|From||Halvorsen Thomas <Thomas.Halvorsen@sintef.no>|
|Subject||st: interpreting the coefficient of the inverse Mills' ratio|
|Date||Sun, 11 May 2003 20:48:36 +0200|
I know this is not a Stata-specific question, but I know no other forum better able to answer this.
I'm running a two-step Heckman procedure creating an inverse Mills' ratio (IMR) variable with the "mills" option.
My question is how the coefficient of this variable is interpreted when used as an ordinary independent variable in the regression equation (second step)?
Is this a correct interpretation?:
When the coefficient of IMR is positive there are unobserved variables that both increase the probability of selection and a higher than average score on the dependent variable. When the coefficient of IMR is negative there are unobserved variables increasing the probability of selection and the probability of a lower than average score on the dependent variable.