Data Environments

LIMDEP and NLOGIT offer a wide range of capabilities for cross section, time series and panel data sets.

Cross Section

There are well over 100 preprogrammed models and procedures for cross section analysis, including descriptive statistics, linear modeling, discrete choice, limited dependent variables, count data, hazard modeling, and so on.  Many special features for matrix algebra based on cross sections, as well as the nonlinear optimization routines allow you to program extended analyses and develop new procedures and model estimators.

Time Series

LIMDEP provides many tools for time series, including ARIMA and ARMAX models and Box-Jenkins identification, estimation and testing procedures.  GARCH models are an extension of the linear model.  More advanced procedures include unit root tests and spectral density estimation.  Dynamic linear models are accommodated using instrumental variable procedures or with one of the panel data estimators.

Panel Data

No program provides a greater range or variety of procedures for panel data analysis.  In addition to common effects (fixed and random) specifications for nearly all of the cross section models, LIMDEP contains latent class and random parameter treatments, again for all models supported by the program.  Among the many other panel data features are:

• Many data transformations, such as group means and deviations, are specific for panel data.
• Bootstrapping may be expanded to panels.
• Several matrix algebra functions are specifically written for panel data.