CLICK HERE to download a zipped file with all the data files, programs and output listed below.

DATA: We thank the authors of the papers listed in the table below for making their data available for empirical illustrations.

STATA: Stata was used for most of the book and the programs reproduce virtually all of the book.
Some Stata programs require user-written addons:
  fmm for finite mixture models (chapter 4, 6, 11)
  countfit for predicted probabilities (chapter 3, 5, 7)
  mtreatreg for multiple treatment models (Chapter 10).
  rcal, simex, qvf and cme for measurement error (Chapter 13).

R: R was used for the nonparametric and semiparametric analysis of chapter 11.
A fairly detailed R program is also given for chapter 3 (basic Poisson and negative binomial regression).
Short R programs to read in data and estimate by Poisson are given for the other chapters.
These usually do not give exactly the same numerical results as the Stata code. e.g. robust standard errors are calculated differently.
The R programs use packages
  foreign to read in a Stata dataset
  sandwich for robust sandwich standard errors
  boot for bootstrap standard errors
  MASS for negative binomial (NB2) regression
  gamlss for negative binomial (NB1 and NB2) regression
  pscl for predicted probabilities from Poisson and NB2 regression
  flexmix for finite mixtures of Poisson
  np for nonparametric and semiparametric estimation

Back to Main Page