Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Heckman sample selection with count and censored outcome equations

From   P AG <>
To   <>
Subject   st: Heckman sample selection with count and censored outcome equations
Date   Thu, 19 Apr 2012 15:31:07 +0000

Hello Statalisters,

I have two different equations I would like to estimate which have sample selection in the first stage which I woul like to account for in the vein of Heckman.

In the first I would like to estimate the number of items selected from a menu, which I only observe for those who choose to make a selection. About 80% choose not to make a selection and receive a default item. For those who do make a purchase most select just one item on the menu. My main outcome equation of interest is how many items are selected from the menu, the size of which varies for groups of participants. I would like to estimates a zero truncated poisson or negative binomial count regression. 

In the second equation, which focuses on a different aspect of the first purchase, I would like to estimate the percentage of foreign owned merchandise in the the total items selected, which will range from 0 to 100 with about 10% of observations at both censoring points. I would like to use a censored regression for this.


I have seen posts on the list for <ssm> which seems to suggest it can estimate a count regression with sample selection which would satisfy the first, but it doesn't seem to in fact do so. I haven't seen anything matching the second in the archive.


I am using Stata 12.


Thanks for any advice.

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index