Windows Vista, the Help systems used by NLOGIT are no longer installed as standard in Windows Vista, and Microsoft has released an update.

The program require 16 megabytes of memory and occupy approximately 10 megabytes on the local hard disk drive.

Coming Soon. Please email us at [email protected] to order your license NLOGIT

Post your comment

Timberlake Consultants