Home  /  Resources & support  /  FAQs  /  Stata 6: Difference between biprobit/heckman and previous STB commands
Note: This FAQ is for users of Stata 6. It is not relevant for more recent versions.

Stata 6: What is the difference between biprobit/heckman and the STB commands?

Title   Stata 6: Comparing STB-33 routines and the Stata 6.0 biprobit and heckprob commands
Author James Hardin, StataCorp

There are several enhancements to the commands for fitting bivariate probit models (biprobit) in Stata 6.0.

  1. One can use weights; fweights, iweights, and pweights are allowed.
  2. There are a number of predictions that are allowed:
    p11 Pr(depvar1=1, depvar2=1); the default
    p10 Pr(depvar1=1, depvar2=0)
    p01 Pr(depvar1=0, depvar2=1)
    p00 Pr(depvar1=0, depvar2=0)
    pmarg1 Pr(depvar1=1); marginal probability equation 1
    pmarg2 Pr(depvar2=1); marginal probability equation 2
    pcond1 Pr(depvar1=1 | depvar2=1)
    pcond2 Pr(depvar2=1 | depvar1=1)
    xb1 linear prediction equation 1
    xb2 linear prediction equation 2
    stdp1 standard error of xb1
    stdp2 standard error of xb2
  3. The correlation parameter is restricted to [-1,1] interval.
  4. The nested model from the STB has been incorporated into the new heckprob command. This new command also offers many predicted values after estimation including
    pmargin predicted probabilities; Pr(depvar=1); the default
    p11 Pr(depvar=1, selvar=1)
    p10 Pr(depvar=1, selvar=0)
    p01 Pr(depvar=0, selvar=1)
    p00 Pr(depvar=0, selvar=0)
    psel Pr(selvar=1)
    pcond Pr(depvar=1 | selvar=1)
    xb linear prediction
    stdp standard error of xb
    xbsel linear prediction in selection equation
    stdpsel standard error of xbsel